Reprinted from Basin Research, 1991, v. 3, p. 123-141.

1

Half-graben basin filling models: New constraints on conti-nental extensional basin developmentRoy W. Schlische

Department of Geological SciencesRutgers The State University of New JerseyNew Brunswick, New Jersey 08903 U.S.A.

ABSTRACTThree end-member models of half-graben development (detachmentfault, domino-style, and fault growth) evolve differently through time andproduce different basin filling patterns. The detachment fault model in-corporates a basin-bounding fault that soles into a subhorizontal detach-ment fault; the change in the rate of increase in the volume of the basinduring uniform fault displacement is zero. Younger strata consistentlypinch out against older synrift strata rather than pre-rift rocks. Both ba-sin-bounding faults and the intervening fault blocks rotate during exten-sion in the domino fault block model; a consequence of this rotation is thatthe change in the rate of increase of the volume of the basin is negativeduring uniform extension. Basin fill commonly forms a fanning wedgeduring fluvial sedimentation, whereas lacustrine strata tend to pinch outagainst older synrift strata. In the fault growth models, basins grow bothwider and longer through time as the basin-bounding faults lengthen anddisplacement accumulates; the change in the rate of increase in basin vol-ume is positive. Fluvial strata progressively onlap pre-rift rocks of thehanging wall block, whereas lacustrine strata pinch out against older flu-vial strata at the center of the basin but onlap pre-rift rocks along the lat-eral edges. These fundamental differences may be useful in discriminatingamong the three end-member models. The transition from fluvial to la-custrine deposition and hanging wall onlap relationships observed in nu-merous continental extensional basins are best explained by the faultgrowth models.

INTRODUCTIONExtensional basins provide perhaps the mostimportant record of the history of the stretch-ing of the crust, for they contain structures tha twere active during sedimentation and stratathat provide the necessary chronologic con-straints for dating the structures. The geometryof the basins themselves furnishes valuableinformation on the mechanisms and amount ofextension. For example, roll-over of hangingwall strata into basin-bounding faults tradi-tionally has been used to infer the presence oflistric faults (Hamblin 1965) as well as to cal-culate the depth to detachment and the amountof horizontal extension (Verrall 1981). Fur-thermore, listric faults that sole into sub-horizontal detachments can accommodate muchgreater amounts of horizontal extension abovethe detachment than, for example, planar nor-mal faults or one generation of domino-style

faults (Wernicke & Burchfiel 1982). Unfortunately,by their very nature, extensional basins commonlyare buried beneath thick sequences of post-riftstrata or under meters of water. Thus, informationon the basins must be obtained by indirect meansand is therefore subject to a wide latitude of inter-pretation.

This paper examines three end-member types ofextensional basin development: (1) the detachmentfault model consists of a listric fault that soles intoa sub-horizontal detachment or kink-style planarfault/subhorizontal detachment complex; (2) dom-ino-style or tilted-fault-block model; and (3) faultgrowth models for planar normal faults resulting inextensional basins that grow both in length andwidth through time. Operating under some simpleassumptions related to the sedimentary infilling ofthe basins, each model of extensional basin devel-opment results in a diagnostic set of features ofstratal geometry (including onlap patterns) and

R.W. Schlische Half-Graben Filling Models

2

stratigraphic successions in continental (non-marine) extensional basins. These stratigraphicfeatures may therefore serve as an additionalconstraint in unraveling the geometry of a ba-sin, its evolutionary development, and themode and magnitude of extension. Given certainstratigraphic relationships commonly observedin numerous continental extensional basins ofvarious ages, the assumptions and general ap-plicability of the three end-member models arecritically evaluated.

EXTENSIONAL BASIN MODELSDetachment Fault ModelsDetachment fault models of extensional basindevelopment have two end-member geometries(Fig. 1A). The first involves listric normalfaults that gradually sole into sub-horizontaldetachments (Wernicke & Burchfiel 1982;Gibbs 1983). In the second case the fault systemhas a kinked geometry consisting of two planarfault segments (Jackson 1987; Groshong 1989).Under both end-member conditions, translationon the sub-horizontal detachment results in po-tential voids between the hanging wall andfootwall blocks, and collapse of the hangingwall results in the formation of a half-graben.The footwall block is assumed to remain pas-sive during extension (Gibbs 1983; Groshong1989). The geometry of the half-graben is gov-erned by (1) the rules of equal-area balancing(Gibbs 1983), (2) the geometry of the fault sys-tem, and (3) the nature of the deformation inthe hanging wall, i.e., collapse along zones ofvertical shear, collapse along antithetic faultsof variable dip angles, and the relativeamounts of bedding-plane shear within thehanging wall block (Gibbs 1983, 1984; White e tal. 1986; Williams & Vann 1987). In general,half-graben become wider and less deep as thedip angle of the antithetic faults along whichthe hanging wall collapses decreases (Crews &McGrew 1990). The dip angle of the border faultand the depth to detachment also strongly in-fluence the geometry of the basin: for the sameamount of net displacement on the horizontaldetachment, basins become narrower and deeperas the dip of the basin-bounding fault anddepth to detachment increase (Morley 1989). Inthe case of listric faults, a roll-over geometryresults in the hanging wall because of the in-creasing size of the potential void between thehanging wall and footwall blocks toward thelistric fault. For the ramp-flat geometry, aflat-bottomed half-graben results because thewidth of the potential void between the hang-

ing wall and footwall blocks is constant over a con-siderable portion of its length.

Given the equal-area balancing assumption, thecross-sectional area of the hanging wall basin isgiven by:A = hd (1)where h is the net displacement on the horizontaldetachment and d is the depth of the detachment.The rate of increase in the cross-sectional area ofthe basin is constant (dA/dh = d), and the changein the rate of area increase (d2A/dh2) is zero. This isa feature unique to the detachment fault models.The volume of the basin also changes similarly(Fig. 1D) since uniform plane strain conditions pre-vail (Gibbs 1983). The uniform plane-strain condi-tion is most likely to be satisfied when the basin isbounded laterally by vertical transfer faults (ter-minology of Gibbs 1984).

