AUTHORS

Shankar Mitra ConocoPhillips School ofGeology and Geophysics University of Okla-homa Norman Oklahoma 73019smitraouedu

Shankar Mitra holds the Monnett Chair andProfessorship in Energy Resources at theUniversity of Oklahoma He received his PhDfrom the Johns Hopkins University in 1977

Debapriya Paul ConocoPhillips School ofGeology and Geophysics University of Okla-

Structural geometry andevolution of releasingand restraining bendsInsights from laser-scannedexperimental modelsShankar Mitra and Debapriya Paul

homa Norman Oklahoma 73019debapriyapaul-1ouedu

Debapriya Paul is a PhD student at the Uni-versity of Oklahoma He received his BScdegree from the University of Calcutta his MScdegree from IIT and his MS degree from theUniversity of Oklahoma

ACKNOWLEDGEMENTS

We thank Sandro Serra Hong bin Xiao RickGroshong Al Lacazette and Ron Nelson for theircomments and suggestions and Shamik Bosefor help with the laser scanning We also ac-knowledge Paradigm for the university licenseof GO-CADThe AAPG Editor thanks the following reviewersfor their work on this paper Richard HGroshong Alfred Lacazette and Ronald ANelson

ABSTRACT

Experimental modeling is used to study the geometry andevolution of structures and related secondary faults along re-leasing bends and offsets and restraining bends on strike-slipfaults The controls of the relative positions of adjacent strike-slip faults on the geometry of the structures and the differencein geometries between bends and offsets are investigated Anew method of laser scanning is used to map the geometryand evolution of the structures and related faults The modelsshow that oblique releasing bends connecting approachingfaults result in spindle-shaped basins whereas transverse bendsresult in more S-shaped or rhomboidal basins Offsets resultin the distribution of strain over a wider area and a largernumber of faults compared with preexisting bends whichresult in fewer well-defined basin-bounding faults Secondaryfaults include R Rprime and Y Riedel shears near the main strike-slip faults and oblique normal faults in the center of the basinFault patterns exhibit en echelon geometrieswith a progressivestep down into the deepest parts of the basin Symmetricasymmetric and double basins may form in any of the struc-tural settings depending on the slip distribution among faultson the basin margins For restraining bends oblique (45deg)bends connecting approaching faults result in spindle-shapeduplifts whereas transverse or oblique (135deg) bends connectingoverlapping faults result in more rhomboidal or rectangularuplifts The fold trends are at increasingly higher angles with

Copyright copy2011 The American Association of Petroleum Geologists All rights reserved

Manuscript received April 12 2010 provisional acceptance June 2 2010 revised manuscript receivedAugust 2 2010 final acceptance September 27 2010DOI10130609271010060

AAPG Bulletin v 95 no 7 (July 2011) pp 1147ndash 1180 1147

the strike faults for transverse and oblique (135deg)bends Secondary faults include en echelon reversefaults which typically form along the steep limbsof asymmetric uplifts normal faults which aretransverse or oblique to the axis of the structureand R Rprime and Y Riedel shears near the main strike-slip faults The aspect ratios of the basins and up-lifts increase with increasing displacement on thestrike-slip faults The results of thesemodels can beused to interpret the structural and fault geom-etries in surface and subsurface structures formedalong strike-slip faults

INTRODUCTION

Strike-slip faults are commonly characterized bybends or offsets marked by faults or fault systemswith an oblique or transverse orientation to theirgeneral trend (Figure 1) A basement strike-slipfault may change position along a bend (Figure 1A)or by an offset of two separate fault segments(Figure 1B) The bends may be oblique or trans-verse to the main fault Depending on their ori-entation relative to the sense of strike slip bends oroffsets in strike-slip faults are referred to as releas-ing or restraining (Figure 1C D) Releasing bendsor offsets form where the sense of shear on a strike-slip fault results in a component of fault-parallelextension (Figure 1C) whereas restraining bendsor offsets form where the shear results in a com-ponent of fault-parallel contraction (Figure 1D)

The types of secondary structures and theirorientations relative to the main fault were firstdescribed by Crowell (1974) and used to explainpull-apart basins and oblique compressive struc-tures mapped by various authors in Californiabasins Subsequently these models have been ap-plied to explain structures in a variety of basinssuch as in the collection of articles compiled bySylvester (1984) Biddle andChristie-Blick (1985)and Cunningham and Mann (2007)

Most studies of restraining and releasing bendsare based on surface seismic or remote-sensed dataBecause data from all of these sources are incom-plete the interpretations are commonlymade basedon simplified conceptual models of these struc-

1148 Structural Geometry and Evolution of Releasing and Rest

tures Knowledge of the geometry of folds andfaults formed in these settings and their progressiveevolution through time would be useful in inter-preting these structures in areas of incomplete dataOne source of information for the detailed in-formation is scaled experimental models of thesestructures

The only published examples of restrainingbends are the sandbox models of McClay andBonora (2001) that investigated the shapes andevolution of structures and faults along preexist-ing jogs modeled above sliding base plates Theseauthors recognized the importance of document-ing the structural relief associated with strike-slipstructures and attempted to do so by buildingpseudondashthree-dimensional (3-D) models fromspaced serial sections Several experimental mod-els of releasing bends have been published Theseinclude the sandboxmodels ofDooley andMcClay(1987) Rahe et al (1998) Basile and Brun (1999)Wu et al (2009) and the clay models of Atmaouiet al (2006) The first four articles all used precutjogs in sliding plates to simulate strike slip whereasAtmaoui et al (2006) used both sliding plates anddistributed shear in their experiments DooleyandMcClay (1997) andWuet al (2009) attemptedto qualitatively explain the variation in structuralrelief resulting from the oblique deformation

The preexisting experimental models all usedsliding metal plates with bends along them to sim-ulate strike-slip deformation These experimentalconfigurations do not model deformation associatedwith preexisting offsets in the basement strike-slipfault which is expected to be quite different fromthat associated with bends Furthermore with theexception of Atmaoui et al (2006) who investi-gated the role of distributed shear in the develop-ment and intersection of Riedel shears they allmodeled the basement as perfectly rigid with thetransition from strike-slip to oblique deformationoccurring along a single preexisting jog

In this article we present experimental claymodels of both restraining and releasing bendswith various geometries of fault bends and faultoffsets The fundamental difference between theexperimental configuration used in these experi-ments with previous models is that the basement

raining Bends

is modeled with stiff clay which is significantlymore rigid than the overlying soft clay represent-ing the sedimentary cover but which allows for asmall component of deformation of the basementespecially in the immediate vicinity of the majorfaults Sliding metal plates which underlie the base-

ment are used only to control the nature of de-formation and the position of the bend or offsetPreexisting vertical cuts in the stiff clay define theinitial geometry of the basement faults This con-figuration also enables us to study the differencebetween releasing offsets and bends which result

Figure 1 (A B) Possible basement faultconfigurations for strike-slip faults (A) Bend(B) Offset (C) Releasing bend (D) Re-straining bend (E) Initial basement faultconfigurations used in experiments forreleasing bends and offsets (F) Initial base-ment fault configurations used in experi-ments for restraining bends Figures 2 and13 show details of the experimental setupfor (E) and (F)

Mitra and Paul 1149

in very different geometries of structures and faultpatterns in pull-apart basins (Figure 1) The specificexperimental configurations depicted in this studyare summarized in panels E and F of Figure 1 anddescribed in more detail in the following sections

The geometry and evolution of secondary faultsfor the top surface are studied in detail In addi-tion using a new technique of laser scanning westudy the mutual relationships between structuralrelief and faulting with progressive deformation incontoured models for the surface The renditionsof the quantitative geometry of the deformed sur-faces are directly comparable to structure contourmaps of natural structures and therefore of mostinterest to petroleum geologists

1150 Structural Geometry and Evolution of Releasing and Rest

The results of the models are compared withnatural examples of releasing and restraining bendsto show how detailed interpretations of the struc-tures can be improved through an understandingof the evolution of the structures

RELEASING BENDS AND OFFSETS

Experimental Approach

Clay experiments are conducted to simulate theformation of releasing bends and offsets in strike-slip faults The main advantage of using clay asthe model material is that it allows a much better

Figure 2 Experimental models conductedfor releasing bends and offsets (A B)Oblique 45deg bend in basement fault(C D) Transverse 90deg bend in basementfault (E F) Transverse 90deg offset inbasement fault For each experiment thefirst figure shows the relationship betweenthe underlying sliding plates and thesecond figure shows the basement andsedimentary cover and the preexisting faultcuts in the basement

raining Bends

definition of fault geometries and evolution and alsoenables the deformation to be studied for ma-ture structures involving steep surfaces The ex-perimental setup consists of two overlapping metalplates one of which is attached to a moving platedriven by motors whereas the other is attached toa fixed plate The lateral motion between the twobase plates drives the strike-slip motion in the over-lying stiff and soft clay representing the basementand sedimentary layers (Figure 2) No shearing oc-curs along the contact between the stiff clay andthe sliding plates Basement is represented by stiffclay with a density of 185 gcc and a thickness of1 in (25 cm) The overlying sedimentary layer isrepresented by soft water-based clay with a densityof 16 to 165 gcc and a thickness of 08 in (2 cm)The density of the clay is adjusted by altering thewater content of the clay so that the clay has theappropriate consistency to deform by faulting with-out developing large cracks No layering is usedwithin the soft clay The displacement rate for allexperiments is 0001 cms corresponding to a strainrate of approximately 3 times 10ndash5 per second

The experimental configurations used in thisstudy differ from previous setups (Dooley andMcClay 1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006Wu et al 2009) in thatstrike-slip motion is induced in both the stiff clay(basement) and the overlying soft clay (sedimentarycover) through sliding base plates at depth There-fore the sliding plates only control the nature ofdeformation and the location of the bend or offsetThis approach enables us to study differences indeformation involving releasing bends and offsetsin the strike-slip faults within the basement whereasprevious experiments modeled only releasing bendsFurthermore using stiff clay to model basementinstead of rigid plates the basement deforms in theimmediate vicinity of the faults to accommodatethe deformation in these areas In previousmodelsthe top of the basement was represented by slidingplates rendering the basement perfectly rigid inthe vicinity of the faults

Two experimental setups for the master faultsin the basement are investigated (1) bends inwhichtwo parallel strike-slip faults are connected by anoblique (Figure 2A B) or transverse (Figure 2C D)

bend in the stiff clay representing the basementand (2) offsets in which two separate strike-slipfault segments in the basement are separated by a90deg offset in the stiff clay (Figure 2E F) The initialconfiguration for the bend setup consists of twovertical cuts in the stiff clay connected by obliqueor transverse bends whereas the configuration forthe offset setup consists of two parallel and offsetvertical cuts In both cases the basement is under-lain by two sliding plates that override each other toprovide the strike-slip motion (Serra and Nelson1988) The bend configuration may represent anoriginal geometry of the strike-slip fault (Crowell1974) or may result from two separate parallel seg-ments that have connected along a transverse oroblique segment The offset configuration morespecifically addresses the linking of two adjacentstrike-slip segments by the development of a trans-fer zone consisting of single or multiple obliquefaults These configurations result in significantdifferences in the geometry and faulting associatedwith the pull-apart basins Two runs of each ex-periment are conducted to test for repeatability

The progressive evolution of structures and sec-ondary faults and their orientations within a strike-slip system is mapped directly from photographstaken during various stages of the experiments Inaddition we use a new technique involving theanalysis of the 3-D geometry of the top surface inthe structures using laser scanning (see also Boseand Mitra 2010) The technology of scanning sur-faces by a laser scanner and developing a virtual3-D model has been in use in numerous industriesparticularly in 3-D animation The same technol-ogy is applied here to gather 3-D information andthereby generate a virtual surface that can be used tovisualize the development of the structure in muchgreater detail and to generate contour maps of sur-faces The scanning size is 135 times 101 in (343 times257 cm) with a resolution of 75 dots per inch(sim0015 in point density) The scanner is placedface down approximately 15 in (38 cm) verticallyabove the clay cake and top surfaces are scanned atapproximately equal increments of deformation

The sensor captures the laser beam that is pro-jectedon the surface and thecoordinatesof thepointsare computed by the scanner using a triangulation

Mitra and Paul 1151

method (Figure 3A) The known parameters arethe angle a at which the laser points out its beamon the surface the angle b at which the sensor col-lects the laser beam (b is known if the focal lengthand the pixel size of the sensor are given) and thetriangulation distance between the sensor and lasersource (Petrov et al 1998) Because all the geo-metric parameters are known the coordinates x yand z on the surface are calculated using trigono-metric methods The laser source produces a lineand all data along that line are collected at the sametime (Figure 3B) Moreover twin arrays of four laserbeams are used as source in this instrument for

1152 Structural Geometry and Evolution of Releasing and Rest

cross-validating each data point by measuring it atleast twice The origin or reference coordinate sys-tem is determined automatically during each run bythe first point scanned on the surface It is thereforenot necessary to move the scanner during the lengthof the experiment to spatially position the scannedsurfaces in the same reference coordinate system

The scanned data are exported as point files intothe 3-D modeling software GoCAD (Bose andMitra 2010) The point clouds obtained (Figure 3C)are then used to build a surface Fault polygons arebuilt from visible breaks in the modeled surface(Figure 3D) and the geometries of the polygons

Figure 3 Laser scanning of experimental clay models (A) Schematic diagram of the working principle of the laser scanner withthe triangulation formed by the laser source the detector and the point on the clay surface from which the laser beam is reflected(B) Schematic view of the laser scanner projecting a line and sweeping across the clay surface at a constant velocity (from Bose andMitra 2010 used with permission from AAPG) (C) Cloud of points obtained from the scanned surface and visualized in GO-CAD(D) Modeled surface in GO-CAD cut by faults (E) Oblique photograph of the top of the clay surface

raining Bends

are checked against the traced fault cuts on photo-graphs taken of the clay surface which provide de-tails of the fault geometry (Figure 3E) Themodeledsurface is cut by the fault polygons and contouredto depict their topology (Figure 3D) The contouredfaulted surface provides an accurate quantitativerendition of the structural geometry which can becompared directly with structure contour maps onnatural structures

Experimental Results

Oblique (45deg) BendInitial movement on the base plate results in a broaddepression oblique to the preexisting basementstrike-slip faults in the area above the obliquebend (Figures 4A 5A) This depression is boundedby two sets of faults dipping toward each otherand trending approximately 42deg to the basement

Figure 4 Clay models showing faults in the sedimentary cover for a releasing bend with an oblique (45deg) bend connecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Insets in this and other figures for experimentalmodels show initial fault cuts in the stiff clay Panels (AndashF) show the progressive evolution of the structure with increasing strike-slipdisplacement Rose diagrams show the lengths of faults for different trends in 5deg increments as a percentage of the total length of faults

Mitra and Paul 1153

Figure 5 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with an oblique (45deg) bend connectingtwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement Contours in this and subsequent diagrams are in intervals of 04 mm Every fifth contour is shown in black Color gridsreflect structural elevation with red and yellow representing high areas and blue and green representing low areas Note the formation oftwo diametrically opposite lows in the pull-apart basin (G H) Cross sections through top of clay showing final basin geometry and shift inpolarity for cross sections AAprime and BBprime Locations of the cross sections are shown in (F)

1154 Structural Geometry and Evolution of Releasing and Restraining Bends

strike-slip faults that define the boundaries of thepull-apart basin The angular measurements citedpreviously and in all experiments on pull-apart ba-sins are made in a counterclockwise direction fromthe strike-slip faults The individual faults are madeup of two or more right-stepping en echelon seg-ments some of which eventually interact to formlonger faults With progressive deformation thesefaults systems are rotated slightly by shear

Immediately above the main vertical strike-slipfaults short segments of steep faults of two mainsets trending approximately 15 and 80deg to the base-ment strike-slip faults develop (Figure 4B C) Thesefaults are interpreted to be the R and Rprime Riedelshears associated with strike-slip faulting describedand experimentally modeled by Tchalenko (1970)Tchalenko and Ambraseys (1970) Wilcox et al(1973) and Harding (1985)

Figure 6 Clay models showing faults in the sedimentary cover for a releasing bend with a transverse (90deg) bend connecting two strike-slip fault segments in the basement Sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution of thestructure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

Mitra and Paul 1155

Figure 7 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with a transverse (90deg) bend con-necting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slip displacement

1156 Structural Geometry and Evolution of Releasing and Restraining Bends

The normal faults in the central part of thebasin curve into the orientations of the R shearsin the vicinity of the basement strike-slip faults(Figures 4B C 5B C) The resulting fault systemshave a sigmoidal geometry Rose diagrams of faultorientations show the progressive developmentof scatter because of the development of curvedfaults with varying components of normal andstrike-slip motion (Figure 4AndashC) Twomajor strike-slip faults develop at the surface directly above thepreexisting cuts in the basement after about 08 cmof displacement

Continuing deformation results in additional setsof major normal faults inward and approximatelyparallel to the original pair of fault systems Addi-tional fault systems with the same overall sigmoidalgeometry form in other parts of the transfer zoneand outside the central depression (Figures 4BndashE5BndashE) In the final stages of deformation the lon-gest pair of oblique faults connects with the twomain systems of strike-slip faults thus defining theboundary of the pull-apart basin (Figures 4F 5F)The final rose diagram shows significant scatter oftrends ranging from north to west because of thevariation from strike-slip to normal fault trends witha maximum density in an approximately north-west orientation (Figure 4F)

Examination of the contour patterns (Figure 5)suggests that the structural elevation steps downquite steeply across the normal faults bounding thepull-apart basin but drops more gradually from thebounding system of strike-slip faults to the centerof the basin The central part consists of two sep-arate depressions separated by a local high thatdevelop during the late stages (Figure 5E F) Theaxis connecting the centers of the two depressionsis approximately parallel to the trend of the mainstrike-slip faults Cross sections through the twodepressions show a flip in polarity because of achange in fault displacement along trend withinthe basin (Figure 5G H)

Transverse (90deg) BendFor a transverse bend (90deg) the normal fault seg-ments in the transfer zone form at a higher angle(58ndash78deg) than in the case of the oblique jog (45deg)and show left- and right-stepping patterns on op-

posite sides of the transfer zone (Figures 6A B 7AB) Immediately above the basement strike-slipfaults R and Rprime Riedel shears develop as a series ofen echelon faults making angles of 15 and 82deg to themain faults respectively The normal faults in thecenter of the pull-apart basin curve sharply clock-wise toward the orientation of the R shears in thevicinity of the main strike-slip fault zones along theboundaries of the pull-apart basin (Figures 6C D7C D)With continuing shear the major strike-slipfaults break through to the surface as Y shears de-fining the boundaries of pull-apart basin (Figures 6EF 7E F) As in the case of the oblique-bend ex-periment multiple sets of faults form in the pull-apart basin especially close to the main strike-slipfault zone Theorientations of the faults showawidevariation in trends ranging from northndashnorthwest towestndashnorthwest with the most dominant trend inthe center of the basin being about northwest in thefinal stage (Figure 6F)

Because of the sharper bends and changes inthe trends of the faults in the transfer zone the pull-apart basin associated with a transverse bend has amore S-shaped or rhomboidal geometry than thespindle-shaped geometry in the case of an obliquebend The contours show that the deepest part ofthe basin is located toward one of the major transferfaults so that the pull-apart basin has a more asym-metric shape with a sharper drop of contours onone side (Figure 7E F) The sides of the basin aremarked by a more gradual change in elevation withthe structure contours cut by secondary faults Dif-ferent basin shapes may result depending upon theslip distribution among the different faults

Transverse (90deg) OffsetThis setup has two cuts in the basement that even-tually form the bounding strike-slip faults offsetalong a transfer zone The main purpose of con-ducting this experiment was to understand the dif-ference in the geometry of the pull-apart basinwhenno connecting bend exists in the basement faultThe faults in the transfer zone develop with on-going strike-slip deformation independent of a con-nected basement fault (Figures 8 9) Our experi-ments show that a transverse offset results in theoptimum configuration for the development of a

Mitra and Paul 1157

transfer zone of connecting faults Experiments in-volving overlapping or approaching patterns do notresult in well-defined pull-apart basins because theyboth involve a large transfer zone area This suggeststhat for two approaching strike-slip fault segmentsconnecting faults in the transfer zone will mostlikely start to develop when the end points of thetwo strike-slip segments are directly opposite eachother in a transverse offset position

1158 Structural Geometry and Evolution of Releasing and Rest

Initial deformation results in the formation ofa large number of short en echelon normal faults inthe area of the transfer zone (Figures 8A B 9A B)The central parts of the faults trend at angles of 53to 63deg to themain strike-slip fault zone At the sametime a series of en echelon R and Rprime Riedel shearsform above the main basement faults at angles of18 and 70deg with respect to the basement strike-slip faults respectivelyWith increasing strike-slip

Figure 8 Clay models showing faults in the sedimentary cover for a releasing offset with a transverse (90deg) offset between two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution ofthe structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

raining Bends

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

the strike faults for transverse and oblique (135deg)bends Secondary faults include en echelon reversefaults which typically form along the steep limbsof asymmetric uplifts normal faults which aretransverse or oblique to the axis of the structureand R Rprime and Y Riedel shears near the main strike-slip faults The aspect ratios of the basins and up-lifts increase with increasing displacement on thestrike-slip faults The results of thesemodels can beused to interpret the structural and fault geom-etries in surface and subsurface structures formedalong strike-slip faults

INTRODUCTION

Strike-slip faults are commonly characterized bybends or offsets marked by faults or fault systemswith an oblique or transverse orientation to theirgeneral trend (Figure 1) A basement strike-slipfault may change position along a bend (Figure 1A)or by an offset of two separate fault segments(Figure 1B) The bends may be oblique or trans-verse to the main fault Depending on their ori-entation relative to the sense of strike slip bends oroffsets in strike-slip faults are referred to as releas-ing or restraining (Figure 1C D) Releasing bendsor offsets form where the sense of shear on a strike-slip fault results in a component of fault-parallelextension (Figure 1C) whereas restraining bendsor offsets form where the shear results in a com-ponent of fault-parallel contraction (Figure 1D)

The types of secondary structures and theirorientations relative to the main fault were firstdescribed by Crowell (1974) and used to explainpull-apart basins and oblique compressive struc-tures mapped by various authors in Californiabasins Subsequently these models have been ap-plied to explain structures in a variety of basinssuch as in the collection of articles compiled bySylvester (1984) Biddle andChristie-Blick (1985)and Cunningham and Mann (2007)

Most studies of restraining and releasing bendsare based on surface seismic or remote-sensed dataBecause data from all of these sources are incom-plete the interpretations are commonlymade basedon simplified conceptual models of these struc-

1148 Structural Geometry and Evolution of Releasing and Rest

tures Knowledge of the geometry of folds andfaults formed in these settings and their progressiveevolution through time would be useful in inter-preting these structures in areas of incomplete dataOne source of information for the detailed in-formation is scaled experimental models of thesestructures

The only published examples of restrainingbends are the sandbox models of McClay andBonora (2001) that investigated the shapes andevolution of structures and faults along preexist-ing jogs modeled above sliding base plates Theseauthors recognized the importance of document-ing the structural relief associated with strike-slipstructures and attempted to do so by buildingpseudondashthree-dimensional (3-D) models fromspaced serial sections Several experimental mod-els of releasing bends have been published Theseinclude the sandboxmodels ofDooley andMcClay(1987) Rahe et al (1998) Basile and Brun (1999)Wu et al (2009) and the clay models of Atmaouiet al (2006) The first four articles all used precutjogs in sliding plates to simulate strike slip whereasAtmaoui et al (2006) used both sliding plates anddistributed shear in their experiments DooleyandMcClay (1997) andWuet al (2009) attemptedto qualitatively explain the variation in structuralrelief resulting from the oblique deformation

The preexisting experimental models all usedsliding metal plates with bends along them to sim-ulate strike-slip deformation These experimentalconfigurations do not model deformation associatedwith preexisting offsets in the basement strike-slipfault which is expected to be quite different fromthat associated with bends Furthermore with theexception of Atmaoui et al (2006) who investi-gated the role of distributed shear in the develop-ment and intersection of Riedel shears they allmodeled the basement as perfectly rigid with thetransition from strike-slip to oblique deformationoccurring along a single preexisting jog

In this article we present experimental claymodels of both restraining and releasing bendswith various geometries of fault bends and faultoffsets The fundamental difference between theexperimental configuration used in these experi-ments with previous models is that the basement

raining Bends

is modeled with stiff clay which is significantlymore rigid than the overlying soft clay represent-ing the sedimentary cover but which allows for asmall component of deformation of the basementespecially in the immediate vicinity of the majorfaults Sliding metal plates which underlie the base-

ment are used only to control the nature of de-formation and the position of the bend or offsetPreexisting vertical cuts in the stiff clay define theinitial geometry of the basement faults This con-figuration also enables us to study the differencebetween releasing offsets and bends which result

Figure 1 (A B) Possible basement faultconfigurations for strike-slip faults (A) Bend(B) Offset (C) Releasing bend (D) Re-straining bend (E) Initial basement faultconfigurations used in experiments forreleasing bends and offsets (F) Initial base-ment fault configurations used in experi-ments for restraining bends Figures 2 and13 show details of the experimental setupfor (E) and (F)

Mitra and Paul 1149

in very different geometries of structures and faultpatterns in pull-apart basins (Figure 1) The specificexperimental configurations depicted in this studyare summarized in panels E and F of Figure 1 anddescribed in more detail in the following sections

The geometry and evolution of secondary faultsfor the top surface are studied in detail In addi-tion using a new technique of laser scanning westudy the mutual relationships between structuralrelief and faulting with progressive deformation incontoured models for the surface The renditionsof the quantitative geometry of the deformed sur-faces are directly comparable to structure contourmaps of natural structures and therefore of mostinterest to petroleum geologists

1150 Structural Geometry and Evolution of Releasing and Rest

The results of the models are compared withnatural examples of releasing and restraining bendsto show how detailed interpretations of the struc-tures can be improved through an understandingof the evolution of the structures

RELEASING BENDS AND OFFSETS

Experimental Approach

Clay experiments are conducted to simulate theformation of releasing bends and offsets in strike-slip faults The main advantage of using clay asthe model material is that it allows a much better

Figure 2 Experimental models conductedfor releasing bends and offsets (A B)Oblique 45deg bend in basement fault(C D) Transverse 90deg bend in basementfault (E F) Transverse 90deg offset inbasement fault For each experiment thefirst figure shows the relationship betweenthe underlying sliding plates and thesecond figure shows the basement andsedimentary cover and the preexisting faultcuts in the basement

raining Bends

definition of fault geometries and evolution and alsoenables the deformation to be studied for ma-ture structures involving steep surfaces The ex-perimental setup consists of two overlapping metalplates one of which is attached to a moving platedriven by motors whereas the other is attached toa fixed plate The lateral motion between the twobase plates drives the strike-slip motion in the over-lying stiff and soft clay representing the basementand sedimentary layers (Figure 2) No shearing oc-curs along the contact between the stiff clay andthe sliding plates Basement is represented by stiffclay with a density of 185 gcc and a thickness of1 in (25 cm) The overlying sedimentary layer isrepresented by soft water-based clay with a densityof 16 to 165 gcc and a thickness of 08 in (2 cm)The density of the clay is adjusted by altering thewater content of the clay so that the clay has theappropriate consistency to deform by faulting with-out developing large cracks No layering is usedwithin the soft clay The displacement rate for allexperiments is 0001 cms corresponding to a strainrate of approximately 3 times 10ndash5 per second

The experimental configurations used in thisstudy differ from previous setups (Dooley andMcClay 1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006Wu et al 2009) in thatstrike-slip motion is induced in both the stiff clay(basement) and the overlying soft clay (sedimentarycover) through sliding base plates at depth There-fore the sliding plates only control the nature ofdeformation and the location of the bend or offsetThis approach enables us to study differences indeformation involving releasing bends and offsetsin the strike-slip faults within the basement whereasprevious experiments modeled only releasing bendsFurthermore using stiff clay to model basementinstead of rigid plates the basement deforms in theimmediate vicinity of the faults to accommodatethe deformation in these areas In previousmodelsthe top of the basement was represented by slidingplates rendering the basement perfectly rigid inthe vicinity of the faults

Two experimental setups for the master faultsin the basement are investigated (1) bends inwhichtwo parallel strike-slip faults are connected by anoblique (Figure 2A B) or transverse (Figure 2C D)

bend in the stiff clay representing the basementand (2) offsets in which two separate strike-slipfault segments in the basement are separated by a90deg offset in the stiff clay (Figure 2E F) The initialconfiguration for the bend setup consists of twovertical cuts in the stiff clay connected by obliqueor transverse bends whereas the configuration forthe offset setup consists of two parallel and offsetvertical cuts In both cases the basement is under-lain by two sliding plates that override each other toprovide the strike-slip motion (Serra and Nelson1988) The bend configuration may represent anoriginal geometry of the strike-slip fault (Crowell1974) or may result from two separate parallel seg-ments that have connected along a transverse oroblique segment The offset configuration morespecifically addresses the linking of two adjacentstrike-slip segments by the development of a trans-fer zone consisting of single or multiple obliquefaults These configurations result in significantdifferences in the geometry and faulting associatedwith the pull-apart basins Two runs of each ex-periment are conducted to test for repeatability

The progressive evolution of structures and sec-ondary faults and their orientations within a strike-slip system is mapped directly from photographstaken during various stages of the experiments Inaddition we use a new technique involving theanalysis of the 3-D geometry of the top surface inthe structures using laser scanning (see also Boseand Mitra 2010) The technology of scanning sur-faces by a laser scanner and developing a virtual3-D model has been in use in numerous industriesparticularly in 3-D animation The same technol-ogy is applied here to gather 3-D information andthereby generate a virtual surface that can be used tovisualize the development of the structure in muchgreater detail and to generate contour maps of sur-faces The scanning size is 135 times 101 in (343 times257 cm) with a resolution of 75 dots per inch(sim0015 in point density) The scanner is placedface down approximately 15 in (38 cm) verticallyabove the clay cake and top surfaces are scanned atapproximately equal increments of deformation

The sensor captures the laser beam that is pro-jectedon the surface and thecoordinatesof thepointsare computed by the scanner using a triangulation

Mitra and Paul 1151

method (Figure 3A) The known parameters arethe angle a at which the laser points out its beamon the surface the angle b at which the sensor col-lects the laser beam (b is known if the focal lengthand the pixel size of the sensor are given) and thetriangulation distance between the sensor and lasersource (Petrov et al 1998) Because all the geo-metric parameters are known the coordinates x yand z on the surface are calculated using trigono-metric methods The laser source produces a lineand all data along that line are collected at the sametime (Figure 3B) Moreover twin arrays of four laserbeams are used as source in this instrument for

1152 Structural Geometry and Evolution of Releasing and Rest

cross-validating each data point by measuring it atleast twice The origin or reference coordinate sys-tem is determined automatically during each run bythe first point scanned on the surface It is thereforenot necessary to move the scanner during the lengthof the experiment to spatially position the scannedsurfaces in the same reference coordinate system

The scanned data are exported as point files intothe 3-D modeling software GoCAD (Bose andMitra 2010) The point clouds obtained (Figure 3C)are then used to build a surface Fault polygons arebuilt from visible breaks in the modeled surface(Figure 3D) and the geometries of the polygons

Figure 3 Laser scanning of experimental clay models (A) Schematic diagram of the working principle of the laser scanner withthe triangulation formed by the laser source the detector and the point on the clay surface from which the laser beam is reflected(B) Schematic view of the laser scanner projecting a line and sweeping across the clay surface at a constant velocity (from Bose andMitra 2010 used with permission from AAPG) (C) Cloud of points obtained from the scanned surface and visualized in GO-CAD(D) Modeled surface in GO-CAD cut by faults (E) Oblique photograph of the top of the clay surface

raining Bends

are checked against the traced fault cuts on photo-graphs taken of the clay surface which provide de-tails of the fault geometry (Figure 3E) Themodeledsurface is cut by the fault polygons and contouredto depict their topology (Figure 3D) The contouredfaulted surface provides an accurate quantitativerendition of the structural geometry which can becompared directly with structure contour maps onnatural structures

Experimental Results

Oblique (45deg) BendInitial movement on the base plate results in a broaddepression oblique to the preexisting basementstrike-slip faults in the area above the obliquebend (Figures 4A 5A) This depression is boundedby two sets of faults dipping toward each otherand trending approximately 42deg to the basement

Figure 4 Clay models showing faults in the sedimentary cover for a releasing bend with an oblique (45deg) bend connecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Insets in this and other figures for experimentalmodels show initial fault cuts in the stiff clay Panels (AndashF) show the progressive evolution of the structure with increasing strike-slipdisplacement Rose diagrams show the lengths of faults for different trends in 5deg increments as a percentage of the total length of faults

Mitra and Paul 1153

Figure 5 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with an oblique (45deg) bend connectingtwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement Contours in this and subsequent diagrams are in intervals of 04 mm Every fifth contour is shown in black Color gridsreflect structural elevation with red and yellow representing high areas and blue and green representing low areas Note the formation oftwo diametrically opposite lows in the pull-apart basin (G H) Cross sections through top of clay showing final basin geometry and shift inpolarity for cross sections AAprime and BBprime Locations of the cross sections are shown in (F)

1154 Structural Geometry and Evolution of Releasing and Restraining Bends

strike-slip faults that define the boundaries of thepull-apart basin The angular measurements citedpreviously and in all experiments on pull-apart ba-sins are made in a counterclockwise direction fromthe strike-slip faults The individual faults are madeup of two or more right-stepping en echelon seg-ments some of which eventually interact to formlonger faults With progressive deformation thesefaults systems are rotated slightly by shear

Immediately above the main vertical strike-slipfaults short segments of steep faults of two mainsets trending approximately 15 and 80deg to the base-ment strike-slip faults develop (Figure 4B C) Thesefaults are interpreted to be the R and Rprime Riedelshears associated with strike-slip faulting describedand experimentally modeled by Tchalenko (1970)Tchalenko and Ambraseys (1970) Wilcox et al(1973) and Harding (1985)

Figure 6 Clay models showing faults in the sedimentary cover for a releasing bend with a transverse (90deg) bend connecting two strike-slip fault segments in the basement Sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution of thestructure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

Mitra and Paul 1155

Figure 7 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with a transverse (90deg) bend con-necting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slip displacement

1156 Structural Geometry and Evolution of Releasing and Restraining Bends

The normal faults in the central part of thebasin curve into the orientations of the R shearsin the vicinity of the basement strike-slip faults(Figures 4B C 5B C) The resulting fault systemshave a sigmoidal geometry Rose diagrams of faultorientations show the progressive developmentof scatter because of the development of curvedfaults with varying components of normal andstrike-slip motion (Figure 4AndashC) Twomajor strike-slip faults develop at the surface directly above thepreexisting cuts in the basement after about 08 cmof displacement

Continuing deformation results in additional setsof major normal faults inward and approximatelyparallel to the original pair of fault systems Addi-tional fault systems with the same overall sigmoidalgeometry form in other parts of the transfer zoneand outside the central depression (Figures 4BndashE5BndashE) In the final stages of deformation the lon-gest pair of oblique faults connects with the twomain systems of strike-slip faults thus defining theboundary of the pull-apart basin (Figures 4F 5F)The final rose diagram shows significant scatter oftrends ranging from north to west because of thevariation from strike-slip to normal fault trends witha maximum density in an approximately north-west orientation (Figure 4F)

Examination of the contour patterns (Figure 5)suggests that the structural elevation steps downquite steeply across the normal faults bounding thepull-apart basin but drops more gradually from thebounding system of strike-slip faults to the centerof the basin The central part consists of two sep-arate depressions separated by a local high thatdevelop during the late stages (Figure 5E F) Theaxis connecting the centers of the two depressionsis approximately parallel to the trend of the mainstrike-slip faults Cross sections through the twodepressions show a flip in polarity because of achange in fault displacement along trend withinthe basin (Figure 5G H)

Transverse (90deg) BendFor a transverse bend (90deg) the normal fault seg-ments in the transfer zone form at a higher angle(58ndash78deg) than in the case of the oblique jog (45deg)and show left- and right-stepping patterns on op-

posite sides of the transfer zone (Figures 6A B 7AB) Immediately above the basement strike-slipfaults R and Rprime Riedel shears develop as a series ofen echelon faults making angles of 15 and 82deg to themain faults respectively The normal faults in thecenter of the pull-apart basin curve sharply clock-wise toward the orientation of the R shears in thevicinity of the main strike-slip fault zones along theboundaries of the pull-apart basin (Figures 6C D7C D)With continuing shear the major strike-slipfaults break through to the surface as Y shears de-fining the boundaries of pull-apart basin (Figures 6EF 7E F) As in the case of the oblique-bend ex-periment multiple sets of faults form in the pull-apart basin especially close to the main strike-slipfault zone Theorientations of the faults showawidevariation in trends ranging from northndashnorthwest towestndashnorthwest with the most dominant trend inthe center of the basin being about northwest in thefinal stage (Figure 6F)

Because of the sharper bends and changes inthe trends of the faults in the transfer zone the pull-apart basin associated with a transverse bend has amore S-shaped or rhomboidal geometry than thespindle-shaped geometry in the case of an obliquebend The contours show that the deepest part ofthe basin is located toward one of the major transferfaults so that the pull-apart basin has a more asym-metric shape with a sharper drop of contours onone side (Figure 7E F) The sides of the basin aremarked by a more gradual change in elevation withthe structure contours cut by secondary faults Dif-ferent basin shapes may result depending upon theslip distribution among the different faults

Transverse (90deg) OffsetThis setup has two cuts in the basement that even-tually form the bounding strike-slip faults offsetalong a transfer zone The main purpose of con-ducting this experiment was to understand the dif-ference in the geometry of the pull-apart basinwhenno connecting bend exists in the basement faultThe faults in the transfer zone develop with on-going strike-slip deformation independent of a con-nected basement fault (Figures 8 9) Our experi-ments show that a transverse offset results in theoptimum configuration for the development of a

Mitra and Paul 1157

transfer zone of connecting faults Experiments in-volving overlapping or approaching patterns do notresult in well-defined pull-apart basins because theyboth involve a large transfer zone area This suggeststhat for two approaching strike-slip fault segmentsconnecting faults in the transfer zone will mostlikely start to develop when the end points of thetwo strike-slip segments are directly opposite eachother in a transverse offset position

1158 Structural Geometry and Evolution of Releasing and Rest

Initial deformation results in the formation ofa large number of short en echelon normal faults inthe area of the transfer zone (Figures 8A B 9A B)The central parts of the faults trend at angles of 53to 63deg to themain strike-slip fault zone At the sametime a series of en echelon R and Rprime Riedel shearsform above the main basement faults at angles of18 and 70deg with respect to the basement strike-slip faults respectivelyWith increasing strike-slip

Figure 8 Clay models showing faults in the sedimentary cover for a releasing offset with a transverse (90deg) offset between two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution ofthe structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

raining Bends

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

is modeled with stiff clay which is significantlymore rigid than the overlying soft clay represent-ing the sedimentary cover but which allows for asmall component of deformation of the basementespecially in the immediate vicinity of the majorfaults Sliding metal plates which underlie the base-

ment are used only to control the nature of de-formation and the position of the bend or offsetPreexisting vertical cuts in the stiff clay define theinitial geometry of the basement faults This con-figuration also enables us to study the differencebetween releasing offsets and bends which result