Domino-Style Fault Block ModelIn this model of extensional basin development,first described by Emmons & Garrey (1910) for theBasin and Range of the western United States, boththe faults and the blocks between the faults rotateabove a detachment horizon during extension. Theindividual faults may be planar or listric, and thedetachment need not be horizontal (Wernicke &Burchfiel 1982; Axen 1988). Under the simplest con-ditions (uniform fault spacing, horizontal detach-ment horizon) Thompson (1960) and Wernicke &Burchfiel (1982) presented a relationship amongthe amount of horizontal extension (β), the initialdip of the normal faults (ϕi), the subsequent dip ofthe faults (ϕ), the dip of horizons that were ini-tially horizontal (θ), the original horizontal com-ponent of fault spacing (F') and the subsequenthorizontal component of fault spacing (F) (Fig. 1B):

β = ( )FF'

= [ ]sin ( )ϕ + θsin ϕ = [ ]sin ϕ

i

sin ( )ϕi- θ

(2)The cross-sectional area of the triangular-shapedhalf-graben is given by:

A = 12

[ ]F' 2β sin θ (3)

Since the faults decrease in dip with continued ex-tension, each increment of cross-sectional area issmaller than the preceding one under conditions ofuniform extension, as can be verified by equation(3). The domino fault block models are assumed tobe uniform plane strain; again, the simplest geo-logical example is a basin bounded along-strike byvertical transfer faults. Given the uniform planestrain condition, the change in the rate of increasein basin volume during uniform extension must be

R.W. Schlische Half-Graben Filling Models

3

negative, i.e., the rate at which new basin vol-ume is added must decrease during uniformlyincreasing extension (Fig. 1D). Furthermore,because of the rotation of the faults, therecomes a point when additional slip on verylow-angle normal faults becomes impossible(Sibson 1985). A new generation of moresteeply-dipping domino-style normal faultsmay then form. Multiple generations of domino-style faults have been documented by Proffett(1977) and Miller et al. (1983) in the Basin andRange of Nevada. Further elaborations on thedomino-style model of extension are found inBarr (1987), Mandl (1987), Axen (1988), andSclater & Célérier (1989).

Fault Growth ModelsStudies of small normal faults in British coalfields and larger normal faults in the NorthSea imaged on a closely-spaced grid of seismiclines have shown that the displacement onthese faults is generally greatest at or near thecenter of the fault and decreases to zero at itsends (Barnett et al. 1987; Walsh & Watterson1987, 1988, 1989; Gibson et al. 1989). Lateralvariations in fault displacement were alsomeasured on a normal fault in the Kenya riftvalley (Fig. 2A; Chapman et al. 1978) and fol-lowing the 1959 Hebgen lake earthquake(Fraser et al. 1964). Displacement of sedimen-tary horizons by blind normal faults decreasesin a direction normal to the fault surface, bothin the hanging wall and the footwall, produc-ing a reverse-drag geometry (Barnett et a l .1987; Gibson et al. 1989). In the hanging wall,this reverse drag resembles “roll-over,” eventhough it is associated with planar normalfaults. The reverse drag geometry is also pro-duced for synsedimentary faults in which thenormal fault intersects the earth’s free surface(Gibson et al. 1989). In this situation, however,the absolute magnitude of hanging wall subsi-dence is greater than footwall uplift. The roll-over radius increases as the displacement in-creases. Changes in elevation following the1983 Borah Peak Earthquake in Idaho, showingboth hanging wall subsidence and a considera-bly smaller footwall uplift (Stein & Barrientos1985), are also consistent with transversevariations in displacement (Fig. 2B).

When fault length (L) is plotted againstmean displacement (Dm) on a log-log plot, lin-ear relationships are evident (Fig. 2C). Wat-terson (1986), Walsh & Watterson (1987, 1988),and Cowie & Scholz (1990) interpreted theserelationships in terms of fault growth: short

faults with small maximum displacements growinto longer faults with larger maximum displace-ments. Watterson (1986) and Walsh & Watterson(1987, 1988) grouped together all dip-slip faultsfrom a variety of tectonic settings and lithologiesand derived a universal scaling law ofL= cD1/2 (4)where D is the maximum fault displacement and cis some constant of proportionality, primarily de-pendent on rock properties. Cowie & Scholz (1990)argued that length-displacement data fall intodiscrete populations and are better interpreted interms of the scaling lawL = D/ε (5)where ε is the linear strain and is dependent on rocktype and tectonic environment. Models employingthe first scaling law subsequently will be referredto as fault growth model I; models using the secondscaling law will be known as fault growth model II.

Sedimentary basins bounded by normal faultsthat obey either fault growth model are expectedto grow in length and width through time (Fig. 1C).Schlische (1990b) noted that the early Mesozoichalf-graben basins of eastern North America ap-pear to be deepest adjacent to the center of the bor-der fault system and decrease in depth toward a l ledges of the basin. The thicknesses of fixed-periodlacustrine cycles within the Newark basin vary ina similar manner and may therefore act as a proxyof variations in fault displacement (Schlische1990a). The finite lengths of these basins or basincomplexes is also consistent with the notion of de-creasing displacement toward the ends of thefaults. In the Basin and Range, the elevations ofthe ranges [resulting from footwall uplift along therange-bounding normal faults (Wernicke & Axen1988)] may be a proxy of variations in fault dis-placement. The ranges are, of course, of finitelength and generally are higher at their centers,apparently in general agreement with the faultgrowth models.

Gibson et al. (1989) used relationships of faultgrowth model I to generate model half-graben. Spe-cifically, the along-strike dimension of the basin isgiven by the fault length L:

L = { }8SD[ ]7πG16( )∆σ

212

(6)where G is the shear modulus, ∆σ is the stress dropafter each seismic event, S is the increment bywhich slip increases after each slip event (neces-sary for the fault to grow and for the growth se-quence to match the observational data), and D isthe maximum displacement given by

R.W. Schlische Half-Graben Filling Models

4

D = 12

( )n ( )n - 1 ( )S (7)

where n is the nth slip event (Walsh & Wat-terson 1988). The total number of slip eventsmay be estimated by knowing or assuming thetotal time of active faulting and the averagerepeat time between slip events. The displace-ment d at any point on the fault surface is givenby

d = 2D( )1 - rR { }[ ]1

2 ( )1 + rR

2- ( )r

R2

12

(8)where R is radius of the fault (L/2) and r is thedistance on the fault from the center of thefault (after Gibson et al. 1989). Displacementnormal to the fault surface, d', is given by

d' = D{ }e[ ]-5.5( )r' /R' - 0.004( )r' /R' (9)where r' is the distance from the fault meas-ured normal to the fault and R' is the roll-overradius (after Gibson et al. 1989). The roll-overradius is given by

R' = R'max ( )dD

12

(10)where R'max is the maximum roll-over radius,taken by Gibson et al . (1989) to be equal to themean of the major and minor radii of the faultsurface ellipse. The ratio of the major to minorradii of the ellipse typically ranges from 1.25to 3 (Gibson et al. 1989).