Figure 1 (A B) Possible basement faultconfigurations for strike-slip faults (A) Bend(B) Offset (C) Releasing bend (D) Re-straining bend (E) Initial basement faultconfigurations used in experiments forreleasing bends and offsets (F) Initial base-ment fault configurations used in experi-ments for restraining bends Figures 2 and13 show details of the experimental setupfor (E) and (F)

Mitra and Paul 1149

in very different geometries of structures and faultpatterns in pull-apart basins (Figure 1) The specificexperimental configurations depicted in this studyare summarized in panels E and F of Figure 1 anddescribed in more detail in the following sections

The geometry and evolution of secondary faultsfor the top surface are studied in detail In addi-tion using a new technique of laser scanning westudy the mutual relationships between structuralrelief and faulting with progressive deformation incontoured models for the surface The renditionsof the quantitative geometry of the deformed sur-faces are directly comparable to structure contourmaps of natural structures and therefore of mostinterest to petroleum geologists

1150 Structural Geometry and Evolution of Releasing and Rest

The results of the models are compared withnatural examples of releasing and restraining bendsto show how detailed interpretations of the struc-tures can be improved through an understandingof the evolution of the structures

RELEASING BENDS AND OFFSETS

Experimental Approach

Clay experiments are conducted to simulate theformation of releasing bends and offsets in strike-slip faults The main advantage of using clay asthe model material is that it allows a much better

Figure 2 Experimental models conductedfor releasing bends and offsets (A B)Oblique 45deg bend in basement fault(C D) Transverse 90deg bend in basementfault (E F) Transverse 90deg offset inbasement fault For each experiment thefirst figure shows the relationship betweenthe underlying sliding plates and thesecond figure shows the basement andsedimentary cover and the preexisting faultcuts in the basement

raining Bends

definition of fault geometries and evolution and alsoenables the deformation to be studied for ma-ture structures involving steep surfaces The ex-perimental setup consists of two overlapping metalplates one of which is attached to a moving platedriven by motors whereas the other is attached toa fixed plate The lateral motion between the twobase plates drives the strike-slip motion in the over-lying stiff and soft clay representing the basementand sedimentary layers (Figure 2) No shearing oc-curs along the contact between the stiff clay andthe sliding plates Basement is represented by stiffclay with a density of 185 gcc and a thickness of1 in (25 cm) The overlying sedimentary layer isrepresented by soft water-based clay with a densityof 16 to 165 gcc and a thickness of 08 in (2 cm)The density of the clay is adjusted by altering thewater content of the clay so that the clay has theappropriate consistency to deform by faulting with-out developing large cracks No layering is usedwithin the soft clay The displacement rate for allexperiments is 0001 cms corresponding to a strainrate of approximately 3 times 10ndash5 per second

The experimental configurations used in thisstudy differ from previous setups (Dooley andMcClay 1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006Wu et al 2009) in thatstrike-slip motion is induced in both the stiff clay(basement) and the overlying soft clay (sedimentarycover) through sliding base plates at depth There-fore the sliding plates only control the nature ofdeformation and the location of the bend or offsetThis approach enables us to study differences indeformation involving releasing bends and offsetsin the strike-slip faults within the basement whereasprevious experiments modeled only releasing bendsFurthermore using stiff clay to model basementinstead of rigid plates the basement deforms in theimmediate vicinity of the faults to accommodatethe deformation in these areas In previousmodelsthe top of the basement was represented by slidingplates rendering the basement perfectly rigid inthe vicinity of the faults

Two experimental setups for the master faultsin the basement are investigated (1) bends inwhichtwo parallel strike-slip faults are connected by anoblique (Figure 2A B) or transverse (Figure 2C D)

bend in the stiff clay representing the basementand (2) offsets in which two separate strike-slipfault segments in the basement are separated by a90deg offset in the stiff clay (Figure 2E F) The initialconfiguration for the bend setup consists of twovertical cuts in the stiff clay connected by obliqueor transverse bends whereas the configuration forthe offset setup consists of two parallel and offsetvertical cuts In both cases the basement is under-lain by two sliding plates that override each other toprovide the strike-slip motion (Serra and Nelson1988) The bend configuration may represent anoriginal geometry of the strike-slip fault (Crowell1974) or may result from two separate parallel seg-ments that have connected along a transverse oroblique segment The offset configuration morespecifically addresses the linking of two adjacentstrike-slip segments by the development of a trans-fer zone consisting of single or multiple obliquefaults These configurations result in significantdifferences in the geometry and faulting associatedwith the pull-apart basins Two runs of each ex-periment are conducted to test for repeatability

The progressive evolution of structures and sec-ondary faults and their orientations within a strike-slip system is mapped directly from photographstaken during various stages of the experiments Inaddition we use a new technique involving theanalysis of the 3-D geometry of the top surface inthe structures using laser scanning (see also Boseand Mitra 2010) The technology of scanning sur-faces by a laser scanner and developing a virtual3-D model has been in use in numerous industriesparticularly in 3-D animation The same technol-ogy is applied here to gather 3-D information andthereby generate a virtual surface that can be used tovisualize the development of the structure in muchgreater detail and to generate contour maps of sur-faces The scanning size is 135 times 101 in (343 times257 cm) with a resolution of 75 dots per inch(sim0015 in point density) The scanner is placedface down approximately 15 in (38 cm) verticallyabove the clay cake and top surfaces are scanned atapproximately equal increments of deformation

The sensor captures the laser beam that is pro-jectedon the surface and thecoordinatesof thepointsare computed by the scanner using a triangulation

Mitra and Paul 1151

method (Figure 3A) The known parameters arethe angle a at which the laser points out its beamon the surface the angle b at which the sensor col-lects the laser beam (b is known if the focal lengthand the pixel size of the sensor are given) and thetriangulation distance between the sensor and lasersource (Petrov et al 1998) Because all the geo-metric parameters are known the coordinates x yand z on the surface are calculated using trigono-metric methods The laser source produces a lineand all data along that line are collected at the sametime (Figure 3B) Moreover twin arrays of four laserbeams are used as source in this instrument for

1152 Structural Geometry and Evolution of Releasing and Rest

cross-validating each data point by measuring it atleast twice The origin or reference coordinate sys-tem is determined automatically during each run bythe first point scanned on the surface It is thereforenot necessary to move the scanner during the lengthof the experiment to spatially position the scannedsurfaces in the same reference coordinate system

The scanned data are exported as point files intothe 3-D modeling software GoCAD (Bose andMitra 2010) The point clouds obtained (Figure 3C)are then used to build a surface Fault polygons arebuilt from visible breaks in the modeled surface(Figure 3D) and the geometries of the polygons

Figure 3 Laser scanning of experimental clay models (A) Schematic diagram of the working principle of the laser scanner withthe triangulation formed by the laser source the detector and the point on the clay surface from which the laser beam is reflected(B) Schematic view of the laser scanner projecting a line and sweeping across the clay surface at a constant velocity (from Bose andMitra 2010 used with permission from AAPG) (C) Cloud of points obtained from the scanned surface and visualized in GO-CAD(D) Modeled surface in GO-CAD cut by faults (E) Oblique photograph of the top of the clay surface

raining Bends

are checked against the traced fault cuts on photo-graphs taken of the clay surface which provide de-tails of the fault geometry (Figure 3E) Themodeledsurface is cut by the fault polygons and contouredto depict their topology (Figure 3D) The contouredfaulted surface provides an accurate quantitativerendition of the structural geometry which can becompared directly with structure contour maps onnatural structures

Experimental Results

Oblique (45deg) BendInitial movement on the base plate results in a broaddepression oblique to the preexisting basementstrike-slip faults in the area above the obliquebend (Figures 4A 5A) This depression is boundedby two sets of faults dipping toward each otherand trending approximately 42deg to the basement

Figure 4 Clay models showing faults in the sedimentary cover for a releasing bend with an oblique (45deg) bend connecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Insets in this and other figures for experimentalmodels show initial fault cuts in the stiff clay Panels (AndashF) show the progressive evolution of the structure with increasing strike-slipdisplacement Rose diagrams show the lengths of faults for different trends in 5deg increments as a percentage of the total length of faults

Mitra and Paul 1153

Figure 5 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with an oblique (45deg) bend connectingtwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement Contours in this and subsequent diagrams are in intervals of 04 mm Every fifth contour is shown in black Color gridsreflect structural elevation with red and yellow representing high areas and blue and green representing low areas Note the formation oftwo diametrically opposite lows in the pull-apart basin (G H) Cross sections through top of clay showing final basin geometry and shift inpolarity for cross sections AAprime and BBprime Locations of the cross sections are shown in (F)

1154 Structural Geometry and Evolution of Releasing and Restraining Bends

strike-slip faults that define the boundaries of thepull-apart basin The angular measurements citedpreviously and in all experiments on pull-apart ba-sins are made in a counterclockwise direction fromthe strike-slip faults The individual faults are madeup of two or more right-stepping en echelon seg-ments some of which eventually interact to formlonger faults With progressive deformation thesefaults systems are rotated slightly by shear

Immediately above the main vertical strike-slipfaults short segments of steep faults of two mainsets trending approximately 15 and 80deg to the base-ment strike-slip faults develop (Figure 4B C) Thesefaults are interpreted to be the R and Rprime Riedelshears associated with strike-slip faulting describedand experimentally modeled by Tchalenko (1970)Tchalenko and Ambraseys (1970) Wilcox et al(1973) and Harding (1985)

Figure 6 Clay models showing faults in the sedimentary cover for a releasing bend with a transverse (90deg) bend connecting two strike-slip fault segments in the basement Sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution of thestructure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

Mitra and Paul 1155

Figure 7 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with a transverse (90deg) bend con-necting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slip displacement

1156 Structural Geometry and Evolution of Releasing and Restraining Bends

The normal faults in the central part of thebasin curve into the orientations of the R shearsin the vicinity of the basement strike-slip faults(Figures 4B C 5B C) The resulting fault systemshave a sigmoidal geometry Rose diagrams of faultorientations show the progressive developmentof scatter because of the development of curvedfaults with varying components of normal andstrike-slip motion (Figure 4AndashC) Twomajor strike-slip faults develop at the surface directly above thepreexisting cuts in the basement after about 08 cmof displacement

Continuing deformation results in additional setsof major normal faults inward and approximatelyparallel to the original pair of fault systems Addi-tional fault systems with the same overall sigmoidalgeometry form in other parts of the transfer zoneand outside the central depression (Figures 4BndashE5BndashE) In the final stages of deformation the lon-gest pair of oblique faults connects with the twomain systems of strike-slip faults thus defining theboundary of the pull-apart basin (Figures 4F 5F)The final rose diagram shows significant scatter oftrends ranging from north to west because of thevariation from strike-slip to normal fault trends witha maximum density in an approximately north-west orientation (Figure 4F)

Examination of the contour patterns (Figure 5)suggests that the structural elevation steps downquite steeply across the normal faults bounding thepull-apart basin but drops more gradually from thebounding system of strike-slip faults to the centerof the basin The central part consists of two sep-arate depressions separated by a local high thatdevelop during the late stages (Figure 5E F) Theaxis connecting the centers of the two depressionsis approximately parallel to the trend of the mainstrike-slip faults Cross sections through the twodepressions show a flip in polarity because of achange in fault displacement along trend withinthe basin (Figure 5G H)

Transverse (90deg) BendFor a transverse bend (90deg) the normal fault seg-ments in the transfer zone form at a higher angle(58ndash78deg) than in the case of the oblique jog (45deg)and show left- and right-stepping patterns on op-

posite sides of the transfer zone (Figures 6A B 7AB) Immediately above the basement strike-slipfaults R and Rprime Riedel shears develop as a series ofen echelon faults making angles of 15 and 82deg to themain faults respectively The normal faults in thecenter of the pull-apart basin curve sharply clock-wise toward the orientation of the R shears in thevicinity of the main strike-slip fault zones along theboundaries of the pull-apart basin (Figures 6C D7C D)With continuing shear the major strike-slipfaults break through to the surface as Y shears de-fining the boundaries of pull-apart basin (Figures 6EF 7E F) As in the case of the oblique-bend ex-periment multiple sets of faults form in the pull-apart basin especially close to the main strike-slipfault zone Theorientations of the faults showawidevariation in trends ranging from northndashnorthwest towestndashnorthwest with the most dominant trend inthe center of the basin being about northwest in thefinal stage (Figure 6F)

Because of the sharper bends and changes inthe trends of the faults in the transfer zone the pull-apart basin associated with a transverse bend has amore S-shaped or rhomboidal geometry than thespindle-shaped geometry in the case of an obliquebend The contours show that the deepest part ofthe basin is located toward one of the major transferfaults so that the pull-apart basin has a more asym-metric shape with a sharper drop of contours onone side (Figure 7E F) The sides of the basin aremarked by a more gradual change in elevation withthe structure contours cut by secondary faults Dif-ferent basin shapes may result depending upon theslip distribution among the different faults

Transverse (90deg) OffsetThis setup has two cuts in the basement that even-tually form the bounding strike-slip faults offsetalong a transfer zone The main purpose of con-ducting this experiment was to understand the dif-ference in the geometry of the pull-apart basinwhenno connecting bend exists in the basement faultThe faults in the transfer zone develop with on-going strike-slip deformation independent of a con-nected basement fault (Figures 8 9) Our experi-ments show that a transverse offset results in theoptimum configuration for the development of a

Mitra and Paul 1157

transfer zone of connecting faults Experiments in-volving overlapping or approaching patterns do notresult in well-defined pull-apart basins because theyboth involve a large transfer zone area This suggeststhat for two approaching strike-slip fault segmentsconnecting faults in the transfer zone will mostlikely start to develop when the end points of thetwo strike-slip segments are directly opposite eachother in a transverse offset position

1158 Structural Geometry and Evolution of Releasing and Rest

Initial deformation results in the formation ofa large number of short en echelon normal faults inthe area of the transfer zone (Figures 8A B 9A B)The central parts of the faults trend at angles of 53to 63deg to themain strike-slip fault zone At the sametime a series of en echelon R and Rprime Riedel shearsform above the main basement faults at angles of18 and 70deg with respect to the basement strike-slip faults respectivelyWith increasing strike-slip

Figure 8 Clay models showing faults in the sedimentary cover for a releasing offset with a transverse (90deg) offset between two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution ofthe structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

raining Bends

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

in very different geometries of structures and faultpatterns in pull-apart basins (Figure 1) The specificexperimental configurations depicted in this studyare summarized in panels E and F of Figure 1 anddescribed in more detail in the following sections

The geometry and evolution of secondary faultsfor the top surface are studied in detail In addi-tion using a new technique of laser scanning westudy the mutual relationships between structuralrelief and faulting with progressive deformation incontoured models for the surface The renditionsof the quantitative geometry of the deformed sur-faces are directly comparable to structure contourmaps of natural structures and therefore of mostinterest to petroleum geologists

1150 Structural Geometry and Evolution of Releasing and Rest

The results of the models are compared withnatural examples of releasing and restraining bendsto show how detailed interpretations of the struc-tures can be improved through an understandingof the evolution of the structures

RELEASING BENDS AND OFFSETS

Experimental Approach

Clay experiments are conducted to simulate theformation of releasing bends and offsets in strike-slip faults The main advantage of using clay asthe model material is that it allows a much better

Figure 2 Experimental models conductedfor releasing bends and offsets (A B)Oblique 45deg bend in basement fault(C D) Transverse 90deg bend in basementfault (E F) Transverse 90deg offset inbasement fault For each experiment thefirst figure shows the relationship betweenthe underlying sliding plates and thesecond figure shows the basement andsedimentary cover and the preexisting faultcuts in the basement

raining Bends

definition of fault geometries and evolution and alsoenables the deformation to be studied for ma-ture structures involving steep surfaces The ex-perimental setup consists of two overlapping metalplates one of which is attached to a moving platedriven by motors whereas the other is attached toa fixed plate The lateral motion between the twobase plates drives the strike-slip motion in the over-lying stiff and soft clay representing the basementand sedimentary layers (Figure 2) No shearing oc-curs along the contact between the stiff clay andthe sliding plates Basement is represented by stiffclay with a density of 185 gcc and a thickness of1 in (25 cm) The overlying sedimentary layer isrepresented by soft water-based clay with a densityof 16 to 165 gcc and a thickness of 08 in (2 cm)The density of the clay is adjusted by altering thewater content of the clay so that the clay has theappropriate consistency to deform by faulting with-out developing large cracks No layering is usedwithin the soft clay The displacement rate for allexperiments is 0001 cms corresponding to a strainrate of approximately 3 times 10ndash5 per second

The experimental configurations used in thisstudy differ from previous setups (Dooley andMcClay 1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006Wu et al 2009) in thatstrike-slip motion is induced in both the stiff clay(basement) and the overlying soft clay (sedimentarycover) through sliding base plates at depth There-fore the sliding plates only control the nature ofdeformation and the location of the bend or offsetThis approach enables us to study differences indeformation involving releasing bends and offsetsin the strike-slip faults within the basement whereasprevious experiments modeled only releasing bendsFurthermore using stiff clay to model basementinstead of rigid plates the basement deforms in theimmediate vicinity of the faults to accommodatethe deformation in these areas In previousmodelsthe top of the basement was represented by slidingplates rendering the basement perfectly rigid inthe vicinity of the faults

Two experimental setups for the master faultsin the basement are investigated (1) bends inwhichtwo parallel strike-slip faults are connected by anoblique (Figure 2A B) or transverse (Figure 2C D)

bend in the stiff clay representing the basementand (2) offsets in which two separate strike-slipfault segments in the basement are separated by a90deg offset in the stiff clay (Figure 2E F) The initialconfiguration for the bend setup consists of twovertical cuts in the stiff clay connected by obliqueor transverse bends whereas the configuration forthe offset setup consists of two parallel and offsetvertical cuts In both cases the basement is under-lain by two sliding plates that override each other toprovide the strike-slip motion (Serra and Nelson1988) The bend configuration may represent anoriginal geometry of the strike-slip fault (Crowell1974) or may result from two separate parallel seg-ments that have connected along a transverse oroblique segment The offset configuration morespecifically addresses the linking of two adjacentstrike-slip segments by the development of a trans-fer zone consisting of single or multiple obliquefaults These configurations result in significantdifferences in the geometry and faulting associatedwith the pull-apart basins Two runs of each ex-periment are conducted to test for repeatability

The progressive evolution of structures and sec-ondary faults and their orientations within a strike-slip system is mapped directly from photographstaken during various stages of the experiments Inaddition we use a new technique involving theanalysis of the 3-D geometry of the top surface inthe structures using laser scanning (see also Boseand Mitra 2010) The technology of scanning sur-faces by a laser scanner and developing a virtual3-D model has been in use in numerous industriesparticularly in 3-D animation The same technol-ogy is applied here to gather 3-D information andthereby generate a virtual surface that can be used tovisualize the development of the structure in muchgreater detail and to generate contour maps of sur-faces The scanning size is 135 times 101 in (343 times257 cm) with a resolution of 75 dots per inch(sim0015 in point density) The scanner is placedface down approximately 15 in (38 cm) verticallyabove the clay cake and top surfaces are scanned atapproximately equal increments of deformation

The sensor captures the laser beam that is pro-jectedon the surface and thecoordinatesof thepointsare computed by the scanner using a triangulation

Mitra and Paul 1151

method (Figure 3A) The known parameters arethe angle a at which the laser points out its beamon the surface the angle b at which the sensor col-lects the laser beam (b is known if the focal lengthand the pixel size of the sensor are given) and thetriangulation distance between the sensor and lasersource (Petrov et al 1998) Because all the geo-metric parameters are known the coordinates x yand z on the surface are calculated using trigono-metric methods The laser source produces a lineand all data along that line are collected at the sametime (Figure 3B) Moreover twin arrays of four laserbeams are used as source in this instrument for