The preceding equations assumed that thenormal faults were blind and consequently thedisplacement of horizons was distributedequally in the footwall and hanging wallblocks. For non-vertical synsedimentary faults(where the fault intersects the free surface ofthe earth), there is an asymmetry betweenhanging wall and footwall displacement of ho-rizons, with the asymmetry increasing as thefault dip decreases. The percent contribution ofhanging wall displacement is given byHW = 110 - (2ϕ/3) (11)where ϕ is the dip of the fault in degrees (Gib-son et al. 1989). Maximum displacement is as-sumed to occur at the free surface.

A necessary condition of the fault growthmodel I of basin evolution is that the majorityof the displacement on the basin bounding faultaccumulates during the final time increments ofextension because the amount of slip is increas-ing arithmetically from slip event to slip event[equation (7)]. Gibson et al. (1989) infilled their

model basins assuming that sedimentation alwayskept up with subsidence. Therefore, with this con-straint, the sedimentation rate must increasethrough time, and younger strata will always pro-gressively onlap pre-rift rocks of the hanging wallblock (Gibson et al. 1989). A corollary to this is tha tthe increment by which the cross-sectional area ofthe basin increases after each uniform increment oftime itself must increase.

Using the scaling law of fault growth model II ,maximum fault displacement is given byD = εL (12)Maximum displacement may also be represented byD = qn (13)where q is the displacement per slip event, and n isthe number of slip events. The duration of basinformation is given byT = tn (14)where T is the time since the onset of active fault-ing and t is the repeat time of slip events, which istaken to be constant. Equations 8-11 of fault growthmodel I are also used here to describe the displace-ment of horizons measured on the fault surface (d),the displacement of horizons measured normal tothe fault surface (d'), the roll-over radius (R'), andpercent contribution of the hanging wall to the to-tal displacement (HW).

In addition to the different scaling laws used todescribe the relationship between maximum dis-placement and fault length (equations 4 and 12),the main difference between the two models restswith the manner in which fault displacement ac-cumulates through time if the repeat time of slipevents is uniform. Fault growth model I requiresthat the maximum displacement per slip event in-crease with time. Consequently, the model basinsstart out shallow and exponentially deepenthrough time. In fault growth model II, the maxi-mum displacement per slip event remains un-changed through time, and the model basins line-arly deepen through time. In fault growth model I ,the amount of slip per slip event is proportional tothe length of the fault. This may be reasonable dur-ing the early stages of fault growth, when slip onsmall faults ruptures the entire fault, yet is insuffi-cient to relieve all of the accumulated stress; theexcess stress may then be used to fracture intactrock, lengthening the fault. However, at some time,when the fault has reached a critical length, therupture length and displacement per slip event mustbecome independent of the total length of thefault—the basis for the second fault growth model.Rupture events occurring near the ends of the faultwill still fracture unslipped rock, increasing thelength of the fault. Whichever model is operating,

R.W. Schlische Half-Graben Filling Models

5

the change in the rate of volume increase ispositive (Fig. 1D).

In all three models of extensional basin de-velopment discussed above, the width of thebasin (measured perpendicular to the basin-bounding fault) naturally increases during faultslip, and hence the volume of the basin con-stantly increases. Under conditions of uniformfault displacement, the change in the rate ofincrease of the basin volume is zero in the caseof the areally-balanced detachment fault mod-els. For the domino model, the change in therate of increase in basin volume is negative. Inthe two fault growth models, the change in therate of increase in basin volume is positive.Uniform plane strain conditions are assumed toprevail in both the domino and detachmentfault models, and therefore the length of thebasin must remain fixed throughout its devel-opment; that length may be considered the dis-tance between two adjacent vertical transferfaults. In contrast, the length of the basin in-creases in the fault growth model. Based on thestratigraphic constraints established below,these significant differences among the modelslead to the development of distinctive stratalgeometries and stratigraphic successions.

STRATIGRAPHIC CONSTRAINTSFOR NON-MARINE BASINS

As documented by Schlische & Olsen (1990)and Lambiase (1991), numerous continental (ter-restrial or non-marine) extensional basins (Ta-ble 1) display all or part of a tripartite strati-graphic sequence consisting of (1) a lowermost,predominantly fluvial unit; (2) a middle, pre-dominantly lacustrine interval, commonly con-taining deep-water facies; and (3) an upper unitconsisting of shallow-water lacustrine stratacommonly overlain by predominantly fluvialstrata. Many half-graben also display onlap ofyounger strata onto pre-rift rocks (Table 2; An-derson et al. 1983; Leeder & Gawthorpe 1987;Gibson et al. 1989; Schlische & Olsen 1990), in-dicating that the depositional surface area wasprogressively increasing through time as thebasin filled.

In non-marine extensional basins, there aretwo fundamentally different depositional set-tings that have tectonostratigraphic signifi-cance (Schlische & Olsen 1990). In the first casethe volume of sediment available to fill thebasin during a specific interval of time (knownsubsequently as the volumetric sedimentationrate) exceeds the capacity of the basin. Thebasin can therefore always fill to its lowestoutlet, with both excess sediment and water

leaving the basin through its lowest-elevation out-let. The predominant depositional environment con-sists of through-flowing streams, although otherfluvial systems will be present along the margins ofthe basin, and some minor, local ponding of watermay occur. Under this depositional regime, sedi-mentation rate (the thickness of sediment depos-ited during a given increment of time) is equal tothe prevailing subsidence rate (amount of verticalmotion during a given increment of time). This de-positional regime will subsequently be referred toas fluvial.

In the second case the volumetric sedimentationrate is less than the rate at which the basin createsnew space for sediment as a result of continued ex-tension and/or an increase in elevation of the outletof the basin (perhaps a consequence of damming bylava flows). The basin is therefore sedimentstarved: all sediment that enters the basin remainsin the basin. If the inflow rate of water is greaterthan the evaporation and groundwater dischargerates, then a lake can occupy the space between thedepositional surface and the lowest outlet. Depend-ing on the volume of water available, the basinmay be hydrologically open (some water flows out)or closed [all water that flows into the basin leavesonly by evaporation or seepage (Smoot 1985)]. La-custrine sedimentation predominates, althoughfluvial and deltaic sedimentation may be commonaround the margins of the lake, particularly duringlow lake levels driven by climatic fluctuations(e.g., Olsen 1986). Sedimentation rate always willbe less than the subsidence rate. This depositionalregime subsequently will be known as lacustrine.