1152 Structural Geometry and Evolution of Releasing and Rest

cross-validating each data point by measuring it atleast twice The origin or reference coordinate sys-tem is determined automatically during each run bythe first point scanned on the surface It is thereforenot necessary to move the scanner during the lengthof the experiment to spatially position the scannedsurfaces in the same reference coordinate system

The scanned data are exported as point files intothe 3-D modeling software GoCAD (Bose andMitra 2010) The point clouds obtained (Figure 3C)are then used to build a surface Fault polygons arebuilt from visible breaks in the modeled surface(Figure 3D) and the geometries of the polygons

Figure 3 Laser scanning of experimental clay models (A) Schematic diagram of the working principle of the laser scanner withthe triangulation formed by the laser source the detector and the point on the clay surface from which the laser beam is reflected(B) Schematic view of the laser scanner projecting a line and sweeping across the clay surface at a constant velocity (from Bose andMitra 2010 used with permission from AAPG) (C) Cloud of points obtained from the scanned surface and visualized in GO-CAD(D) Modeled surface in GO-CAD cut by faults (E) Oblique photograph of the top of the clay surface

raining Bends

are checked against the traced fault cuts on photo-graphs taken of the clay surface which provide de-tails of the fault geometry (Figure 3E) Themodeledsurface is cut by the fault polygons and contouredto depict their topology (Figure 3D) The contouredfaulted surface provides an accurate quantitativerendition of the structural geometry which can becompared directly with structure contour maps onnatural structures

Experimental Results

Oblique (45deg) BendInitial movement on the base plate results in a broaddepression oblique to the preexisting basementstrike-slip faults in the area above the obliquebend (Figures 4A 5A) This depression is boundedby two sets of faults dipping toward each otherand trending approximately 42deg to the basement

Figure 4 Clay models showing faults in the sedimentary cover for a releasing bend with an oblique (45deg) bend connecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Insets in this and other figures for experimentalmodels show initial fault cuts in the stiff clay Panels (AndashF) show the progressive evolution of the structure with increasing strike-slipdisplacement Rose diagrams show the lengths of faults for different trends in 5deg increments as a percentage of the total length of faults

Mitra and Paul 1153

Figure 5 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with an oblique (45deg) bend connectingtwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement Contours in this and subsequent diagrams are in intervals of 04 mm Every fifth contour is shown in black Color gridsreflect structural elevation with red and yellow representing high areas and blue and green representing low areas Note the formation oftwo diametrically opposite lows in the pull-apart basin (G H) Cross sections through top of clay showing final basin geometry and shift inpolarity for cross sections AAprime and BBprime Locations of the cross sections are shown in (F)

1154 Structural Geometry and Evolution of Releasing and Restraining Bends

strike-slip faults that define the boundaries of thepull-apart basin The angular measurements citedpreviously and in all experiments on pull-apart ba-sins are made in a counterclockwise direction fromthe strike-slip faults The individual faults are madeup of two or more right-stepping en echelon seg-ments some of which eventually interact to formlonger faults With progressive deformation thesefaults systems are rotated slightly by shear

Immediately above the main vertical strike-slipfaults short segments of steep faults of two mainsets trending approximately 15 and 80deg to the base-ment strike-slip faults develop (Figure 4B C) Thesefaults are interpreted to be the R and Rprime Riedelshears associated with strike-slip faulting describedand experimentally modeled by Tchalenko (1970)Tchalenko and Ambraseys (1970) Wilcox et al(1973) and Harding (1985)

Figure 6 Clay models showing faults in the sedimentary cover for a releasing bend with a transverse (90deg) bend connecting two strike-slip fault segments in the basement Sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution of thestructure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

Mitra and Paul 1155

Figure 7 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with a transverse (90deg) bend con-necting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slip displacement

1156 Structural Geometry and Evolution of Releasing and Restraining Bends

The normal faults in the central part of thebasin curve into the orientations of the R shearsin the vicinity of the basement strike-slip faults(Figures 4B C 5B C) The resulting fault systemshave a sigmoidal geometry Rose diagrams of faultorientations show the progressive developmentof scatter because of the development of curvedfaults with varying components of normal andstrike-slip motion (Figure 4AndashC) Twomajor strike-slip faults develop at the surface directly above thepreexisting cuts in the basement after about 08 cmof displacement

Continuing deformation results in additional setsof major normal faults inward and approximatelyparallel to the original pair of fault systems Addi-tional fault systems with the same overall sigmoidalgeometry form in other parts of the transfer zoneand outside the central depression (Figures 4BndashE5BndashE) In the final stages of deformation the lon-gest pair of oblique faults connects with the twomain systems of strike-slip faults thus defining theboundary of the pull-apart basin (Figures 4F 5F)The final rose diagram shows significant scatter oftrends ranging from north to west because of thevariation from strike-slip to normal fault trends witha maximum density in an approximately north-west orientation (Figure 4F)

Examination of the contour patterns (Figure 5)suggests that the structural elevation steps downquite steeply across the normal faults bounding thepull-apart basin but drops more gradually from thebounding system of strike-slip faults to the centerof the basin The central part consists of two sep-arate depressions separated by a local high thatdevelop during the late stages (Figure 5E F) Theaxis connecting the centers of the two depressionsis approximately parallel to the trend of the mainstrike-slip faults Cross sections through the twodepressions show a flip in polarity because of achange in fault displacement along trend withinthe basin (Figure 5G H)

Transverse (90deg) BendFor a transverse bend (90deg) the normal fault seg-ments in the transfer zone form at a higher angle(58ndash78deg) than in the case of the oblique jog (45deg)and show left- and right-stepping patterns on op-

posite sides of the transfer zone (Figures 6A B 7AB) Immediately above the basement strike-slipfaults R and Rprime Riedel shears develop as a series ofen echelon faults making angles of 15 and 82deg to themain faults respectively The normal faults in thecenter of the pull-apart basin curve sharply clock-wise toward the orientation of the R shears in thevicinity of the main strike-slip fault zones along theboundaries of the pull-apart basin (Figures 6C D7C D)With continuing shear the major strike-slipfaults break through to the surface as Y shears de-fining the boundaries of pull-apart basin (Figures 6EF 7E F) As in the case of the oblique-bend ex-periment multiple sets of faults form in the pull-apart basin especially close to the main strike-slipfault zone Theorientations of the faults showawidevariation in trends ranging from northndashnorthwest towestndashnorthwest with the most dominant trend inthe center of the basin being about northwest in thefinal stage (Figure 6F)

Because of the sharper bends and changes inthe trends of the faults in the transfer zone the pull-apart basin associated with a transverse bend has amore S-shaped or rhomboidal geometry than thespindle-shaped geometry in the case of an obliquebend The contours show that the deepest part ofthe basin is located toward one of the major transferfaults so that the pull-apart basin has a more asym-metric shape with a sharper drop of contours onone side (Figure 7E F) The sides of the basin aremarked by a more gradual change in elevation withthe structure contours cut by secondary faults Dif-ferent basin shapes may result depending upon theslip distribution among the different faults

Transverse (90deg) OffsetThis setup has two cuts in the basement that even-tually form the bounding strike-slip faults offsetalong a transfer zone The main purpose of con-ducting this experiment was to understand the dif-ference in the geometry of the pull-apart basinwhenno connecting bend exists in the basement faultThe faults in the transfer zone develop with on-going strike-slip deformation independent of a con-nected basement fault (Figures 8 9) Our experi-ments show that a transverse offset results in theoptimum configuration for the development of a

Mitra and Paul 1157

transfer zone of connecting faults Experiments in-volving overlapping or approaching patterns do notresult in well-defined pull-apart basins because theyboth involve a large transfer zone area This suggeststhat for two approaching strike-slip fault segmentsconnecting faults in the transfer zone will mostlikely start to develop when the end points of thetwo strike-slip segments are directly opposite eachother in a transverse offset position

1158 Structural Geometry and Evolution of Releasing and Rest

Initial deformation results in the formation ofa large number of short en echelon normal faults inthe area of the transfer zone (Figures 8A B 9A B)The central parts of the faults trend at angles of 53to 63deg to themain strike-slip fault zone At the sametime a series of en echelon R and Rprime Riedel shearsform above the main basement faults at angles of18 and 70deg with respect to the basement strike-slip faults respectivelyWith increasing strike-slip

Figure 8 Clay models showing faults in the sedimentary cover for a releasing offset with a transverse (90deg) offset between two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution ofthe structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

raining Bends

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

definition of fault geometries and evolution and alsoenables the deformation to be studied for ma-ture structures involving steep surfaces The ex-perimental setup consists of two overlapping metalplates one of which is attached to a moving platedriven by motors whereas the other is attached toa fixed plate The lateral motion between the twobase plates drives the strike-slip motion in the over-lying stiff and soft clay representing the basementand sedimentary layers (Figure 2) No shearing oc-curs along the contact between the stiff clay andthe sliding plates Basement is represented by stiffclay with a density of 185 gcc and a thickness of1 in (25 cm) The overlying sedimentary layer isrepresented by soft water-based clay with a densityof 16 to 165 gcc and a thickness of 08 in (2 cm)The density of the clay is adjusted by altering thewater content of the clay so that the clay has theappropriate consistency to deform by faulting with-out developing large cracks No layering is usedwithin the soft clay The displacement rate for allexperiments is 0001 cms corresponding to a strainrate of approximately 3 times 10ndash5 per second

The experimental configurations used in thisstudy differ from previous setups (Dooley andMcClay 1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006Wu et al 2009) in thatstrike-slip motion is induced in both the stiff clay(basement) and the overlying soft clay (sedimentarycover) through sliding base plates at depth There-fore the sliding plates only control the nature ofdeformation and the location of the bend or offsetThis approach enables us to study differences indeformation involving releasing bends and offsetsin the strike-slip faults within the basement whereasprevious experiments modeled only releasing bendsFurthermore using stiff clay to model basementinstead of rigid plates the basement deforms in theimmediate vicinity of the faults to accommodatethe deformation in these areas In previousmodelsthe top of the basement was represented by slidingplates rendering the basement perfectly rigid inthe vicinity of the faults

Two experimental setups for the master faultsin the basement are investigated (1) bends inwhichtwo parallel strike-slip faults are connected by anoblique (Figure 2A B) or transverse (Figure 2C D)

bend in the stiff clay representing the basementand (2) offsets in which two separate strike-slipfault segments in the basement are separated by a90deg offset in the stiff clay (Figure 2E F) The initialconfiguration for the bend setup consists of twovertical cuts in the stiff clay connected by obliqueor transverse bends whereas the configuration forthe offset setup consists of two parallel and offsetvertical cuts In both cases the basement is under-lain by two sliding plates that override each other toprovide the strike-slip motion (Serra and Nelson1988) The bend configuration may represent anoriginal geometry of the strike-slip fault (Crowell1974) or may result from two separate parallel seg-ments that have connected along a transverse oroblique segment The offset configuration morespecifically addresses the linking of two adjacentstrike-slip segments by the development of a trans-fer zone consisting of single or multiple obliquefaults These configurations result in significantdifferences in the geometry and faulting associatedwith the pull-apart basins Two runs of each ex-periment are conducted to test for repeatability

The progressive evolution of structures and sec-ondary faults and their orientations within a strike-slip system is mapped directly from photographstaken during various stages of the experiments Inaddition we use a new technique involving theanalysis of the 3-D geometry of the top surface inthe structures using laser scanning (see also Boseand Mitra 2010) The technology of scanning sur-faces by a laser scanner and developing a virtual3-D model has been in use in numerous industriesparticularly in 3-D animation The same technol-ogy is applied here to gather 3-D information andthereby generate a virtual surface that can be used tovisualize the development of the structure in muchgreater detail and to generate contour maps of sur-faces The scanning size is 135 times 101 in (343 times257 cm) with a resolution of 75 dots per inch(sim0015 in point density) The scanner is placedface down approximately 15 in (38 cm) verticallyabove the clay cake and top surfaces are scanned atapproximately equal increments of deformation

The sensor captures the laser beam that is pro-jectedon the surface and thecoordinatesof thepointsare computed by the scanner using a triangulation

Mitra and Paul 1151

method (Figure 3A) The known parameters arethe angle a at which the laser points out its beamon the surface the angle b at which the sensor col-lects the laser beam (b is known if the focal lengthand the pixel size of the sensor are given) and thetriangulation distance between the sensor and lasersource (Petrov et al 1998) Because all the geo-metric parameters are known the coordinates x yand z on the surface are calculated using trigono-metric methods The laser source produces a lineand all data along that line are collected at the sametime (Figure 3B) Moreover twin arrays of four laserbeams are used as source in this instrument for

1152 Structural Geometry and Evolution of Releasing and Rest

cross-validating each data point by measuring it atleast twice The origin or reference coordinate sys-tem is determined automatically during each run bythe first point scanned on the surface It is thereforenot necessary to move the scanner during the lengthof the experiment to spatially position the scannedsurfaces in the same reference coordinate system

The scanned data are exported as point files intothe 3-D modeling software GoCAD (Bose andMitra 2010) The point clouds obtained (Figure 3C)are then used to build a surface Fault polygons arebuilt from visible breaks in the modeled surface(Figure 3D) and the geometries of the polygons

Figure 3 Laser scanning of experimental clay models (A) Schematic diagram of the working principle of the laser scanner withthe triangulation formed by the laser source the detector and the point on the clay surface from which the laser beam is reflected(B) Schematic view of the laser scanner projecting a line and sweeping across the clay surface at a constant velocity (from Bose andMitra 2010 used with permission from AAPG) (C) Cloud of points obtained from the scanned surface and visualized in GO-CAD(D) Modeled surface in GO-CAD cut by faults (E) Oblique photograph of the top of the clay surface

raining Bends

are checked against the traced fault cuts on photo-graphs taken of the clay surface which provide de-tails of the fault geometry (Figure 3E) Themodeledsurface is cut by the fault polygons and contouredto depict their topology (Figure 3D) The contouredfaulted surface provides an accurate quantitativerendition of the structural geometry which can becompared directly with structure contour maps onnatural structures

Experimental Results

Oblique (45deg) BendInitial movement on the base plate results in a broaddepression oblique to the preexisting basementstrike-slip faults in the area above the obliquebend (Figures 4A 5A) This depression is boundedby two sets of faults dipping toward each otherand trending approximately 42deg to the basement

Figure 4 Clay models showing faults in the sedimentary cover for a releasing bend with an oblique (45deg) bend connecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Insets in this and other figures for experimentalmodels show initial fault cuts in the stiff clay Panels (AndashF) show the progressive evolution of the structure with increasing strike-slipdisplacement Rose diagrams show the lengths of faults for different trends in 5deg increments as a percentage of the total length of faults

Mitra and Paul 1153

Figure 5 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with an oblique (45deg) bend connectingtwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement Contours in this and subsequent diagrams are in intervals of 04 mm Every fifth contour is shown in black Color gridsreflect structural elevation with red and yellow representing high areas and blue and green representing low areas Note the formation oftwo diametrically opposite lows in the pull-apart basin (G H) Cross sections through top of clay showing final basin geometry and shift inpolarity for cross sections AAprime and BBprime Locations of the cross sections are shown in (F)

1154 Structural Geometry and Evolution of Releasing and Restraining Bends

strike-slip faults that define the boundaries of thepull-apart basin The angular measurements citedpreviously and in all experiments on pull-apart ba-sins are made in a counterclockwise direction fromthe strike-slip faults The individual faults are madeup of two or more right-stepping en echelon seg-ments some of which eventually interact to formlonger faults With progressive deformation thesefaults systems are rotated slightly by shear

Immediately above the main vertical strike-slipfaults short segments of steep faults of two mainsets trending approximately 15 and 80deg to the base-ment strike-slip faults develop (Figure 4B C) Thesefaults are interpreted to be the R and Rprime Riedelshears associated with strike-slip faulting describedand experimentally modeled by Tchalenko (1970)Tchalenko and Ambraseys (1970) Wilcox et al(1973) and Harding (1985)

Figure 6 Clay models showing faults in the sedimentary cover for a releasing bend with a transverse (90deg) bend connecting two strike-slip fault segments in the basement Sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution of thestructure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

Mitra and Paul 1155

Figure 7 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with a transverse (90deg) bend con-necting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slip displacement

1156 Structural Geometry and Evolution of Releasing and Restraining Bends

The normal faults in the central part of thebasin curve into the orientations of the R shearsin the vicinity of the basement strike-slip faults(Figures 4B C 5B C) The resulting fault systemshave a sigmoidal geometry Rose diagrams of faultorientations show the progressive developmentof scatter because of the development of curvedfaults with varying components of normal andstrike-slip motion (Figure 4AndashC) Twomajor strike-slip faults develop at the surface directly above thepreexisting cuts in the basement after about 08 cmof displacement

Continuing deformation results in additional setsof major normal faults inward and approximatelyparallel to the original pair of fault systems Addi-tional fault systems with the same overall sigmoidalgeometry form in other parts of the transfer zoneand outside the central depression (Figures 4BndashE5BndashE) In the final stages of deformation the lon-gest pair of oblique faults connects with the twomain systems of strike-slip faults thus defining theboundary of the pull-apart basin (Figures 4F 5F)The final rose diagram shows significant scatter oftrends ranging from north to west because of thevariation from strike-slip to normal fault trends witha maximum density in an approximately north-west orientation (Figure 4F)

Examination of the contour patterns (Figure 5)suggests that the structural elevation steps downquite steeply across the normal faults bounding thepull-apart basin but drops more gradually from thebounding system of strike-slip faults to the centerof the basin The central part consists of two sep-arate depressions separated by a local high thatdevelop during the late stages (Figure 5E F) Theaxis connecting the centers of the two depressionsis approximately parallel to the trend of the mainstrike-slip faults Cross sections through the twodepressions show a flip in polarity because of achange in fault displacement along trend withinthe basin (Figure 5G H)

Transverse (90deg) BendFor a transverse bend (90deg) the normal fault seg-ments in the transfer zone form at a higher angle(58ndash78deg) than in the case of the oblique jog (45deg)and show left- and right-stepping patterns on op-

posite sides of the transfer zone (Figures 6A B 7AB) Immediately above the basement strike-slipfaults R and Rprime Riedel shears develop as a series ofen echelon faults making angles of 15 and 82deg to themain faults respectively The normal faults in thecenter of the pull-apart basin curve sharply clock-wise toward the orientation of the R shears in thevicinity of the main strike-slip fault zones along theboundaries of the pull-apart basin (Figures 6C D7C D)With continuing shear the major strike-slipfaults break through to the surface as Y shears de-fining the boundaries of pull-apart basin (Figures 6EF 7E F) As in the case of the oblique-bend ex-periment multiple sets of faults form in the pull-apart basin especially close to the main strike-slipfault zone Theorientations of the faults showawidevariation in trends ranging from northndashnorthwest towestndashnorthwest with the most dominant trend inthe center of the basin being about northwest in thefinal stage (Figure 6F)

Because of the sharper bends and changes inthe trends of the faults in the transfer zone the pull-apart basin associated with a transverse bend has amore S-shaped or rhomboidal geometry than thespindle-shaped geometry in the case of an obliquebend The contours show that the deepest part ofthe basin is located toward one of the major transferfaults so that the pull-apart basin has a more asym-metric shape with a sharper drop of contours onone side (Figure 7E F) The sides of the basin aremarked by a more gradual change in elevation withthe structure contours cut by secondary faults Dif-ferent basin shapes may result depending upon theslip distribution among the different faults

Transverse (90deg) OffsetThis setup has two cuts in the basement that even-tually form the bounding strike-slip faults offsetalong a transfer zone The main purpose of con-ducting this experiment was to understand the dif-ference in the geometry of the pull-apart basinwhenno connecting bend exists in the basement faultThe faults in the transfer zone develop with on-going strike-slip deformation independent of a con-nected basement fault (Figures 8 9) Our experi-ments show that a transverse offset results in theoptimum configuration for the development of a