Given that fluvial deposition, which requiresonly a slope and a supply of sediment, is the mostcommon depositional regime in a continental set-ting, it is not surprising that the basal deposits inmany rift basins are fluvial. As discussed above,lacustrine deposition requires special conditions. Ina half-graben basin, the fluvial-lacustrine transi-tion might occur as a consequence of decreasing thevolume of sediment entering the basin per unit timeor increasing the capacity of the basin, perhaps asa result of some tectonic event (Lambiase 1991). Incontrast, Schlische & Olsen (1990) argued that thetripartite stratigraphy and hanging wall onlaprelationships might simply reflect the infilling ofa basin that is growing in size through time. Even i fthe volumetric sedimentation rate is constant, ini-tial fluvial sedimentation will eventually giveway to lacustrine sedimentation, reflecting thecritical moment when the available supply ofsediment no longer can completely fill the growingbasin and a lake could occupy the space between thedepositional surface area and the outlet of the ba-sin. Eventually, as basin subsidence waned near the

R.W. Schlische Half-Graben Filling Models

6

end of the extensional period, the basin couldfill to its lowest outlet, and fluvial sedimenta-tion could return. A critical constraint on theextensional basin evolution and filling modelspresented below will be the necessary condi-tions (if any) under which the tripartite strati-graphic architecture and the hanging wall on-lap patterns develop.

Sedimentation rates are generally difficult todetermine in most continental strata, given theuncertainties in age control. However, in theNewark basin of eastern North America,Schlische & Olsen (1990) used the thicknessesof fixed-period Milankovitch cycles to calcu-late sedimentation rates in lacustrine strata.Although the data, which were collected fromdifferent localities in an asymmetric basin, ex-hibited some scatter, Schlische & Olsen noted ageneral decrease in sedimentation rates afterthe onset of lacustrine sedimentation in Triassicstrata, followed by a five-fold increase in la t -est Triassic and Early Jurassic strata, followedonce again by a decrease. Similar patterns areobserved in other early Mesozoic basins of east-ern North America. In the subsequent fillingmodels, particular attention will be paid totemporal variations in sedimentation rates.

Before attempting to reproduce these strati-graphic features in the three end-member mod-els of extensional basin development, it is bestto define some simple terms. Synrift strata arespecifically those strata that accumulated inthe fault-bounded basin while the faults wereactive, in contrast to post-rift strata, whichaccumulated after fault-controlled subsidence(generally during regional thermal subsidence)and which may have accumulated within thetopographic basin but probably accumulatedover a much wider area. The pre-rift—sediment contact marks the boundary be-tween pre-rift rocks and synrift strata; it gener-ally is not equivalent to the breakupunconformity, which commonly is present be-tween synrift and post-rift strata. In the threeextensional basin models described in this pa-per, the pre-rift—sediment contact is alwaysassumed to be horizontal before the start of ex-tension. Onlap, as defined here, is a variety ofstratigraphic termination in which synriftstratal units terminate against pre-rift rocksand/or more hingeward (i.e., away from thebasin boundary fault in the hanging wall block)than previously deposited synrift strata.Pinchout refers to the termination of a givensynrift stratal wedge against the previouslydeposited synrift stratal wedge. A fanningwedge is a form of stratigraphic termination in

which all stratigraphic horizons converge at thesame point (in two dimensions) or the same line (inthree dimensions).

CONTINENTAL HALF-GRABENFILLING MODELSIn this section the consequences of filling exten-sional basins with sediments during their develop-ment for each of three end-member types of basinsare explored. In each instance a given amount offault displacement results in basin subsidence. Pre-vious basin filling models assumed that the basinwould be completely and instantaneously infilledfollowing each increment of subsidence (Gibson e tal. 1989; Crews & McGrew 1990); given the strati-graphic arguments presented in the previous sec-tion, this would correspond to perennial fluvialsedimentation, but only in a terrestrial setting. Barr(1987) presented filling models for marine hal f -graben, assuming that the basins always filled tosea level. The approach taken here is similar tothat espoused by Schlische & Olsen (1990) for con-tinental full-graben: the half-graben is only filledwith the volume of sediment available per unittime (volumetric sedimentation rate). If the se-lected volume of sediment exceeds the capacity ofthe basin, fluvial sediments are deposited; other-wise lacustrine sediments accumulate. For the uni-form plane strain linked fault and domino models,every cross section of the basin drawn perpendicularto the border fault has the same cross-sectionalarea, and therefore the cross-sectional equivalentof a volumetric sedimentation rate can be used. Pre-viously deposited synrift strata are deformed dur-ing each subsequent increment of extension by pre-serving their cross-sectional areas, although bedlengths in the case of the linked fault and faultgrowth models must necessarily increase. The f i l l -ing process is then repeated, again using the sameavailable volume of sediment. Although a uniformsediment-supply rate may be unreasonable, particu-larly over very short time intervals, this approachwas chosen for simplicity and to highlight the in-herent tectonic differences among the three models.In all cases the stages of extensional basin evolutionare shown for uniformly-spaced time increments (5M.yr.). Although this approach is again geologi-cally unreasonable, it nevertheless does illustratein a convenient format the same general featuresthat are produced if the increments of extension andfilling are much smaller. For simplicity, the effectsof erosion or sediment compaction are not consid-ered. The elevation of the outlet of the basin(taken to be the elevation of the earth’s surfaceprior to extension) is assumed to remain fixed withrespect to the same surface in an adjacent region notaffected by extension. Because fluvial deposits

R.W. Schlische Half-Graben Filling Models

7

commonly form the basal sequences in numerouscontinental extensional basins, the modelsshown in Figs 3-7 emphasize the case of initialfluvial sedimentation, although, for the sakeof completeness, the effects of initial lacustrinesedimentation are also shown in some cases.Although volumetric sedimentation rate or itscross-sectional equivalent are held constant inall of the models illustrated, the qualitativeeffects of varying the sediment supply are alsodiscussed below. Special attention is paid toresulting stratal geometry, transitions betweenfluvial and lacustrine sedimentation, andsedimentation rates.

Detachment Fault ModelsThe results of filling half-graben that formedunder the assumptions of the detachment faultmodel are shown for a listric border fault in Fig.3 and for a planar border fault in Fig. 4. In a l lcases illustrated, the hanging wall collapsedalong antithetic faults dipping at 60°.

Following the first increment of fault dis-placement in Fig. 3B, the volume of the hal f -graben is less than the available volume ofsediment, and the half-graben filled com-pletely with fluvial sediment. Sedimentationrates progressively decrease toward the hingeof the basin and precisely mimic the subsidencerates. Since the second increment of displace-ment is equal to the first, area balance dictatesthat the newly created volume of the hal f -graben be equal to that of the first increment.Since the volumetric sedimentation rate is con-stant, the basin again completely fills withfluvial sediments. Notice that the youngerwedge of sediments pinches out against theolder wedge. This is because the footwall, thebasin-bounding fault, and the depocenter of thebasin remain fixed during extension, but thehinge of the basin migrates away from the ba-sin-bounding fault. This pattern of fluvialsedimentation and pinchout of younger strataagainst older strata would continue as long asthe displacement rate was uniform. As shown inthe third increment of displacement in Fig. 3B,however, a doubling of the amount of displace-ment also doubles the incremental volume ofthe basin, which now exceeds the volume ofsediments available. Lacustrine deposition oc-curs. Note that (1) the lacustrine wedge ofsediment pinches against older fluvial strata,(2) the maximum sedimentation rate in the l a -custrine wedge is higher than the maxima ofthe two older fluvial wedges, and (3) the depo-sitional surface area of the lacustrine wedge isless than for the fluvial wedges, requiring a

higher transverse gradient in sedimentation rates.Given initial fluvial sedimentation, lacustrinesedimentation can only occur if there is an increasein the extension rate and/or if the volumetric sedi-mentation rate decreases.

Figure 3C records the filling of an extensional ba-sin under conditions of accelerating fault displace-ment rates (to better facilitate comparisons withfault growth model I). The chosen volumetric sedi-mentation rate results in fluvial sedimentation fol-lowing the first two increments of displacement.After the third increment, the basin is of such a sizethat lacustrine sedimentation occurs. In general,under conditions of accelerated extension, youngerunits consistently pinch out against older units, themaximum sedimentation rate in younger units ishigher than in older units, and a transition fromfluvial to lacustrine is predicted if extension con-tinues long enough and if the effects of accelerateddisplacement overcome the effects of any increasein the volumetric sedimentation rate.

In the case illustrated in Fig. 3D, the volume ofsediments is less than the capacity of the basin, thebasin is sediment starved, and lacustrine sedimen-tation occurs from the outset. Under uniform faultdisplacement rates and constant volumetric sedi-mentation rate, lacustrine sedimentation continues,the basin becomes progressively more sedimentstarved, and the gap between the outlet of the ba-sin and the depositional surface progressively in-creases. As a consequence, maximum sedimentationrates in successively younger lacustrine units in-crease, and the depositional surfaces decrease inarea. As shown in the last stage of Fig. 3D, a de-crease in the displacement rate permits the young-est strata, which are fluvial, to onlap pre-riftrocks; the maximum sedimentation rate shows adecrease compared to the previous stratigraphicwedge. A decrease in the displacement rate or acessation of extension altogether is therefore thelikely explanation of why fluvial strata cap manystratigraphic sequences in continental extensionalbasins (Schlische & Olsen 1990).

The filling model shown in Fig. 4 for a kinkedlinked fault system is virtually identical to thelistric fault case illustrated in Fig. 3B. Younger flu-vial strata pinch out against older fluvial strata,and the transition from fluvial to lacustrine deposi-tion is possible only following an increase in thedisplacement rate or a decrease in the availablevolume of sediment. The major difference betweenthese two models is in the variation in sedimenta-tion rates within a sedimentary wedge. In the caseof the listric basin-bounding fault, sedimentationrates progressively increase from the hinge of thebasin toward the border fault. Because a f lat-bottomed graben develops in the case of a planar

R.W. Schlische Half-Graben Filling Models

8

basin-bounding fault, sedimentation rates a tthe time of deposition are constant over a largeregion. Because of hanging wall deformationduring continued extension, considerable thin-ning occurs within the stratal wedge. Its effectis to make stratal thicknesses and consequentlysedimentation rates even more uniformthroughout the wedge.

Domino-Style Fault Block ModelFilling models for domino-style half-grabenare shown in Fig. 5. In the first case considered(Fig. 5B), the volume of the basin is less thanthe available volume of sediment, resulting influvial sedimentation with excess sedimentleaving the basin. After the next uniform in-crement of extension, the available volume ofthe basin is even less than the preceding one,and fluvial sedimentation must continue. Notethat the maximum sedimentation rate of theyounger fluvial wedge is less than in the pre-ceding wedge. Furthermore, fluvial depositsdefine a fanning wedge geometry with a l lstratal horizons converging at the same line.This is because both the basin-bounding faultsand the intervening fault blocks all rotate dur-ing extension, and the edge of the basin is de-fined by the position of the next normal faultduring fluvial sedimentation. As shown in laststage of Fig. 5B, lacustrine sedimentation occursbecause of an increase in the extension rate. Thelacustrine wedge pinches out against previ-ously-deposited fluvial strata. The fluvial-lacustrine transition may also be occasioned bya decrease in the volumetric sedimentationrate. Figure 5C illustrates the effects of accel-erated extension on basin filling. The pattern isnearly identical to that of Fig. 5B except tha tthe maximum sedimentation rate progressivelyincreases in younger sedimentary wedges.

The filling history shown in Fig. 5D is foruniform extension and initially results in asediment-starved basin and lacustrine deposi-tion. The stratal wedges onlap pre-rift rocks.Because the volume of space created after eachincrement of extension is decreasing throughtime, the basin progressively becomes lesssediment starved, the depositional surface areadecreases, the maximum sedimentation ratedecreases, and younger strata continue to onlappre-rift rocks. As shown in the last stage of Fig.5D, the negative change in the rate of volumeincrease results in a basin in which the supplyof sediment exceeds the capacity of the basin,causing fluvial sedimentation. In this fluvialwedge, the maximum sedimentation rate is lessthan in all underlying units. If extension were to

continue at the same rate and with the same volu-metric sedimentation rate, the fluvial depositswould form a fanning wedge.

Fault Growth ModelsAn application of fault growth model I to an evolv-ing extensional basin is illustrated in Fig. 6A. Thefollowing parameters are used: final maximum dis-placement, D = 15,625 m; duration of active faulting, T = 25 M.yr.; repeat time of slip events, 5000 yr;shear modulus, G = 3.0 x 1010 Pa; stress drop, ∆σ =3.0 x 106 Pa; ratio of major and minor radii of faultsurface ellipse, 2; and dip of the fault, ϕ = 60°. Theshear modulus, stress drop, fault ellipse ratio areaverage values suggested by Walsh and Watterson(1987) and Gibson et al . (1989). The total number ofslip events (n = 4000) is obvious and used in equation(7) to determine the increment by which slip in-creases from seismic event to another , S = 1.25 mm.Using these parameters equation (6) calculates thelength of the fault zone as a function of maximumdisplacement, equation (8) determines how dis-placement varies with along-strike distance fromthe fault center, equations (9) and (10) describe howdisplacement varies normal to the fault surface,and equation (11) calculates the percentage contri-bution of hanging wall displacement. In all casessynsedimentary faulting is assumed (maximum dis-placement occurs at the originally horizontal freesurface at the initiation of faulting), and thus a l lthe calculated displacements are for the pre-rift—sediment contact. The displacement of thepre-rift—sediment contact in the hanging wallblock (hanging wall subsidence) are then contouredand are shown for time increments of 5 M.yr. in Fig.6A. Displacement of the same contact in the foot-wall block (footwall uplift) is not shown.