Mitra and Paul 1157

transfer zone of connecting faults Experiments in-volving overlapping or approaching patterns do notresult in well-defined pull-apart basins because theyboth involve a large transfer zone area This suggeststhat for two approaching strike-slip fault segmentsconnecting faults in the transfer zone will mostlikely start to develop when the end points of thetwo strike-slip segments are directly opposite eachother in a transverse offset position

1158 Structural Geometry and Evolution of Releasing and Rest

Initial deformation results in the formation ofa large number of short en echelon normal faults inthe area of the transfer zone (Figures 8A B 9A B)The central parts of the faults trend at angles of 53to 63deg to themain strike-slip fault zone At the sametime a series of en echelon R and Rprime Riedel shearsform above the main basement faults at angles of18 and 70deg with respect to the basement strike-slip faults respectivelyWith increasing strike-slip

Figure 8 Clay models showing faults in the sedimentary cover for a releasing offset with a transverse (90deg) offset between two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution ofthe structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

raining Bends

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

method (Figure 3A) The known parameters arethe angle a at which the laser points out its beamon the surface the angle b at which the sensor col-lects the laser beam (b is known if the focal lengthand the pixel size of the sensor are given) and thetriangulation distance between the sensor and lasersource (Petrov et al 1998) Because all the geo-metric parameters are known the coordinates x yand z on the surface are calculated using trigono-metric methods The laser source produces a lineand all data along that line are collected at the sametime (Figure 3B) Moreover twin arrays of four laserbeams are used as source in this instrument for

1152 Structural Geometry and Evolution of Releasing and Rest

cross-validating each data point by measuring it atleast twice The origin or reference coordinate sys-tem is determined automatically during each run bythe first point scanned on the surface It is thereforenot necessary to move the scanner during the lengthof the experiment to spatially position the scannedsurfaces in the same reference coordinate system

The scanned data are exported as point files intothe 3-D modeling software GoCAD (Bose andMitra 2010) The point clouds obtained (Figure 3C)are then used to build a surface Fault polygons arebuilt from visible breaks in the modeled surface(Figure 3D) and the geometries of the polygons

Figure 3 Laser scanning of experimental clay models (A) Schematic diagram of the working principle of the laser scanner withthe triangulation formed by the laser source the detector and the point on the clay surface from which the laser beam is reflected(B) Schematic view of the laser scanner projecting a line and sweeping across the clay surface at a constant velocity (from Bose andMitra 2010 used with permission from AAPG) (C) Cloud of points obtained from the scanned surface and visualized in GO-CAD(D) Modeled surface in GO-CAD cut by faults (E) Oblique photograph of the top of the clay surface

raining Bends

are checked against the traced fault cuts on photo-graphs taken of the clay surface which provide de-tails of the fault geometry (Figure 3E) Themodeledsurface is cut by the fault polygons and contouredto depict their topology (Figure 3D) The contouredfaulted surface provides an accurate quantitativerendition of the structural geometry which can becompared directly with structure contour maps onnatural structures

Experimental Results

Oblique (45deg) BendInitial movement on the base plate results in a broaddepression oblique to the preexisting basementstrike-slip faults in the area above the obliquebend (Figures 4A 5A) This depression is boundedby two sets of faults dipping toward each otherand trending approximately 42deg to the basement

Figure 4 Clay models showing faults in the sedimentary cover for a releasing bend with an oblique (45deg) bend connecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Insets in this and other figures for experimentalmodels show initial fault cuts in the stiff clay Panels (AndashF) show the progressive evolution of the structure with increasing strike-slipdisplacement Rose diagrams show the lengths of faults for different trends in 5deg increments as a percentage of the total length of faults

Mitra and Paul 1153

Figure 5 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with an oblique (45deg) bend connectingtwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement Contours in this and subsequent diagrams are in intervals of 04 mm Every fifth contour is shown in black Color gridsreflect structural elevation with red and yellow representing high areas and blue and green representing low areas Note the formation oftwo diametrically opposite lows in the pull-apart basin (G H) Cross sections through top of clay showing final basin geometry and shift inpolarity for cross sections AAprime and BBprime Locations of the cross sections are shown in (F)

1154 Structural Geometry and Evolution of Releasing and Restraining Bends

strike-slip faults that define the boundaries of thepull-apart basin The angular measurements citedpreviously and in all experiments on pull-apart ba-sins are made in a counterclockwise direction fromthe strike-slip faults The individual faults are madeup of two or more right-stepping en echelon seg-ments some of which eventually interact to formlonger faults With progressive deformation thesefaults systems are rotated slightly by shear

Immediately above the main vertical strike-slipfaults short segments of steep faults of two mainsets trending approximately 15 and 80deg to the base-ment strike-slip faults develop (Figure 4B C) Thesefaults are interpreted to be the R and Rprime Riedelshears associated with strike-slip faulting describedand experimentally modeled by Tchalenko (1970)Tchalenko and Ambraseys (1970) Wilcox et al(1973) and Harding (1985)

Figure 6 Clay models showing faults in the sedimentary cover for a releasing bend with a transverse (90deg) bend connecting two strike-slip fault segments in the basement Sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution of thestructure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

Mitra and Paul 1155

Figure 7 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with a transverse (90deg) bend con-necting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slip displacement

1156 Structural Geometry and Evolution of Releasing and Restraining Bends

The normal faults in the central part of thebasin curve into the orientations of the R shearsin the vicinity of the basement strike-slip faults(Figures 4B C 5B C) The resulting fault systemshave a sigmoidal geometry Rose diagrams of faultorientations show the progressive developmentof scatter because of the development of curvedfaults with varying components of normal andstrike-slip motion (Figure 4AndashC) Twomajor strike-slip faults develop at the surface directly above thepreexisting cuts in the basement after about 08 cmof displacement

Continuing deformation results in additional setsof major normal faults inward and approximatelyparallel to the original pair of fault systems Addi-tional fault systems with the same overall sigmoidalgeometry form in other parts of the transfer zoneand outside the central depression (Figures 4BndashE5BndashE) In the final stages of deformation the lon-gest pair of oblique faults connects with the twomain systems of strike-slip faults thus defining theboundary of the pull-apart basin (Figures 4F 5F)The final rose diagram shows significant scatter oftrends ranging from north to west because of thevariation from strike-slip to normal fault trends witha maximum density in an approximately north-west orientation (Figure 4F)

Examination of the contour patterns (Figure 5)suggests that the structural elevation steps downquite steeply across the normal faults bounding thepull-apart basin but drops more gradually from thebounding system of strike-slip faults to the centerof the basin The central part consists of two sep-arate depressions separated by a local high thatdevelop during the late stages (Figure 5E F) Theaxis connecting the centers of the two depressionsis approximately parallel to the trend of the mainstrike-slip faults Cross sections through the twodepressions show a flip in polarity because of achange in fault displacement along trend withinthe basin (Figure 5G H)

Transverse (90deg) BendFor a transverse bend (90deg) the normal fault seg-ments in the transfer zone form at a higher angle(58ndash78deg) than in the case of the oblique jog (45deg)and show left- and right-stepping patterns on op-

posite sides of the transfer zone (Figures 6A B 7AB) Immediately above the basement strike-slipfaults R and Rprime Riedel shears develop as a series ofen echelon faults making angles of 15 and 82deg to themain faults respectively The normal faults in thecenter of the pull-apart basin curve sharply clock-wise toward the orientation of the R shears in thevicinity of the main strike-slip fault zones along theboundaries of the pull-apart basin (Figures 6C D7C D)With continuing shear the major strike-slipfaults break through to the surface as Y shears de-fining the boundaries of pull-apart basin (Figures 6EF 7E F) As in the case of the oblique-bend ex-periment multiple sets of faults form in the pull-apart basin especially close to the main strike-slipfault zone Theorientations of the faults showawidevariation in trends ranging from northndashnorthwest towestndashnorthwest with the most dominant trend inthe center of the basin being about northwest in thefinal stage (Figure 6F)

Because of the sharper bends and changes inthe trends of the faults in the transfer zone the pull-apart basin associated with a transverse bend has amore S-shaped or rhomboidal geometry than thespindle-shaped geometry in the case of an obliquebend The contours show that the deepest part ofthe basin is located toward one of the major transferfaults so that the pull-apart basin has a more asym-metric shape with a sharper drop of contours onone side (Figure 7E F) The sides of the basin aremarked by a more gradual change in elevation withthe structure contours cut by secondary faults Dif-ferent basin shapes may result depending upon theslip distribution among the different faults

Transverse (90deg) OffsetThis setup has two cuts in the basement that even-tually form the bounding strike-slip faults offsetalong a transfer zone The main purpose of con-ducting this experiment was to understand the dif-ference in the geometry of the pull-apart basinwhenno connecting bend exists in the basement faultThe faults in the transfer zone develop with on-going strike-slip deformation independent of a con-nected basement fault (Figures 8 9) Our experi-ments show that a transverse offset results in theoptimum configuration for the development of a

Mitra and Paul 1157

transfer zone of connecting faults Experiments in-volving overlapping or approaching patterns do notresult in well-defined pull-apart basins because theyboth involve a large transfer zone area This suggeststhat for two approaching strike-slip fault segmentsconnecting faults in the transfer zone will mostlikely start to develop when the end points of thetwo strike-slip segments are directly opposite eachother in a transverse offset position

1158 Structural Geometry and Evolution of Releasing and Rest

Initial deformation results in the formation ofa large number of short en echelon normal faults inthe area of the transfer zone (Figures 8A B 9A B)The central parts of the faults trend at angles of 53to 63deg to themain strike-slip fault zone At the sametime a series of en echelon R and Rprime Riedel shearsform above the main basement faults at angles of18 and 70deg with respect to the basement strike-slip faults respectivelyWith increasing strike-slip

Figure 8 Clay models showing faults in the sedimentary cover for a releasing offset with a transverse (90deg) offset between two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution ofthe structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

raining Bends

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

are checked against the traced fault cuts on photo-graphs taken of the clay surface which provide de-tails of the fault geometry (Figure 3E) Themodeledsurface is cut by the fault polygons and contouredto depict their topology (Figure 3D) The contouredfaulted surface provides an accurate quantitativerendition of the structural geometry which can becompared directly with structure contour maps onnatural structures

Experimental Results

Oblique (45deg) BendInitial movement on the base plate results in a broaddepression oblique to the preexisting basementstrike-slip faults in the area above the obliquebend (Figures 4A 5A) This depression is boundedby two sets of faults dipping toward each otherand trending approximately 42deg to the basement

Figure 4 Clay models showing faults in the sedimentary cover for a releasing bend with an oblique (45deg) bend connecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Insets in this and other figures for experimentalmodels show initial fault cuts in the stiff clay Panels (AndashF) show the progressive evolution of the structure with increasing strike-slipdisplacement Rose diagrams show the lengths of faults for different trends in 5deg increments as a percentage of the total length of faults

Mitra and Paul 1153

Figure 5 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with an oblique (45deg) bend connectingtwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement Contours in this and subsequent diagrams are in intervals of 04 mm Every fifth contour is shown in black Color gridsreflect structural elevation with red and yellow representing high areas and blue and green representing low areas Note the formation oftwo diametrically opposite lows in the pull-apart basin (G H) Cross sections through top of clay showing final basin geometry and shift inpolarity for cross sections AAprime and BBprime Locations of the cross sections are shown in (F)

1154 Structural Geometry and Evolution of Releasing and Restraining Bends

strike-slip faults that define the boundaries of thepull-apart basin The angular measurements citedpreviously and in all experiments on pull-apart ba-sins are made in a counterclockwise direction fromthe strike-slip faults The individual faults are madeup of two or more right-stepping en echelon seg-ments some of which eventually interact to formlonger faults With progressive deformation thesefaults systems are rotated slightly by shear

Immediately above the main vertical strike-slipfaults short segments of steep faults of two mainsets trending approximately 15 and 80deg to the base-ment strike-slip faults develop (Figure 4B C) Thesefaults are interpreted to be the R and Rprime Riedelshears associated with strike-slip faulting describedand experimentally modeled by Tchalenko (1970)Tchalenko and Ambraseys (1970) Wilcox et al(1973) and Harding (1985)

Figure 6 Clay models showing faults in the sedimentary cover for a releasing bend with a transverse (90deg) bend connecting two strike-slip fault segments in the basement Sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution of thestructure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

Mitra and Paul 1155

Figure 7 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with a transverse (90deg) bend con-necting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slip displacement

1156 Structural Geometry and Evolution of Releasing and Restraining Bends

The normal faults in the central part of thebasin curve into the orientations of the R shearsin the vicinity of the basement strike-slip faults(Figures 4B C 5B C) The resulting fault systemshave a sigmoidal geometry Rose diagrams of faultorientations show the progressive developmentof scatter because of the development of curvedfaults with varying components of normal andstrike-slip motion (Figure 4AndashC) Twomajor strike-slip faults develop at the surface directly above thepreexisting cuts in the basement after about 08 cmof displacement

Continuing deformation results in additional setsof major normal faults inward and approximatelyparallel to the original pair of fault systems Addi-tional fault systems with the same overall sigmoidalgeometry form in other parts of the transfer zoneand outside the central depression (Figures 4BndashE5BndashE) In the final stages of deformation the lon-gest pair of oblique faults connects with the twomain systems of strike-slip faults thus defining theboundary of the pull-apart basin (Figures 4F 5F)The final rose diagram shows significant scatter oftrends ranging from north to west because of thevariation from strike-slip to normal fault trends witha maximum density in an approximately north-west orientation (Figure 4F)

Examination of the contour patterns (Figure 5)suggests that the structural elevation steps downquite steeply across the normal faults bounding thepull-apart basin but drops more gradually from thebounding system of strike-slip faults to the centerof the basin The central part consists of two sep-arate depressions separated by a local high thatdevelop during the late stages (Figure 5E F) Theaxis connecting the centers of the two depressionsis approximately parallel to the trend of the mainstrike-slip faults Cross sections through the twodepressions show a flip in polarity because of achange in fault displacement along trend withinthe basin (Figure 5G H)

Transverse (90deg) BendFor a transverse bend (90deg) the normal fault seg-ments in the transfer zone form at a higher angle(58ndash78deg) than in the case of the oblique jog (45deg)and show left- and right-stepping patterns on op-

posite sides of the transfer zone (Figures 6A B 7AB) Immediately above the basement strike-slipfaults R and Rprime Riedel shears develop as a series ofen echelon faults making angles of 15 and 82deg to themain faults respectively The normal faults in thecenter of the pull-apart basin curve sharply clock-wise toward the orientation of the R shears in thevicinity of the main strike-slip fault zones along theboundaries of the pull-apart basin (Figures 6C D7C D)With continuing shear the major strike-slipfaults break through to the surface as Y shears de-fining the boundaries of pull-apart basin (Figures 6EF 7E F) As in the case of the oblique-bend ex-periment multiple sets of faults form in the pull-apart basin especially close to the main strike-slipfault zone Theorientations of the faults showawidevariation in trends ranging from northndashnorthwest towestndashnorthwest with the most dominant trend inthe center of the basin being about northwest in thefinal stage (Figure 6F)

Because of the sharper bends and changes inthe trends of the faults in the transfer zone the pull-apart basin associated with a transverse bend has amore S-shaped or rhomboidal geometry than thespindle-shaped geometry in the case of an obliquebend The contours show that the deepest part ofthe basin is located toward one of the major transferfaults so that the pull-apart basin has a more asym-metric shape with a sharper drop of contours onone side (Figure 7E F) The sides of the basin aremarked by a more gradual change in elevation withthe structure contours cut by secondary faults Dif-ferent basin shapes may result depending upon theslip distribution among the different faults

Transverse (90deg) OffsetThis setup has two cuts in the basement that even-tually form the bounding strike-slip faults offsetalong a transfer zone The main purpose of con-ducting this experiment was to understand the dif-ference in the geometry of the pull-apart basinwhenno connecting bend exists in the basement faultThe faults in the transfer zone develop with on-going strike-slip deformation independent of a con-nected basement fault (Figures 8 9) Our experi-ments show that a transverse offset results in theoptimum configuration for the development of a

Mitra and Paul 1157

transfer zone of connecting faults Experiments in-volving overlapping or approaching patterns do notresult in well-defined pull-apart basins because theyboth involve a large transfer zone area This suggeststhat for two approaching strike-slip fault segmentsconnecting faults in the transfer zone will mostlikely start to develop when the end points of thetwo strike-slip segments are directly opposite eachother in a transverse offset position

1158 Structural Geometry and Evolution of Releasing and Rest

Initial deformation results in the formation ofa large number of short en echelon normal faults inthe area of the transfer zone (Figures 8A B 9A B)The central parts of the faults trend at angles of 53to 63deg to themain strike-slip fault zone At the sametime a series of en echelon R and Rprime Riedel shearsform above the main basement faults at angles of18 and 70deg with respect to the basement strike-slip faults respectivelyWith increasing strike-slip

Figure 8 Clay models showing faults in the sedimentary cover for a releasing offset with a transverse (90deg) offset between two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution ofthe structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

raining Bends

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

Figure 5 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with an oblique (45deg) bend connectingtwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement Contours in this and subsequent diagrams are in intervals of 04 mm Every fifth contour is shown in black Color gridsreflect structural elevation with red and yellow representing high areas and blue and green representing low areas Note the formation oftwo diametrically opposite lows in the pull-apart basin (G H) Cross sections through top of clay showing final basin geometry and shift inpolarity for cross sections AAprime and BBprime Locations of the cross sections are shown in (F)

1154 Structural Geometry and Evolution of Releasing and Restraining Bends

strike-slip faults that define the boundaries of thepull-apart basin The angular measurements citedpreviously and in all experiments on pull-apart ba-sins are made in a counterclockwise direction fromthe strike-slip faults The individual faults are madeup of two or more right-stepping en echelon seg-ments some of which eventually interact to formlonger faults With progressive deformation thesefaults systems are rotated slightly by shear

Immediately above the main vertical strike-slipfaults short segments of steep faults of two mainsets trending approximately 15 and 80deg to the base-ment strike-slip faults develop (Figure 4B C) Thesefaults are interpreted to be the R and Rprime Riedelshears associated with strike-slip faulting describedand experimentally modeled by Tchalenko (1970)Tchalenko and Ambraseys (1970) Wilcox et al(1973) and Harding (1985)

Figure 6 Clay models showing faults in the sedimentary cover for a releasing bend with a transverse (90deg) bend connecting two strike-slip fault segments in the basement Sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution of thestructure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

Mitra and Paul 1155

Figure 7 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with a transverse (90deg) bend con-necting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slip displacement

1156 Structural Geometry and Evolution of Releasing and Restraining Bends

The normal faults in the central part of thebasin curve into the orientations of the R shearsin the vicinity of the basement strike-slip faults(Figures 4B C 5B C) The resulting fault systemshave a sigmoidal geometry Rose diagrams of faultorientations show the progressive developmentof scatter because of the development of curvedfaults with varying components of normal andstrike-slip motion (Figure 4AndashC) Twomajor strike-slip faults develop at the surface directly above thepreexisting cuts in the basement after about 08 cmof displacement

Continuing deformation results in additional setsof major normal faults inward and approximatelyparallel to the original pair of fault systems Addi-tional fault systems with the same overall sigmoidalgeometry form in other parts of the transfer zoneand outside the central depression (Figures 4BndashE5BndashE) In the final stages of deformation the lon-gest pair of oblique faults connects with the twomain systems of strike-slip faults thus defining theboundary of the pull-apart basin (Figures 4F 5F)The final rose diagram shows significant scatter oftrends ranging from north to west because of thevariation from strike-slip to normal fault trends witha maximum density in an approximately north-west orientation (Figure 4F)