The filling of the evolving half-graben also re-quires elaboration. Sediments are only added aftereach of the 5-M.yr.-long time increments. Other-wise the thicknesses of sediment deposited aftereach of the slip events would have been impossiblysmall to represent and would have resulted in 4000stratal horizons. One can therefore think of theapproach used in this paper as a lumping of indi-vidual stratal horizons into 5-M.yr.-long timeequivalents. The available volume of sediment per5-M.yr.-long time increment, 360 km3, chosen toyield a fluvial-lacustrine transition at T = 10M.yr., is iteratively fitted to the basin. Volumebalancing is achieved by dividing the basin into anumber of boxes of fixed width (17.2 km). Cross-sectional areas are calculated for transects throughthe center-line of each box and converted into vol-umes by multiplying them by the width of theboxes. The sum of all the boxes shown and multi-

R.W. Schlische Half-Graben Filling Models

9

plied by 2 approximates the total volume ofthe sediments. Of course, as the width of theboxes is reduced to infinitely small, the truevolume is approached as a limit.

Figure 6A shows the mapped distribution ofthe youngest sediments for each time increment,and Fig. 6B presents all of the cross sectionsused in the volume-balanced filling process.Following the first 5 M.yr., the available vol-ume of sediments vastly overwhelms the capac-ity of the basin by almost a factor of 10 to 1;fluvial sedimentation results. At 10 M.yr. afterthe start of subsidence, the new capacity of thebasin matches the volume of available sedi-ments; nonetheless, the basin still fills to itslowest outlet with fluvial sediments. Thisyounger wedge of fluvial sediments everywhereonlaps pre-rift rocks. Lacustrine strata fill thesediment-starved basin at 15 M.yr. after theonset of subsidence. Although the along-strikedimension of this sedimentary wedge is ap-proximately the same as for the underlyingunit, the transverse dimension is considerablysmaller, requiring that the maximum thicknessof the wedge to increase over that of the pre-ceding wedge. The lacustrine wedge pinches outagainst the underlying fluvial strata (cross sec-tions a3 and b3, Fig. 5B). A second lacustrinewedge is deposited at 20 M.yr.; it pinches outagainst the underlying lacustrine unit in crosssection a4 but onlaps the younger of the fluvialwedges in cross section b4. The lacustrine wedgedeposited at T=25 M.yr. onlaps fluvial strata incross sections a5 and b5 but onlaps prerift rocksin cross section C5. Because subsidence is concen-trated toward the end of its active history, thebasin becomes progressively more and moresediment-starved. In the “5-series” of cross sec-tions the effects of filling the sediment-starvedbasin (volume-balanced) after the cessation ofactive faulting are explored. The two sedimen-tary wedges are both lacustrine and are bothwider than the last synrift wedge. They onlapolder synrift deposits in cross sections a5 and b5and onlap pre-rift rocks in cross section c5.There is no change in dip angle between thesetwo youngest stratal wedges.

In the two fluvial wedges, the sedimentationrates (thickness/time) are everywhere equal tothe incremental subsidence rates. Sedimenta-tion rates decrease toward the lateral edgesand toward the hanging wall hinge of the ba-sin. The maximum sedimentation rate in eachfluvial wedge increases in progressivelyyounger strata. In progressively younger synriftlacustrine strata, the maximum sedimentationrate is constant or decreases slightly at the cen-

ter of the basin and increases slightly in those crosssections located closer to the lateral edge of thebasin. Within a given lacustrine wedge, sedimenta-tion rates generally increase toward the fault a tthe center of the fault trace. The post-rift lacus-trine units deposited after fault displacementceased record a decrease in maximum sedimentationrate because their depositional surface areas in-crease through time, and thus the thickness ofsediment deposited per unit time decreases. Thesepost-rift strata may be recognized by the large mapregion over which the sedimentation rate is con-stant within a stratal wedge. This is because theseunits were deposited over much of their extent on aflat-surface (the undeformed upper surface of thelast synrift unit).

The case for initial lacustrine sedimentationseems trivial. The basin would be sediment starvedfrom the outset. Because the rate of subsidence in-creases throughout the active life of the basin, thealmost negligible amount of sediment available(assuming a constant supply)—in effect—would beconcentrated at the bottom of a very deep hole.Given the unlikelihood of this scenario coupledwith the slow start to subsidence in the beginning ofits extensional history and the resultant small ca-pacity for accepting sediment, fault growth model Iof basin evolution almost requires that the initialdeposits of a basin be fluvial.

Figure 7 shows the evolutionary and filling his-tory of an extensional basin governed by faultgrowth model II. The following parameters areused: final maximum displacement, D = 12,500 m;final fault length, L = 175,000 m; linear strain, ε =0.07; total duration of extension, Τ = 25 Myr; amountof slip per slip event, q = 1 m; and repeat time forslip events, t = 2000 years. As in the other faultgrowth model, the basin is shown after incrementsof 5 M.yr. The hanging wall block displacement ofan originally horizontal surface is shown by theisodisplacement contours in Fig. 7A. Stratal wedgesare also shown for 5 M.yr. increments. Approximatevolume balancing, using a volumetric sedimentationrate of 650 km3/5 M.yr. (designed to yield a flu-vial-lacustrine transition at T = 10 M.yr.), is againapproximately achieved through the equal subdi-vision of the basin into boxes of width of 17.5 km;cross sections through the center of each box areshown in Fig. 7B.

After the first 5 M.yr. of fault displacement, theavailable volume of sediments exceeds the capac-ity of the basin, and therefore fluvial sedimenta-tion results. These fluvial strata everywhere onlappre-rift rocks. At 10 M.yr. after the start of subsi-dence, the volume of the basin precisely matchesthe available volume of sediments. Fluvial sedi-

R.W. Schlische Half-Graben Filling Models

10

mentation is again the result, and these strataagain everywhere onlap pre-rift rocks. Lacus-trine strata fill the basin 15 M.yr. after the ini-tiation of subsidence. These strata pinch outagainst the underlying fluvial strata. Lacus-trine sedimentation continues following thenext two increment of displacement and basinfilling. These stratal wedges pinchout againstolder lacustrine strata in cross sections a4 anda5, onlap fluvial strata in cross sections b4 andb5, and onlap pre-rift rocks in cross sections c4and c5.

As in the first fault growth model, sedimen-tation rates in fluvial strata equal the incre-mental subsidence rates. Sedimentation ratesgenerally decrease in all directions from centerof the basin immediately adjacent to the borderfault. Maximum sedimentation rate in youngerfluvial wedges is constant, for the maximumincremental subsidence rate is also constant.However, this maximum sedimentation rate isonly found at the center of the basin. In a l lother parts of the basin, sedimentation rates inyounger fluvial strata should increase throughtime. For example, notice that the sedimenta-tion rates in the first fluvial wedge at the posi-tion of the “a-series” of cross sections is consid-erably less than for the second fluvial wedge,despite the fact that the fault displacementrate is uniform. This can be explained as fol-lows. For T = 5 M.yr., cross section a1 is locatedmidway between the center of the basin and itslateral termination. Thus the subsidence at thisposition is approximately one-half of themaximum subsidence at the center of the basin.For T = 10 M.yr., cross section a2 is located twiceas far from the edge of the basin as it is fromthe center of the basin. Thus, for this positionfixed relative to the center of the basin, thesubsidence rate increases through time, eventhough the maximum subsidence rate at thecenter of the basin is uniform. Note that thesevariations in sedimentation rate are only possi-ble in fluvial strata, where the supply of sedi-ment exceeds the capacity of the basin.

Lacustrine strata in all cross sections show adecrease in maximum sedimentation rates up-section because the depositional surface areaincreases through time. Fault growth model Ionly exhibited this trend in the a-series of crosssections. Sedimentation rates within a lacus-trine wedge increase toward the border fault a tthe center of the basin.

As for fault growth model I, the case of ini-tial lacustrine sedimentation is somewhat triv-ial. Progressively younger lacustrine stratawill onlap pre-rift rocks and show a progres-

sive decrease in maximum sedimentation rates, butall sediment will be deposited at the bottom of avery sediment-starved basin.

SUMMARY AND DISCUSSIONAll three end-member models of extensional basindevelopment discussed in this paper are inherentlydifferent. The domino fault block and linked faultsystem models are uniform plane strain and involveno along-strike changes through their evolution;the fault growth models provide for the along-strike dimension of the basin to increase. For a con-stant displacement rate or uniform extension, thechange in the rate at which the basin volume in-creases from one time increment to the next is zerofor the detachment fault model, negative for thedomino fault block model, and positive for faultgrowth model II. Fault growth model I—as pre-sented here—involves an increasing displacementrate through time, which also translates into apositive change in the rate of increase in basin vol-ume. Transitions from fluvial to lacustrine transi-tion in the detachment fault and domino fault blockmodels require either an increase in the displace-ment or extension rate and/or a decrease in theavailable volume of sediment. The fault growthmodels provide a natural transition from fluvial tolacustrine deposition because of the positive changein the rate of increase in basin capacity, althoughthe timing of this transition may be hastened orpostponed by decreasing or increasing, respectively,the sediment supply. Prior to this modeling, flu-vial-lacustrine transitions were seen as a responseto accelerated subsidence and/or decreasing sedi-ment supply (e.g., Lambiase 1991). The faultgrowth models have shown that the transition maysimply be produced in a basin which the change inthe rate of increase in basin capacity is positive.

Each half-graben model also produces its ownunique suites of stratal geometries. The detachmentfault model generally is characterized by pinchoutof younger stratal units against older synrift strata,with the exception of the oldest stratal wedge,which rests directly on pre-rift rocks. This patternapplies to both fluvial and lacustrine strata. In thedomino fault block model fluvial strata define afanning wedge geometry, and younger lacustrinestrata pinchout against older fluvial strata. In abasin in which only lacustrine strata are found,these units onlap pre-rift rocks. The fault growthmodels result in fluvial strata that progressivelyonlap pre-rift rocks and lacustrine strata tha tpinch out against or onlap syn-rift rocks in the cen-ter of the basin and onlap pre-rift rocks at its la t -eral edges.

No particular half-graben model can explain a l lthe major stratigraphic transitions exhibited in the

R.W. Schlische Half-Graben Filling Models

11

tripartite stratigraphy and the hanging wallonlap features observed in many continentalextensional basins (Tables 1 and 2). However,fault growth models come closest. These modelsare particularly successful in explaining hang-ing wall onlap patterns in the fluvial sequenceand the transition from fluvial to lacustrinesedimentation. These models also predict somehanging wall onlap for lacustrine strata in someparts of the basin. They do not specificallypredict a portion of the third element of thetripartite stratigraphy (fluvial deposits).These sequences may be generated during fillingof the sediment-starved basin after subsidencehas stopped or slowed considerably and thebasin has filled to its lowest outlet.

Fault growth model II, with its uniform dis-placement rate, appears to best explain thegeneral decrease in sedimentation rates in suc-cessively younger Triassic lacustrine strata ofthe Newark basin of eastern North America(Schlische & Olsen 1990). An increase in thedisplacement rate (see results of fault growthmodel I) may account for the pronouced increasein sedimentation rates in latest Triassic andEarly Jurassic strata. As discussed by Schlische& Olsen (1990), this increase in displacementrates shortly preceeded a widespread butshortlived igneous extrusion and intrusion eventin eastern North America. The subsequent de-cline in sedimentation rates in Jurassic lacus-trine strata may reflect a return to a more uni-form or decelerating displacement rates.

In addition to explaining most features of thestratigraphic development and stratal geome-try of continental extensional basins, the faultgrowth models are also superior in explainingthe source of sediment for extensional basins.Based on the physiography of the East Africanrift system, Lambiase (1991) concluded tha tsediment is derived principally from the hang-ing wall block, which slopes toward the depo-sitional basin, and from rivers flowing downthe axis of the rift system. Only a minor compo-nent of sediment is derived from the footwallblock, which slopes away from the rift basindue to footwall uplift. Rivers flowing down theslope of the footwall may eventually enter thebasin at its lateral ends, where there is nofootwall uplift, and contribute to the axialsource of sediments. Schlische (1990a) notedthat the provenance of sediment within theMesozoic Newark basin of eastern North Amer-ica supports a similar conclusion. The faultgrowth models specifically predict a hangingwall that slopes toward the basin and a foot-wall that slopes away from the basin, a geome-

try identical to that following a single seismicevent (Fig. 2B; Stein and Barrientos 1985). Theflexural-cantilever model of Kusznir et al. (1991),in which the hanging wall and footwall blocksflexurally/isostatically respond to displacement onplanar normal faults, produces a similar geometry.It should be noted that, for any given basin situatedbetween domino-style fault blocks, the hangingwall also slopes toward the basin and the footwallaway from the basin. The footwall of the basin-bounding fault does not slope away from the basinin the linked fault system model because the foot-wall is assumed to be passive, although this is aproblem with the model and not necessarily withdetachment faults per se.