Examination of the contour patterns (Figure 5)suggests that the structural elevation steps downquite steeply across the normal faults bounding thepull-apart basin but drops more gradually from thebounding system of strike-slip faults to the centerof the basin The central part consists of two sep-arate depressions separated by a local high thatdevelop during the late stages (Figure 5E F) Theaxis connecting the centers of the two depressionsis approximately parallel to the trend of the mainstrike-slip faults Cross sections through the twodepressions show a flip in polarity because of achange in fault displacement along trend withinthe basin (Figure 5G H)

Transverse (90deg) BendFor a transverse bend (90deg) the normal fault seg-ments in the transfer zone form at a higher angle(58ndash78deg) than in the case of the oblique jog (45deg)and show left- and right-stepping patterns on op-

posite sides of the transfer zone (Figures 6A B 7AB) Immediately above the basement strike-slipfaults R and Rprime Riedel shears develop as a series ofen echelon faults making angles of 15 and 82deg to themain faults respectively The normal faults in thecenter of the pull-apart basin curve sharply clock-wise toward the orientation of the R shears in thevicinity of the main strike-slip fault zones along theboundaries of the pull-apart basin (Figures 6C D7C D)With continuing shear the major strike-slipfaults break through to the surface as Y shears de-fining the boundaries of pull-apart basin (Figures 6EF 7E F) As in the case of the oblique-bend ex-periment multiple sets of faults form in the pull-apart basin especially close to the main strike-slipfault zone Theorientations of the faults showawidevariation in trends ranging from northndashnorthwest towestndashnorthwest with the most dominant trend inthe center of the basin being about northwest in thefinal stage (Figure 6F)

Because of the sharper bends and changes inthe trends of the faults in the transfer zone the pull-apart basin associated with a transverse bend has amore S-shaped or rhomboidal geometry than thespindle-shaped geometry in the case of an obliquebend The contours show that the deepest part ofthe basin is located toward one of the major transferfaults so that the pull-apart basin has a more asym-metric shape with a sharper drop of contours onone side (Figure 7E F) The sides of the basin aremarked by a more gradual change in elevation withthe structure contours cut by secondary faults Dif-ferent basin shapes may result depending upon theslip distribution among the different faults

Transverse (90deg) OffsetThis setup has two cuts in the basement that even-tually form the bounding strike-slip faults offsetalong a transfer zone The main purpose of con-ducting this experiment was to understand the dif-ference in the geometry of the pull-apart basinwhenno connecting bend exists in the basement faultThe faults in the transfer zone develop with on-going strike-slip deformation independent of a con-nected basement fault (Figures 8 9) Our experi-ments show that a transverse offset results in theoptimum configuration for the development of a

Mitra and Paul 1157

transfer zone of connecting faults Experiments in-volving overlapping or approaching patterns do notresult in well-defined pull-apart basins because theyboth involve a large transfer zone area This suggeststhat for two approaching strike-slip fault segmentsconnecting faults in the transfer zone will mostlikely start to develop when the end points of thetwo strike-slip segments are directly opposite eachother in a transverse offset position

1158 Structural Geometry and Evolution of Releasing and Rest

Initial deformation results in the formation ofa large number of short en echelon normal faults inthe area of the transfer zone (Figures 8A B 9A B)The central parts of the faults trend at angles of 53to 63deg to themain strike-slip fault zone At the sametime a series of en echelon R and Rprime Riedel shearsform above the main basement faults at angles of18 and 70deg with respect to the basement strike-slip faults respectivelyWith increasing strike-slip

Figure 8 Clay models showing faults in the sedimentary cover for a releasing offset with a transverse (90deg) offset between two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution ofthe structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

raining Bends

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

strike-slip faults that define the boundaries of thepull-apart basin The angular measurements citedpreviously and in all experiments on pull-apart ba-sins are made in a counterclockwise direction fromthe strike-slip faults The individual faults are madeup of two or more right-stepping en echelon seg-ments some of which eventually interact to formlonger faults With progressive deformation thesefaults systems are rotated slightly by shear

Immediately above the main vertical strike-slipfaults short segments of steep faults of two mainsets trending approximately 15 and 80deg to the base-ment strike-slip faults develop (Figure 4B C) Thesefaults are interpreted to be the R and Rprime Riedelshears associated with strike-slip faulting describedand experimentally modeled by Tchalenko (1970)Tchalenko and Ambraseys (1970) Wilcox et al(1973) and Harding (1985)

Figure 6 Clay models showing faults in the sedimentary cover for a releasing bend with a transverse (90deg) bend connecting two strike-slip fault segments in the basement Sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution of thestructure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

Mitra and Paul 1155

Figure 7 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with a transverse (90deg) bend con-necting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slip displacement

1156 Structural Geometry and Evolution of Releasing and Restraining Bends

The normal faults in the central part of thebasin curve into the orientations of the R shearsin the vicinity of the basement strike-slip faults(Figures 4B C 5B C) The resulting fault systemshave a sigmoidal geometry Rose diagrams of faultorientations show the progressive developmentof scatter because of the development of curvedfaults with varying components of normal andstrike-slip motion (Figure 4AndashC) Twomajor strike-slip faults develop at the surface directly above thepreexisting cuts in the basement after about 08 cmof displacement

Continuing deformation results in additional setsof major normal faults inward and approximatelyparallel to the original pair of fault systems Addi-tional fault systems with the same overall sigmoidalgeometry form in other parts of the transfer zoneand outside the central depression (Figures 4BndashE5BndashE) In the final stages of deformation the lon-gest pair of oblique faults connects with the twomain systems of strike-slip faults thus defining theboundary of the pull-apart basin (Figures 4F 5F)The final rose diagram shows significant scatter oftrends ranging from north to west because of thevariation from strike-slip to normal fault trends witha maximum density in an approximately north-west orientation (Figure 4F)

Examination of the contour patterns (Figure 5)suggests that the structural elevation steps downquite steeply across the normal faults bounding thepull-apart basin but drops more gradually from thebounding system of strike-slip faults to the centerof the basin The central part consists of two sep-arate depressions separated by a local high thatdevelop during the late stages (Figure 5E F) Theaxis connecting the centers of the two depressionsis approximately parallel to the trend of the mainstrike-slip faults Cross sections through the twodepressions show a flip in polarity because of achange in fault displacement along trend withinthe basin (Figure 5G H)

Transverse (90deg) BendFor a transverse bend (90deg) the normal fault seg-ments in the transfer zone form at a higher angle(58ndash78deg) than in the case of the oblique jog (45deg)and show left- and right-stepping patterns on op-

posite sides of the transfer zone (Figures 6A B 7AB) Immediately above the basement strike-slipfaults R and Rprime Riedel shears develop as a series ofen echelon faults making angles of 15 and 82deg to themain faults respectively The normal faults in thecenter of the pull-apart basin curve sharply clock-wise toward the orientation of the R shears in thevicinity of the main strike-slip fault zones along theboundaries of the pull-apart basin (Figures 6C D7C D)With continuing shear the major strike-slipfaults break through to the surface as Y shears de-fining the boundaries of pull-apart basin (Figures 6EF 7E F) As in the case of the oblique-bend ex-periment multiple sets of faults form in the pull-apart basin especially close to the main strike-slipfault zone Theorientations of the faults showawidevariation in trends ranging from northndashnorthwest towestndashnorthwest with the most dominant trend inthe center of the basin being about northwest in thefinal stage (Figure 6F)

Because of the sharper bends and changes inthe trends of the faults in the transfer zone the pull-apart basin associated with a transverse bend has amore S-shaped or rhomboidal geometry than thespindle-shaped geometry in the case of an obliquebend The contours show that the deepest part ofthe basin is located toward one of the major transferfaults so that the pull-apart basin has a more asym-metric shape with a sharper drop of contours onone side (Figure 7E F) The sides of the basin aremarked by a more gradual change in elevation withthe structure contours cut by secondary faults Dif-ferent basin shapes may result depending upon theslip distribution among the different faults

Transverse (90deg) OffsetThis setup has two cuts in the basement that even-tually form the bounding strike-slip faults offsetalong a transfer zone The main purpose of con-ducting this experiment was to understand the dif-ference in the geometry of the pull-apart basinwhenno connecting bend exists in the basement faultThe faults in the transfer zone develop with on-going strike-slip deformation independent of a con-nected basement fault (Figures 8 9) Our experi-ments show that a transverse offset results in theoptimum configuration for the development of a

Mitra and Paul 1157

transfer zone of connecting faults Experiments in-volving overlapping or approaching patterns do notresult in well-defined pull-apart basins because theyboth involve a large transfer zone area This suggeststhat for two approaching strike-slip fault segmentsconnecting faults in the transfer zone will mostlikely start to develop when the end points of thetwo strike-slip segments are directly opposite eachother in a transverse offset position

1158 Structural Geometry and Evolution of Releasing and Rest

Initial deformation results in the formation ofa large number of short en echelon normal faults inthe area of the transfer zone (Figures 8A B 9A B)The central parts of the faults trend at angles of 53to 63deg to themain strike-slip fault zone At the sametime a series of en echelon R and Rprime Riedel shearsform above the main basement faults at angles of18 and 70deg with respect to the basement strike-slip faults respectivelyWith increasing strike-slip

Figure 8 Clay models showing faults in the sedimentary cover for a releasing offset with a transverse (90deg) offset between two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution ofthe structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

raining Bends

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

Figure 7 Three-dimensional geometry of the top of the sedimentary cover for a releasing bend with a transverse (90deg) bend con-necting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slip displacement

1156 Structural Geometry and Evolution of Releasing and Restraining Bends

The normal faults in the central part of thebasin curve into the orientations of the R shearsin the vicinity of the basement strike-slip faults(Figures 4B C 5B C) The resulting fault systemshave a sigmoidal geometry Rose diagrams of faultorientations show the progressive developmentof scatter because of the development of curvedfaults with varying components of normal andstrike-slip motion (Figure 4AndashC) Twomajor strike-slip faults develop at the surface directly above thepreexisting cuts in the basement after about 08 cmof displacement

Continuing deformation results in additional setsof major normal faults inward and approximatelyparallel to the original pair of fault systems Addi-tional fault systems with the same overall sigmoidalgeometry form in other parts of the transfer zoneand outside the central depression (Figures 4BndashE5BndashE) In the final stages of deformation the lon-gest pair of oblique faults connects with the twomain systems of strike-slip faults thus defining theboundary of the pull-apart basin (Figures 4F 5F)The final rose diagram shows significant scatter oftrends ranging from north to west because of thevariation from strike-slip to normal fault trends witha maximum density in an approximately north-west orientation (Figure 4F)

Examination of the contour patterns (Figure 5)suggests that the structural elevation steps downquite steeply across the normal faults bounding thepull-apart basin but drops more gradually from thebounding system of strike-slip faults to the centerof the basin The central part consists of two sep-arate depressions separated by a local high thatdevelop during the late stages (Figure 5E F) Theaxis connecting the centers of the two depressionsis approximately parallel to the trend of the mainstrike-slip faults Cross sections through the twodepressions show a flip in polarity because of achange in fault displacement along trend withinthe basin (Figure 5G H)

Transverse (90deg) BendFor a transverse bend (90deg) the normal fault seg-ments in the transfer zone form at a higher angle(58ndash78deg) than in the case of the oblique jog (45deg)and show left- and right-stepping patterns on op-

posite sides of the transfer zone (Figures 6A B 7AB) Immediately above the basement strike-slipfaults R and Rprime Riedel shears develop as a series ofen echelon faults making angles of 15 and 82deg to themain faults respectively The normal faults in thecenter of the pull-apart basin curve sharply clock-wise toward the orientation of the R shears in thevicinity of the main strike-slip fault zones along theboundaries of the pull-apart basin (Figures 6C D7C D)With continuing shear the major strike-slipfaults break through to the surface as Y shears de-fining the boundaries of pull-apart basin (Figures 6EF 7E F) As in the case of the oblique-bend ex-periment multiple sets of faults form in the pull-apart basin especially close to the main strike-slipfault zone Theorientations of the faults showawidevariation in trends ranging from northndashnorthwest towestndashnorthwest with the most dominant trend inthe center of the basin being about northwest in thefinal stage (Figure 6F)

Because of the sharper bends and changes inthe trends of the faults in the transfer zone the pull-apart basin associated with a transverse bend has amore S-shaped or rhomboidal geometry than thespindle-shaped geometry in the case of an obliquebend The contours show that the deepest part ofthe basin is located toward one of the major transferfaults so that the pull-apart basin has a more asym-metric shape with a sharper drop of contours onone side (Figure 7E F) The sides of the basin aremarked by a more gradual change in elevation withthe structure contours cut by secondary faults Dif-ferent basin shapes may result depending upon theslip distribution among the different faults

Transverse (90deg) OffsetThis setup has two cuts in the basement that even-tually form the bounding strike-slip faults offsetalong a transfer zone The main purpose of con-ducting this experiment was to understand the dif-ference in the geometry of the pull-apart basinwhenno connecting bend exists in the basement faultThe faults in the transfer zone develop with on-going strike-slip deformation independent of a con-nected basement fault (Figures 8 9) Our experi-ments show that a transverse offset results in theoptimum configuration for the development of a

Mitra and Paul 1157

transfer zone of connecting faults Experiments in-volving overlapping or approaching patterns do notresult in well-defined pull-apart basins because theyboth involve a large transfer zone area This suggeststhat for two approaching strike-slip fault segmentsconnecting faults in the transfer zone will mostlikely start to develop when the end points of thetwo strike-slip segments are directly opposite eachother in a transverse offset position

1158 Structural Geometry and Evolution of Releasing and Rest

Initial deformation results in the formation ofa large number of short en echelon normal faults inthe area of the transfer zone (Figures 8A B 9A B)The central parts of the faults trend at angles of 53to 63deg to themain strike-slip fault zone At the sametime a series of en echelon R and Rprime Riedel shearsform above the main basement faults at angles of18 and 70deg with respect to the basement strike-slip faults respectivelyWith increasing strike-slip

Figure 8 Clay models showing faults in the sedimentary cover for a releasing offset with a transverse (90deg) offset between two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution ofthe structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

raining Bends

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

The normal faults in the central part of thebasin curve into the orientations of the R shearsin the vicinity of the basement strike-slip faults(Figures 4B C 5B C) The resulting fault systemshave a sigmoidal geometry Rose diagrams of faultorientations show the progressive developmentof scatter because of the development of curvedfaults with varying components of normal andstrike-slip motion (Figure 4AndashC) Twomajor strike-slip faults develop at the surface directly above thepreexisting cuts in the basement after about 08 cmof displacement

Continuing deformation results in additional setsof major normal faults inward and approximatelyparallel to the original pair of fault systems Addi-tional fault systems with the same overall sigmoidalgeometry form in other parts of the transfer zoneand outside the central depression (Figures 4BndashE5BndashE) In the final stages of deformation the lon-gest pair of oblique faults connects with the twomain systems of strike-slip faults thus defining theboundary of the pull-apart basin (Figures 4F 5F)The final rose diagram shows significant scatter oftrends ranging from north to west because of thevariation from strike-slip to normal fault trends witha maximum density in an approximately north-west orientation (Figure 4F)

Examination of the contour patterns (Figure 5)suggests that the structural elevation steps downquite steeply across the normal faults bounding thepull-apart basin but drops more gradually from thebounding system of strike-slip faults to the centerof the basin The central part consists of two sep-arate depressions separated by a local high thatdevelop during the late stages (Figure 5E F) Theaxis connecting the centers of the two depressionsis approximately parallel to the trend of the mainstrike-slip faults Cross sections through the twodepressions show a flip in polarity because of achange in fault displacement along trend withinthe basin (Figure 5G H)

Transverse (90deg) BendFor a transverse bend (90deg) the normal fault seg-ments in the transfer zone form at a higher angle(58ndash78deg) than in the case of the oblique jog (45deg)and show left- and right-stepping patterns on op-

posite sides of the transfer zone (Figures 6A B 7AB) Immediately above the basement strike-slipfaults R and Rprime Riedel shears develop as a series ofen echelon faults making angles of 15 and 82deg to themain faults respectively The normal faults in thecenter of the pull-apart basin curve sharply clock-wise toward the orientation of the R shears in thevicinity of the main strike-slip fault zones along theboundaries of the pull-apart basin (Figures 6C D7C D)With continuing shear the major strike-slipfaults break through to the surface as Y shears de-fining the boundaries of pull-apart basin (Figures 6EF 7E F) As in the case of the oblique-bend ex-periment multiple sets of faults form in the pull-apart basin especially close to the main strike-slipfault zone Theorientations of the faults showawidevariation in trends ranging from northndashnorthwest towestndashnorthwest with the most dominant trend inthe center of the basin being about northwest in thefinal stage (Figure 6F)

Because of the sharper bends and changes inthe trends of the faults in the transfer zone the pull-apart basin associated with a transverse bend has amore S-shaped or rhomboidal geometry than thespindle-shaped geometry in the case of an obliquebend The contours show that the deepest part ofthe basin is located toward one of the major transferfaults so that the pull-apart basin has a more asym-metric shape with a sharper drop of contours onone side (Figure 7E F) The sides of the basin aremarked by a more gradual change in elevation withthe structure contours cut by secondary faults Dif-ferent basin shapes may result depending upon theslip distribution among the different faults

Transverse (90deg) OffsetThis setup has two cuts in the basement that even-tually form the bounding strike-slip faults offsetalong a transfer zone The main purpose of con-ducting this experiment was to understand the dif-ference in the geometry of the pull-apart basinwhenno connecting bend exists in the basement faultThe faults in the transfer zone develop with on-going strike-slip deformation independent of a con-nected basement fault (Figures 8 9) Our experi-ments show that a transverse offset results in theoptimum configuration for the development of a

Mitra and Paul 1157

transfer zone of connecting faults Experiments in-volving overlapping or approaching patterns do notresult in well-defined pull-apart basins because theyboth involve a large transfer zone area This suggeststhat for two approaching strike-slip fault segmentsconnecting faults in the transfer zone will mostlikely start to develop when the end points of thetwo strike-slip segments are directly opposite eachother in a transverse offset position

1158 Structural Geometry and Evolution of Releasing and Rest

Initial deformation results in the formation ofa large number of short en echelon normal faults inthe area of the transfer zone (Figures 8A B 9A B)The central parts of the faults trend at angles of 53to 63deg to themain strike-slip fault zone At the sametime a series of en echelon R and Rprime Riedel shearsform above the main basement faults at angles of18 and 70deg with respect to the basement strike-slip faults respectivelyWith increasing strike-slip

Figure 8 Clay models showing faults in the sedimentary cover for a releasing offset with a transverse (90deg) offset between two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution ofthe structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

raining Bends

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

transfer zone of connecting faults Experiments in-volving overlapping or approaching patterns do notresult in well-defined pull-apart basins because theyboth involve a large transfer zone area This suggeststhat for two approaching strike-slip fault segmentsconnecting faults in the transfer zone will mostlikely start to develop when the end points of thetwo strike-slip segments are directly opposite eachother in a transverse offset position

1158 Structural Geometry and Evolution of Releasing and Rest

Initial deformation results in the formation ofa large number of short en echelon normal faults inthe area of the transfer zone (Figures 8A B 9A B)The central parts of the faults trend at angles of 53to 63deg to themain strike-slip fault zone At the sametime a series of en echelon R and Rprime Riedel shearsform above the main basement faults at angles of18 and 70deg with respect to the basement strike-slip faults respectivelyWith increasing strike-slip

Figure 8 Clay models showing faults in the sedimentary cover for a releasing offset with a transverse (90deg) offset between two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is left lateral Panels AndashF show the progressive evolution ofthe structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments as apercentage of the total length of faults

raining Bends

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

Figure 9 Three-dimensional geometry of the top of the sedimentary cover for a releasing offset with a transverse (90deg) offset betweentwo strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasing strike-slipdisplacement

Mitra and Paul 1159

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

1160 Structural Geometry and Evolution of Releasing and Restraining Bends

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

deformation some of the normal faults propagatelaterally and connect with the R shears to formsigmoidal fault systems (Figures 8CndashF 9CndashF)

The final geometry is defined by a larger num-ber of small faults in the transfer zone comparedwith the experiments involving bends (Figures 8F9F) The reason for this difference is that the base-ment deformation is distributed over a wider zonethan along a single connecting bend The horizoncontours are discontinuous and terminate along thefaults The faults show a complete range of distri-butions from northndashnorthwest to westndashnorthwest(Figure 8F) The deepest part of the basin is formedin the northeast corner of the pull-apart basin Againwe attribute this to the distribution of slip amongthe faults instead of the experimental configuration

Comparison of Pull-Apart Geometries for DifferentExperimental ConfigurationsPanels A toC of Figure 10 show the final geometriesof the oblique bend transverse bend and transverseoffset configurations with approximately the sameamount of strike-slip deformation Cross sectionsthrough the basin are shown in Figure 10D to FThe main differences in the geometries of the struc-tures and the faults are that (1) an oblique bendresults in a spindle-shaped basin whereas a trans-verse bend or offset result in more S-shaped orrhomboid-shaped basins (2) the bend configura-tion results in a narrower and deeper basin withfewer normal faults that accommodate the exten-sion whereas the offset configuration results in awider and shallower basin with a transfer zone con-sisting of a large number of faults (3) the connec-tivity of faults is defined by fewer longer faults forthe connected setup whereas the offset configura-tion results in amore diffuse connected zone definedby a large number of shorter faults (Figure 10GndashI)

and (4) the deepest part of the basin is located indifferent locations in various experiments depend-ing on the symmetry of the basin and the distribu-tion of slip among the different faults In all cases itwas found that the central parts of the faults wentthrough minimum rotation whereas their marginswere rotated considerably into the strike-slip zoneswith ongoing deformation especially in the trans-verse bend and offset experiments

Comparison with Previous Models of Releasing BendsSeveral sand and clay models of releasing bendshave been published in the past (Dooley andMcClay1997 Rahe et al 1998 Basile and Brun1999 Atmaoui et al 2006 Wu et al 2009) Thecurrent models have several similarities and somekey differenceswith pastmodels in the experimentalsetup andboundary conditions analysis and resultsThe results of the models should be considered ascomplementary to those in previous studies

Most the past experiments used sand to rep-resent the sedimentary units and rigid sliding platesto represent the basement The basement was there-fore modeled as perfectly rigid and unable to de-formOur experimentsuse a two-layermodel drivenby basal sliding plates in which the basement is rep-resented by stiff clay whereas the sedimentary coveris represented by soft clay

This configuration enables us to model bothbends and offsets for pull-apart basins and comparethe differences Furthermore it allows for base-ment deformation in the vicinity of the faults

Several sand models (Dooley and McClay1997 Basile and Brun 1999Wu et al 2009) usedfills to study syngrowth sediments Serial sectionswere used byDooley andMcClay (1997)Wuet al(2009) to study cross sectional fault geometriesThe focus of our studies was the structural relief of

Figure 10 Comparison of final geometries and fault patterns for three releasing bend and offset experiments with approximatelythe same amount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault(C) Transverse (90deg) offset in basement fault Note the smaller angle between normal faults and main strike-slip faults leading to aspindle shape in (A) compared with the higher angle leading to an S or rhomboid shape in (B) and (C) Also the extension is distributedamong fewer major faults and a smaller area in (A) and (B) compared with (C) (DndashF) Cross sections through the top of the clayshowing the final basin geometry for the three experiments Locations of the cross sections are shown in (A) to (C) (GndashI) Faultconnectivity in the three model configurations Note that in (G) and (H) the connectivity is controlled by fewer large basin-boundingfaults whereas in (I) the connectivity is distributed over a larger number of smaller faults

Mitra and Paul 1161

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

1162StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

Figu

re11

Examples

ofreleasingbendsandoffse

ts(A)Right-steppingreleasingoffse

tbetweentheImperialand

Braw

leyfaultsintheImperialValleyCalifornia(redrawnfro

mJohnsonandHadley1976used

with

perm

issionfro

mtheSeism

ologicalSocietyofAm

erica)Themajor

faultsareright-lateralstrike-slip

faults

TheMesquite

Basin

marks

thedeepest

partofthebasin

The

area

was

thesiteofthe1975

Braw

leyearthquakeInsetshow

slocationof(A)a

nd(B)IF=ImperialfaultSAF=SanAn

dreasfaultSJF=SanJacintofault

(B)Transfe

rzone

betweentheClarkandCoyoteCreekright-lateralstrike-slip

faultsintheSanJacin

tofaultzone(m

odified

from

Sharp19671975used

with

perm

issionfro

mthe

CaliforniaGeologicalSurvey)The

area

ofoverlapbetweenthetwofaultsismarkedby

aserieso

fnormalfaultsthattrend

atahigh

angletothestrike-slipfaults

(C)R

ight-steppingoffse

tintherightlateralH

opefaultSouthIslandNe

wZealand(m

odified

fromClayton1966)Theoffse

tresultsintheform

ationofapull-apartbasinw

hose

deepestpointismarkedby

aglaciallake(D)M

apoftheVienna

Basin

show

ingmajorfaults(from

Royden1985used

with

perm

issionfro

mSEPM

)Strike-slipfaultingisdistributed

alonganumberofleft-lateralfaults

thatareinterpretedto

lose

theirstrike-slipcomponent

andbranch

into

aseriesof

enechelonnorm

alfaultsinthebasin

units and their relationship to faults as representedby structure contour maps to enable a direct com-parison with natural structures

For releasing bends the detailed fault geome-tries and their kinematic evolution are somewhatdifferent from those in the previous studies DooleyandMcClay (1997) documented two primary Riedelshears that developed increasing normal separationwith ongoing deformation and eventually curvedinto the primary strike-slip orientation to definethe boundaries of the pull-apart basins Late-stagestrike-slip faults parallel to the bounding strike-slipfaults (Y shears) oblique strike-slip faults travers-ing the pull-apart basins and a few secondary faultsof other orientations were also documented Sim-ilar results were documented by other authorsAtmaoui et al (2006) and Basile and Brun (1999)also documented the formation of some Rprime shearswith progressive deformation Our results show theseparate formation of R Riedel shears in the vicinityof the basement faults the simultaneous forma-tion ofmultiple sets of oblique normal faults in thecentral part of the basin and the eventual con-junction of these fault sets to result in the distinctS-shaped geometries of pull-apart basins They alsoshow the formation of Rprime shears at high angles tothe faults Y shears and oblique strike-slip faults inthe late stages of deformationOverall the numberof fault sets and density of faults documented aregenerally higher than in the sand experiments Therelative densities of the faults and the progressiveevolution of the relative densities are also docu-mented in our studies

Dooley and McClay (1997) showed the transi-tion from spindle or S-shaped geometries in obliqueapproaching bends to more rhomboidal geome-tries in the transverse or oblique overlapping bendsand similar results are obtained in our study In ourexperiments this transition is accompanied by in-creasing curvatures in the faults within the basinresulting from the larger differences in the orienta-tions of the normal faults and R Riedel shears

The two-layer models in this study documentthe difference in geometries between releasingbends and offsets in the basement faults whichhave not been previously modeled The primarydifference is that the pull-apart basin consists of a

Mitra and Paul 1163

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

Figure 12 (A B) Details of the Bockfleiss faults in the Vienna Basin mapped on the basis of subsurface data (from Wessely 1988 reprinted with permission from AAPG) Faults showincreasing heave in lower horizons because of growth (compare A with B) Faults exhibit a typical en echelon and sigmoidal geometry and are separated by ramps with contours at highangles to the faults This pattern is typical of faults formed in pull-apart basins as seen in the experimental model for an oblique bend or offset in (C) (see inset for location of faults)

1164StructuralGeom

etryand

EvolutionofReleasing

andRestraining

Bends

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

larger density of faults In addition using deform-able stiff clay tomodel the basement different shapesof the resulting basins are modeled (Figure 10) Inparticular basins resulting from releasing offsetsresult in a smoother shape of the top of the base-ment as depicted in the map and cross section inFigure 10C and F because the basin shape is de-fined by a large number of low displacement faults

NATURAL EXAMPLES OF RELEASINGBENDS AND OFFSETS

Surface exposures of structures formed along re-leasing bends provide the best natural analogs forcomparison with the experimental analogs becausethe details of the fault patterns and orientations canbe studied in detail The experimental models andthe surface models can be used to interpret sub-surface examples of these structures for whichmore limited data are available The following well-documented surface examples provide a suite ofstructures that can be compared with the observa-tions in the experimental models

Imperial and Brawley Faults ImperialValley California

The Imperial and Brawley faults make up a right-stepping releasing bend or offset on a set of right-lateral faults in the ImperialValleyCalifornia (Johnsonand Hadley 1976) The transfer zone between thetwo faults is primarily marked by two well-definedsets of en echelon normal faults which curve towardeach other (Figure 11A) The faults originating atthe Brawley fault make a smaller angle to the mainfault (~11deg) whereas those originating at the Im-perial fault make a higher angle (closer to 52deg) Thearea was the site of the Brawley swarm of earth-quakes in 1975 The earthquakes originated on theBrawley fault and migrated both north and southon the fault and also occurred on additional faultsin the area of the transfer zone Comparison withthe experimental evolution suggests that the earth-quakes formed along newly developing fault seg-ments within the transfer zones The MesquiteBasin which marks the depocenter within the

pull-apart basin is located closer to the northwestboundary of the basin suggesting an asymmetricflexure within the basin The low density of fault-ing and the presence of short unconnected faultsegments suggest that this configuration representsan early stage of a releasing bend or offset betweenapproaching faults

San JacintondashCoyote Creek Faults

The San Jacinto strike-slip fault zone is character-ized by several en echelon strands (Sharp 19671975) One of the transfer zones between the SanJacinto (Clark) and Coyote Creek faults is well ex-posed in the area to the north of theCoyoteCanyon(Figure 11B) The transfer zone is marked by anoverlap between the two faults The area of overlapcontains several northeast-trending and northwest-dipping normal faults These faults although show-ing some strike curvature at their boundaries showan almost normal trend relative to the boundingstrike-slip faults The dips of the faults shallowwithdepth (Sharp 1975) Sharp (1967) has suggestedthat the total crustal extension has not exceededabout 25 km (16 mi) which he postulates to bethe net right-lateral slip on the northern part ofCoyote Creek fault Our experimental models sug-gest that the geometries in the San JacintondashCoyoteCreek transfer zone are related to the fact that lat-eral and overlapping transfer zones are marked byincreasing angles of the connecting normal faults tothe bounding strike-slip faults

Hope Fault Glen Wye New Zealand

The Hope fault is the southern element of theMarlborough fault system a major right-lateralstrike-slip fault located on South Island New Zea-land (Clayton 1966 Cowan 1990) In the vicinityof the Glen Wye area offset Pleistocene glacial mo-raines and river terraces indicate predominant right-lateral motion (Figure 11C) The area is also markedby a major releasing offset toward the south Thenorthern boundary of the pull-apart basin shows asingle fault whereas the southern boundary is made

Mitra and Paul 1165

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

up of several short discontinuous segments withopposite dips as seen in the early stages of the ex-perimental models Both of these fault systems rep-resent the parts of the strike-slip fault where thestrike-slip component diminishes significantly anda normal component dominates

The termination of the northern strike-slip faultis marked by a zone of curved southndashsouthwestndashdipping en echelon normal faults (A) which stepdown toward a glacial lake in the center of the pull-apart basin These faults have an average trend ofapproximately 24deg to the main fault at their termi-nation within the basin The southern fault segmentis marked by two zones of normal faults B and Cwhich curve asymptotically into the main faultzone Fault zone B consists of shorter fault seg-ments with an average trend of 39deg to the mainfault and terminates at the basin margin whereasfault zone C consists of longer faults with an av-erage trend of 54deg to the main fault and marks thesouthwestern extremity of the pull-apart basin

The multiple zones of faults distributed overa wide area and the discontinuous nature of thesefaults suggest that the basin formed along a right-stepping offset similar to those in Figures 8 and 9

Vienna Basin Austria and Czech Republic

The Vienna Basin is a large pull-apart basin super-imposed on the north-vergent nappes of the outerWest Carpathian flysch belt and partly on the InnerCarpathian belt (Royden1985) The basin containsgas accumulations located in structural-stratigraphictraps within horst blocks in the Sarmatian and Pan-nonian units It is one area where structural trapswithin the pull-apart basin can be directly studiedTheMatzen field is the largestmultipool fieldwithinthe Vienna Basin (Kreutzer 1992) The field pro-duces oil and gas from multiple sand and carbonatereservoir units in a complexly faulted structure

The detailed structure and evolution of thebasin has been documented in detail by Royden(1985) During and after the Badenian synsedi-mentary growth faults with large throws formed(Wessely 1988) The basin formed between twoleft-stepping segments of left-lateral strike-slip

1166 Structural Geometry and Evolution of Releasing and Rest

faults The main strike-slip fault segments sepa-rated thrust faulting in the area east of the basin fromalready thrust-faulted areas to the west Royden(1985) has proposed that the strike-slip faults andrelated normal faults detach within a southeast dip-ping detachment at depth so that extension andstrike slip is restricted mainly to shallow crustallevels whereas Wessely (1988) has suggested thatthe extension may extend into the autochthonousunits underlying the allochthonous thrust sheetsEvidence to compare and evaluate these alternativemodels is not apparent in the surface data throughthe basin

The detailed structure of the basin is depictedin Figure 11D The basin is bounded by two mainstrike-slip faults both of which lose slip as they ap-proach the basin and eventually curve into the pull-apart basin A third strike-slip fault is located in thecentral part of the basin Royden (1985) has sug-gested that the main active segments of the faultsmay have changed somewhat over time but that thetotal slip was distributed along these major faultsand a large number of smaller faults Regionally thearea of the offset is marked by a large number ofen echelon normal fault segments These fault seg-ments curve asymptotically as they approach themain strike-slip segments

The details of the normal faults have been stud-ied by Wessely (1988) based on extensive well datain the basin Several phases of faulting occurred inthe basin The lateMiocene (Badenian and younger)faults which formed after thrusting and duringthe basin evolution typically maintain a constantdip of 40 to 50deg (Wessely 1988) Among thesethe Bockfleiss faults located in the western part ofthe basin show typical en echelon geometry andsigmoidal shapes and progressively drop each unitdown into the deeper part of the basin (Figure 12AB) Because the faults are synsedimentary growthfaults they also show larger throws and heaves indeeper units At the transfer zones between thefaults contours are typically closely spaced and ata high angle to the fault trends as also seen in theexperimental models (Figure 12c) The multiplenormal faults suggest a complex faulted basin formedalong one or more transverse or approaching off-sets between two or more strike-slip faults

raining Bends

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

RESTRAINING BENDS

Experimental Approach

The experimental configuration used for restrain-ing bends is similar to that for releasing bends ex-cept that convergence between the metal plates isused to create right-lateral shear (Figure 13) Theoverlying stiff and soft clay layers are deformedduring the convergence between the metal platesSimilar rates of movement are used as for the re-leasing bends Vertical cuts in the stiff layer aremade directly above the boundaries between themetal plates to represent the initial positions of thebasement strike-slip faults Only bend configura-tions yield the optimum configurations for trans-pressive structures to form in the transfer zone

Therefore only bend configurations are investi-gated in the final configurations Oblique (45deg)transverse (90deg) and oblique (135deg) bend config-urations are all modeled (Figure 13)

The fault geometries are traced directly fromphotographs taken during successive stages of theexperiment In addition laser scans are used tomodel the structures and fault geometries on thetop clay surface

Experimental Results

Oblique (45deg) BendRight lateral slip on a 45deg restraining bend resultsin the formation of an elongate fold at an angle ofapproximately 36deg to the main strike-slip faults(Figures 14A B 15A B) All measurements for

Figure 13 Experimental mod-els conducted for restrainingbends (A B) Oblique 45deg bend inbasement fault (C D) Transverse90deg bend in basement fault (E F)Oblique overlapping 135deg bend inbasement fault For each experi-ment the first figure shows therelationship between the under-lying sliding plates and the sec-ond figure shows the two claylayers and the preexisting faultcuts in the basement

Mitra and Paul 1167

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

the restraining bends are made in a clockwise di-rection from the main strike-slip faults The struc-tural high is cut by several fractures or normalfaults which are oriented approximately normal tothe axis of the structure Immediately above themain strike-slip segments a series of en echelonfaults form at angles of 8 to 18deg to the basementstrike-slip faults These faults are interpreted as R

1168 Structural Geometry and Evolution of Releasing and Rest

Riedel shears emanating from the basement strike-slip faults With progressive deformation the foldbecomes more asymmetric verging to the south-west and develops some northeast-dipping re-verse faults parallel to the structural trend especiallyalong the steep southwest limb of the structure(Figures 14C D 15C D) In the final stages thestructure grows broader and attains its greatest

Figure 14 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique (45deg) bend connecting twostrike-slip fault segments in the basement Sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressive evolutionof the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5deg increments asa percentage of the total length of faults

raining Bends

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

Figure 15 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique (45deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1169

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

structural relief It also shows a small clockwiserotation with the central part trending approxi-mately 26deg to the main strike-slip faults in the finalstage (205 cm of displacement) Linear strike-slipfaults parallel to themain faults (Y shears) also breakthrough the sedimentary cover and bound the struc-ture on two sides (Figures 14E F 15E F) Some

1170 Structural Geometry and Evolution of Releasing and Rest

curved faults at a high angle to the Y shears whichare possible Rprime shears form at the terminations ofthe main faults

The faults exhibit five dominant trends northnorth-northeast andwestndashtrending strike-slip faultsnorth-northwestndashtrending reverse faults and thenortheast-trending normal faults (Figure 14F)

Figure 16 Clay models showing faults in the sedimentary cover for a restraining bend with a transverse (90deg) bend connecting twostrike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show the progressiveevolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for different trends in 5degincrements as a percentage of the total length of faults

raining Bends

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

Figure 17 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with a transverse (90deg) bendconnecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with increasingstrike-slip displacement

Mitra and Paul 1171

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

Transverse (90deg) BendFor a transverse (90degbend) connecting the two strike-slip fault segments initial strike-slip results in theformation of a structural high whose structural axisis at a high angle (~57deg) to the basement strike-slip faults (Figures 16AndashC 17AndashC) A series of enechelon faults (Riedel R shears) forms immediatelyabove the basement strike-slip faults at an angle of

1172 Structural Geometry and Evolution of Releasing and Rest

10deg to the trend of the basement faults These faultsbound the structure to the west and east The nosesof the structures also develop a series of normalfaults trending approximately 65deg relative to theaxis of the structure

With increasing strike slip (Figures 16DndashF 17Dndash

F) the normal faults also propagate to the culmi-nation of the structure maintaining their oblique

Figure 18 Clay models showing faults in the sedimentary cover for a restraining bend with an oblique overlapping (135deg) bendconnecting two strike-slip fault segments in the basement The sense of slip on the strike-slip faults is right lateral Panels AndashF show theprogressive evolution of the structure with increasing strike-slip displacement Rose diagrams show the lengths of faults for differenttrends in 5deg increments as a percentage of the total length of faults

raining Bends

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

Figure 19 Three-dimensional geometry of the top of the sedimentary cover for a restraining bend with an oblique overlapping (135deg)bend connecting two strike-slip fault segments in the basement Panels AndashF show the progressive evolution of the structure with in-creasing strike-slip displacement

Mitra and Paul 1173

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

trend The structure is bounded by the strike-slipfaults on the western and eastern flanks and broad-ens with increasing strike-slip developing a rhom-boid shape The axis of the structure also curves intothe strike-slip faults at both ends In the very finalstages a pair of curved reverse faults bound thestructure on the northeast and southwest flanks(Figure 16F) The main strike-slip faults also breakthrough to the surface forming north-southndashtrendingY shears A few scattered north-northwestndashtrendingstrike-slip faults also develop The axis of the struc-ture in the central part rotates clockwise tomake anangle of 50deg with the main strike-slip faults in thefinal stage (212 cm [083 in] of displacement)

Four main trends of faults are present in the fi-nal stage a north-northeast trend for the en echelonstrike-slip faults (R shears) a north trend for the Yshears an approximate northeast trend for the nor-mal faults and a weak northwest trend for the re-verse faults (Figure 16F)

1174 Structural Geometry and Evolution of Releasing and Rest

Oblique Overlapping (135deg) BendAn oblique 135deg bend signifies a scenario where thetwo basement fault segments have significant over-lap and are connected along an oblique jog In thissetting the structure initially forms almost at rightangles to the main strike-slip faults We considerthis the least common of the three configurationsmodeled in the experiments because a connectingoblique jog with an acute angle or a transverse jogis likely to develop before the development of anoverlap of the two basement strike-slip faults

Initial deformation results in the formation of astructural high whose central part is almost trans-verse to the trend of the basement strike-slip faultsA series of en echelon strike-slip faults (R shears)form above the basement strike-slip faults at anangle of 5 to 10deg and curve away from the centralstructure (Figures 18AndashC 19AndashC) Continuing de-formation results in the formation of a set of faultstrending 75deg (Rprime shears) that curve into the structure

Figure 20 Comparison of final geometries and fault patterns for three restraining bend experiments with approximately the sameamount of strike-slip deformation (A) Oblique (45deg) bend in basement fault (B) Transverse (90deg) bend in basement fault (C) Over-lapping oblique (135deg) bend in basement fault Note the lower angle between axis of the structure and the main strike-slip faults leadingto a more elliptical shape in (A) compared with the higher angle leading to a rhomboid shape in (B) and rectangular shape in (C)

raining Bends

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

from the strike-slip faults Some normal faultstrending obliquely (57deg) to the axis of the struc-ture also form The ends of the structures alsodevelop narrow segments or tails that curve intothe trend of the strike-slip faults Some of thestrike-slip fault segments develop depressions anda significant component of normal separation Inthe late stages Y shears develop above the base-ment strike-slip faults and a few north-northwestndashtrending strike-slip faults traverse the structure(Figures 18DndashF 19DndashF)

The final fault pattern consists of a main trendof north-northeastndashtrending strike-slip faults (Rshears) east-northeastndashtrending Rprime shears north-trending Y shears and normal faults trending north-east (Figure 18F)

Comparison of Restraining Bend ModelsThe three restraining bend experimental modelsall resulted in the formation of a central high in thebend area two sets of en echelon strike-slip faults(R and Rprime shears) as well as Y shears immediatelyabove the basement strike-slip faults (Figure 20)The structural highs all developed narrow tails thatswung into the trend of the strike-slip faults Someextension was also observed along the strike-slipfault segments in the experiments The primarydifferences were that (1) the trends of the struc-tures changed from an acute trend to the strike-slip faults for the 45deg bend to an almost transversetrend for the 135deg bend (2) the normal faultsformed on the structure were almost transverse tothe structure for the 45deg bend but oblique in thecase of the transverse bend to almost parallel inthe case of the 135deg bend and (3) the shapes ofthe structures changed from more elliptical andspindle-shaped in the case of the 45deg bend to morerhomboidal or rectangular geometries with tails inthe case of the transverse and 135deg bends Overallthe incidence of faulting was significantly less thanfor the releasing bend experiments

Comparison of Results with Previous Models ofRestraining BendsMcClay and Bonora (2001) conducted a series ofsand experiments to model structures formed alongrestraining bends for oblique (30deg) transverse and

oblique overlapping (150deg) cases They studied thefault geometrieswith progressive evolution the finalcross sectional geometries and the geometries offaults within sand fills Our experiments investi-gate similar fault configurations but focus on thegeometries of the structures using structural con-tour maps and the relationships of the structures tothe progressively developing faults

InMcClay andBonorarsquos (2001) experiments theoblique (30deg) and transverse cases resulted in rhom-boidal pop-up structures whereas in the obliqueoverlapping (150deg) case the uplift was sigmoidalin shape In all three cases themaximum uplift wasin the center of the stepover zone Fault patternsincluded Riedel shears along the main strike-slipfaults and pairs of reverse faults with strike-slipcomponents bounding the structures Late-stagestrike-slip faults connected the twomain strike-slipfaults In our experiments the structures are asym-metric to symmetric and partially bounded by re-verse faults The axis of the structure forms pro-gressively higher angles in the transverse and obliqueoverlapping (135deg) cases The R Riedel shears arecommon above the main strike-slip faults with Rprimeand Y shears forming at later stages In all experi-ments a set of normal faults forms normal or obliqueto the axis of the structures

NATURAL EXAMPLES OFRESTRAINING BENDS

Ocotillo Badlands Coyote Creek Fault

The Ocotillo Badlands area between segments oftheCoyote Creek fault (Figure 21A) is an exampleof an uplift formed along a left-stepping bend oroffset along a right lateral strike-slip fault (Sharpand Clark 1972) A small overlap occurs betweenthe two strands of the fault (Figure 21A) The areaof the transfer zone is marked by a series of smallerscale anticlines and synclines which form an acuteangle averaging 37 to 40deg with the main strike-slipfaults The structures curve asymptotically into theeastern fault strand The area of greatest uplift is lo-cated in the northwestern part of the transfer zonebetween the two strike-slip faults

Mitra and Paul 1175

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

1176 Structural Geometry and Evolution of Releasing and Restraining Bends

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

The Borrego Mountain earthquake of 1968 re-sulted in rupturing on the preexistingCoyoteCreekfault and related faults The surface ruptures fol-lowed the preexisting structures and had the samesense of slip The fine lines in Figure 21A show thelocations and orientations of surface ruptures formedduring and after earthquake (USGS 1972) Thelocations of the surface ruptures suggest that de-formation occurred both along one of the mainstrike slips and the evolving structures in the re-straining bend The overall structure resembles theexperimental model formed along an oblique re-straining bend (Figures 15 16)

Chainat Ridge Thailand

The Chainat Ridge uplift is located on the MaePing fault zone Thailand (Smith et al 2007) Thestructure has also been referred to as the Chainatduplex (Morley et al 2007 Smith et al 2007)The structure is approximately 100 km (62 mi) inlength and formed on north-southndashtrending KhlongLhan restraining bend on a predominantly northwest-southndashtrending left-lateral fault (Figure 21B) Thearea of the restraining bend is marked by a series ofnorthndashsouth ridges trending about 40deg to the mainstrike-slip faults and exposing Paleozoic and Me-sozoic rocks and bounded by a series of faults Thestructure has been described as a gentle restrain-ing bend and resembles the experimental modelof a moderate-angle oblique jog on a strike-slipfault (Figures 15 16) The primary difference is thenumber of secondary faults and structures in theexample compared with the single large uplift inthe experimental model The wavelengths of foldsand related faults are typically proportional to thethickness of dominant units that are involved inthe deformation Therefore the absence of layer-ing in the experimental models induces the for-

mation of a single large-scale structure instead ofan assemblage of smaller scale structures

DISCUSSION AND CONCLUSIONS

Restraining and releasing bends and offsets can re-sult in a wide array of extensional or compressivestructures The experiments described in this arti-cle document the 3-D geometries of the surfacehorizon and evolution of some of the more com-mon configurations Based on the experimentsseveral conclusions can be made regarding the 3-Dgeometry of the structures The use of laser-scanmodels provides a useful tool for documenting thedetails of the structures in the transfer zones andtheir progressive evolution

Several suggestions have been made regardingthe development and subsequent evolution of thejogs in the main strike fault and related secondarystructures (Crowell 1974 Aydin and Nur 1985)Crowell (1974) suggested that the major strike-slip faults develop bends with ongoing slip andsubsequently develop a secondary branch in thearea of the bend resulting in the formation of afault wedge between the two faults Deformationwithin the wedge results in the formation of pull-apart basins or uplifts depending on the orienta-tion of the fault bend and the sense of strike slipOur approach has been to introduce fault bendsor offsets in the main basement faults and docu-ment the resulting patterns in the overlying sedi-mentary cover For releasing bends or offsets bothconnecting faults and offsets eventually result inthe formation of pull-apart basins For restrainingbends our experiments show that a connectingfault in the basement is necessary in all configura-tions for a well-defined structure to form in thesedimentary cover This suggests that fault offsets

Figure 21 Examples of restraining bends (A) Uplift formed along a restraining bend between two offset left-stepping segments of theCoyote Creek right-lateral fault in California (from Sharp and Clark 1972) The fine lines show surface ruptures formed along the faultsand related secondary faults during the Borrego Mountain earthquake in 1968 (B) The Chainat Ridge uplift formed along the pre-dominantly left-lateral Mae Ping fault zone in Thailand The area of the uplift is marked by a number of northndashsouth ridges exposingPaleozoic and Mesozoic rocks Dark areas represent Tertiary basins (from Smith et al 2007 reprinted with permission from TheGeological Society)

Mitra and Paul 1177

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

will be connected by an oblique or transverse jog atan early stage followed by the development of astructure

Mann et al (1983) andMann (2007) suggestedthat pull-apart basins initiate along releasing bendsin strike-slip faults where the sense of slip is obliqueto the trend of the segment They further suggestedthat with progressive slip the shapes of the pull-apartbasins change from a spindle shape through a lazyZ or S shape to a rhomboid shape Our experi-ments suggest that although the aspect ratios ofthe basins are controlled by the amount of ex-tension the shapes are also dependent on the rel-ative positions of the strike-slip fault terminationsOblique (45deg) bends result in spindle shapes with agentle curvature of normal faults into the strike-slipfaults whereas transverse or oblique overlapping(135deg) bends or offsets result in more S or rhom-boid shapes In addition the presence of preex-isting bends or the early development of the bendsin the basement fault result in narrower and deeperpull-apart basins with fewer basin-bounding faultswhereas the distribution of deformation along anoffset results in awider and deeper basinmarked bya large number of smaller faults

For restraining bends and offsets Mann (2007)proposed that the structures evolve from gentle tosharp shapes with progressive deformation Ourexperimental models show that the geometries ofthe structures are more dependent on the geom-etry of the connecting basement faults (oblique 45degtransverse or oblique overlapping 135deg) and thatthese geometries also control the orientations of themajor uplift Oblique (45deg) bends result in moreelliptical or spindle-shaped structures whereas trans-verse or oblique overlapping (135deg) bends result inrhomboidal or rectangular geometries The anglebetween the axis of the structure and the boundingstrike-slip fault segments increases progressivelyfrom approaching to transverse to overlapping off-sets A small amount of rotation of the structurescommonly occurs with progressive deformation

The nature of transfer of fault slip between thestrike faults and the dip-slip faults in the area ofthe offset or bend controls the nature of the struc-tures The originalmodels of sharp bends byCrowell(1974) involve strike-slip faults that have constant

1178 Structural Geometry and Evolution of Releasing and Rest

slip until they terminate against cross faults with anabrupt transformation from strike-slip to dip-slipfaulting Although this is one end-member possibil-ity our models and many natural examples suggestthat a more common observation is the progressivereduction in the slip on the strike-slip faults as theyapproach the bends or offsets and the progressivetransfer of slip to en echelon normal faults in the caseof releasing bends (see also Royden 1985) and up-lifts in the case of the restraining bends In fact thestrike-slip faults in the sedimentary cover typicallybreak into a series of en echelon faults with verysmall slip in the vicinity of the bends or offsets Typ-ically the bounding systems of strike-slip faults con-nect to only one boundary of the structure in thebend area so that a well-connected systemof faultsbetween all four sides of the structure is very rareThe structures within the transfer zone consist pri-marily of normal faults with en echelon geometriesin the case of releasing bends and offsets and obliqueuplifts cut by reverse faults in the case of restrainingbends Secondary normal faults also form transverseor oblique to the trends of the structures

The results of the experimental models pro-vide details regarding the faults and related struc-tures along restraining bends and releasing bendsand offsets that can be used for interpreting bothsurface and subsurface structures Most releasingand restraining bends are characterized by limiteddata therefore key relationships between the base-ment fault geometries the net displacement on thefaults and the resulting geometries and orientationof folds and secondary faults can be used to improveinterpretations of these structures Characteristicdifferences between structural geometries of pull-apart basins and folds along restraining bends fordifferent basement fault configurations based onthe experimental models can be directly applied tosubsurface structural interpretations

Note however that a superficial resemblancebetween the geometries of natural structures andanalog models is not a sufficient criterion for theassumption that a structure is related to a restrain-ing or releasing bend Branching and rejoining ge-ometries of normal and reverse faults can producesimilar surface patterns of fault traces as thosedocumented for strike-slip structures and oblique

raining Bends

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

structures bounded by such faults can be easilymistaken for strike-slip structures The minimumcriteria for interpreting the structures as strike-slipstructures are that the bounding faults have a sig-nificant strike-slip component and that the struc-tures bounded by the faults formed at the sametime as the faults and are not oblique structures cutby later strike-slip faults

REFERENCES CITED

Atmaoui N N Kukowski B Stoumlckhert and D Koumlnig 2006Initiation and development of pull-apart basins with Riedelshear mechanism Insights from scaled clay experimentsInternational Journal of Earth Sciences v 95 no 2p 225ndash238 doi101007s00531-005-0030-1

Aydin A and A Nur 1985 The types and role of stepoversin strike-slip tectonics in K T Biddle and N Christie-Blick eds Strike-slip deformation basin formationand sedimentation SEPM Special Publication 37 p 35ndash44

Basile C and J-P Brun 1999 Transtensional faulting pat-terns ranging from pull-apart basins to transform conti-nental margins An experimental investigation Journalof Structural Geology v 21 p 23ndash37 doi101016S0191-8141(98)00094-7

Biddle K T and N Christie-Blick eds 1985 Strike-slipdeformation basin formation and sedimentation SEPMSpecial Publication 37 386 p

Bose S and S Mitra 2010 Analog modeling of divergentand convergent transfer zones in listric normal fault sys-tems AAPG Bulletin v 94 p 1425ndash1452

Clayton L 1966 Tectonic depressions along the Hope faulta transcurrent fault in north Canterbury New ZealandNew Zealand Journal of Geology and Geophysics v 9no 1ndash2 p 95ndash104

Cowan H 1990 Late Quaternary displacements on theHope fault at Glynn Wye North Canterbury New Zea-land Journal of Geology and Geophysics v 33 p 285ndash293

Crowell J C 1974 Origin of late Cenozoic basins in south-ern California in W Dickinson ed Tectonics and sedi-mentation SEPM Special Publication 22 p 190ndash204

Cunningham W D and P Mann eds 2007 Tectonics ofstrike-slip restraining and releasing bends The Geologi-cal Society (London) Special Publication 290 482 p

Dooley T and K McClay 1997 Analog modeling of pull-apart basins AAPG Bulletin v 81 no 11 p 1804ndash1826

Harding T P 1985 Petroleum traps associated with wrenchfaults AAPG Bulletin v 58 p 1290ndash1304

Johnson C E and D M Hadley 1976 Tectonic implica-tions of the Brawley earthquake swarm Imperial valleyCalifornia January 1975 Bulletin of the SeismologicalSociety of America v 66 no 4 p 1133ndash1144

Kreutzer N 1992 Matzen field Vienna Basin Austria in

E A Beaumont and N H Foster comp Structuraltraps II AAPG Treatise of Petroleum Geology Atlas ofOil and Gas Fields p 57ndash98

Mann P 2007 Global catalog classification and tectonicorigins of restraining- and releasing bends on active andancient strike-slip fault systems in W D Cunninghamand P Mann eds Tectonics of strike-slip restraining andreleasing bends TheGeological Society (London) SpecialPublication 290 p 13ndash142

Mann P M R Hempton D C Bradley and K Burke1983 Development of pull-apart basins Journal ofGeology v 91 p 529ndash554 doi101086628803

McClay K andM Bonora 2001 Analog models of restrain-ing stepovers in strike-slip fault Systems AAPG Bulletinv 85 no 2 p 233ndash260

Morley C K M Smith A Carter P Charusiri and SChantraprasert 2007 Evolution of deformation stylesat a major restraining bend constraints from cooling his-tories Mae Ping fault zone western Thailand in W DCunningham and P Mann eds in Tectonics of strike-slip restraining and releasing bends The Geological So-ciety (London) Special Publication 290 p 325ndash349

Petrov M A Talapov T Robertson A Lebedev A Zhilyaevand L Polonskiy 1998 Optical 3D digitizers Bringinglife to the virtual world IEEE Computer Graphics andApplications v 18 p 28ndash37 doi10110938674969

Rahe B D A Ferrill and A P Morris 1998 Physical analogmodeling of pull-apart basin evolution Tectonophysicsv 285 no 1ndash2 p 21ndash40 doi101016S0040-1951(97)00193-5

Royden L H 1985 The Vienna Basin A thin-skinned pull-apart basin in K T Biddle and N Christie-Blick edsStrike-slip deformation basin formation sedimentationSEPM Special Publication 37 p 319ndash338

Serra S and R A Nelson 1988 Clay modeling of rift asym-metry and associated structures Tectonophysics v 153p 307ndash312 doi1010160040-1951(88)90023-6

Sharp R V 1967 San Jacinto fault zone in the Peninsularranges of southern California Geological Society ofAmerica Bulletin v 78 no 6 p 705ndash729 doi1011300016-7606(1967)78[705SJFZIT]20CO2

Sharp R V 1975 En echelon fault patterns of the SanJacinto fault zone California Division of Mines and Ge-ology special report 18 p 147ndash152

Sharp R V and M M Clark 1972 Geologic evidence ofprevious faulting near the 1968 rupture on the CoyoteCreek fault U S Geological Survey professional paperreport P 0787 p 131ndash140

Smith M S Chantraprasert C K Morley and I Cartwright2007 Structural geometry and timing of deformation inthe Chainat duplex Thailand in W D Cunninghamand P Mann eds Tectonics of strike-slip restrainingand releasing bends The Geological Society (London)Special Publication 290 p 305ndash323

Sylvester A G comp 1984 Wrench fault tectonicsAAPG Reprint Series 28 374 p

Tchalenko J S 1970 Similarities between shear zones ofdifferent magnitudes Geological Society of AmericaBulletin v 81 p 1625ndash1640 doi1011300016-7606(1970)81[1625SBSZOD]20CO2

Mitra and Paul 1179

Tchalenko J S and N N Ambraseys 1970 Structural anal-ysis of the Dasht-e Bayaz (Iran) earthquake fracturesGeological Society of America Bulletin v 81 p 41ndash60 doi1011300016-7606(1970)81[41SAOTDB]20CO2

US Geological Survey (USGS) 1972 The Borrego moun-tain earthquake of April 9 1968 Geological Survey Pro-fessional paper report 787 207 p

Wessely G 1988 Structure and development of the Vienna

1180 Structural Geometry and Evolution of Releasing and Rest

Basin in Austria in L H Royden and F Horvaacuteth edsThe Pannonian Basin A study in basin evolution AAPGMemoir 45 p 333ndash346

Wilcox R E T P Harding and D R Seely 1973 Basicwrench tectonics AAPG Bulletin v 57 p 74ndash96

Wu J E K McClay P Whitehouse and T Dooley 20094D analog modeling of transtensional pull-apart basinsMarine and Petroleum Geology v 26 no 8 p 1608ndash1623 doi101016jmarpetgeo200806007

raining Bends