Neither the domino fault block nor the detach-ment fault filling model generates strata that on-lap pre-rift rocks during initial fluvial sedimenta-tion nor in any lacustrine strata deposited after theaccumulation of fluvial strata. Both models arecapable of generating the transition from fluvial tolacustrine transition but only through changes inthe displacement or extension rate and/or thevolumetric sedimentation rate. Distinguishing be-tween these two mechanisms is difficult, butchanges in the displacement rate may be recognizedif data on the attitude of strata coupled with chro-nostratigraphic information are available. Sincestrata progressively increase in dip during exten-sion in all models except the detachment faultmodel involving a kinked fault geometry (compareFigs 3, 5, 6B, and 7B with Fig. 4), changes in differ-ential dip between groups of beds representing ap-proximately equal durations of time should indi-cate a change in the extension rate. The modelingalso indicates a general increase in maximum sedi-mentation rate associated with an increase in thedisplacement or extension rate.

Although the detachment fault model and thedomino fault block model are inadequate in ex-plaining some features commonly observed in exten-sional basins, they should not be summarily aban-doned. Particularly for the domino model, there aresome well-documented examples in the geologicrecord (e.g., Proffett 1977). However, given theirdrawbacks and questionable assumptions (notablyuniform plane strain), care should be taken wheninvoking them. The domino model should only beused [for example, in estimating the amount of ex-tension; equation (2)] in areas where the followingkey features are present: (1) numerous normal faultsdip in the same direction and by the same amount ;(2) strata in adjacent (and presumably genetically-related basins) have similar dips, and/or (3) basinsshow a fanning-wedge geometry. Because thechange in the rate of volume increase is negative,domino-style basins should contain predominantly

R.W. Schlische Half-Graben Filling Models

12

fluvial deposits. The detachment fault model,in my opinion, has been particularly overap-plied and therefore should only be invoked incases where both the basin bounding fault andthe detachment are known to exist. The infer-ence of a listric normal fault is not sufficient ininvoking this model because: (1) roll-over ge-ometry, commonly attributed to listric faults(Hamblin 1965; Wernicke & Burchfiel 1982;Gibbs 1983, 1984) can also be generated by pla-nar normal faults (Gibson et al. 1989; see alsoFig. 2B and the flexural-cantilever models ofKusznir et al. 1991) and (2) improperly mi-grated seismic lines may give the appearanceof kinked or listric geometry (Unger 1988).Nonetheless, some of the features of the faultgrowth models can be incorporated into the vet-eran domino and linked fault system models(e.g., Walsh & Watterson 1991).

Given the lessons of the variable fault dis-placement model, it is also clear that careshould be taken in constructing balanced crosssections, for it is unlikely that uniform planestrain conditions prevail on the scale of a singleextensional basin. It remains to be seen if uni-form plane strain is a reasonably good ap-proximation for a region containing numerousbasins that subsided coevally, i.e., where a de-crease in subsidence in one basin is taken up inanother basin parallel to it. The results pre-sented in this paper also suggest that two-dimensional forward stratigraphic modelinghas limited applicability.

Finally, care should be taken in interpretingvery specific results of the basin filling modelspresented here because: (1) the effects of com-paction and erosion of previously depositedsediments were not considered, (2) fluvial de-posits were not allowed to aggrade above theoutlet level of the basin, (3) the displacementand filling increments in all models were unre-alistically large, and (4) the isostatic conse-quences of sediment loading were not consid-ered. Future basin filling models should seek toremedy these deficiencies and should be testedagainst a growing body of fine-scale strati-graphic data for continental extensional basins(e.g., Olsen & Kent 1990). Nonetheless, thesimplifying assumptions used in the modelspresented here should not detract from themain thrust of this paper—that there are in-herent tectonic differences among the threeend-member models, which yield differentstratal geometries and successions in model ba-sins. Although these filling models are clearlynot appropriate in predicting bed-scale strati-

graphy, they are useful in understanding, if notpredicting, basin-scale stratigraphy.

ACKNOWLEDGEMENTSI thank M.H. Anders, G.M. Ashley, N. Christie-Blick, P.A.Cowie, M. Levy, P.E. Olsen, and W.B.F. Ryan for valuablediscussions on this topic and for reviewing the manuscript.D. Barr and J. Watterson, the reviewers for Basin Research,provided extremely useful reviews that corrected errors inthe original manuscript. This paper is an outgrowth of achapter of the author’s Ph.D. dissertation written at Co-lumbia University. Some funding for the research was pro-vided through a Henry A. Rutgers Research Fellowshipgranted by Rutgers University.

REFERENCESANDERSON, R. E., ZOBACK, M. L. & THOMPSON, G. A. (1983)

Implications of selected subsurface data on the structuralform and evolution of some basins in the northern Basinand Range province, Nevada and Utah. Geol. Soc. Amer-ica Bull. 94, 1055-1072.

AXEN, G. J. (1988) The geometry of planar domino-stylenormal faults above a dipping basal detachment. J. Struct.Geol. 10, 405-411.

BARNETT, J. A. M., MORTIMER, J., RIPPON, J. H., WALSH, J. J. &WATTERSON, J. (1987) Displacement geometry in the vol-ume containing a single normal fault. Am. Assoc. Petr.Geol. Bull. 71, 925-937.

BARR, D. (1987) Structural/stratigraphic models for exten-sional basins of half-graben type. J. Struct. Geol. 9, 491-500.

CHAPMAN, G. R., LIPPARD, S. J. & MARTYN, J. E. (1978) Thestratigraphy and structure of the Kamasia Range, KenyaRift Valley. J. Geol. Soc. Lond. 135, 265-281.

COWIE, P. A., & SCHOLZ, C. H. (1990) Fault growth and faulttermination. EOS [Transactions of the American Geo-physical Union] 71, 631.

CREWS, S. G. & MCGREW, A. J. (1990) Influence of struc-tural style on rift basin morphology and non-marine se-quence geometry. Geol. Soc. America Abstr. Programs 22(7), A239.

ELLIOT, D. (1976) The energy balance and deformationmechanism of thrust sheets. Phil. Trans. R. Soc. LondonSer. A 283, 289-312.

EMMONS, W. H. & GARREY, G. H. (1910) General geology.In: Geology and Ore Deposits of the Bullfrog District(RANSOME, F. L. et al.): U.S. Geol. Surv. Bull. 407, 19-89.

FRASER, G. D., WITKIND, I. J. & NELSON, W. H. (1964) A geo-logical interpretation of the epicentral area—the dual-basin concept. U. S. Geol. Surv. Prof. Paper 435, 99-106.

GIBBS, A. D. (1983) Balanced cross-section constructionfrom seismic sections in areas of extensional tectonics. J.Struct. Geol. 5, 153-160.

GIBBS, A. D. (1984) Structural evolution of extensional ba-sin margins. J. Geol. Soc. Lond. 141, 609-620.

GIBSON, J. R., WALSH, J. J. & WATTERSON, J. (1989) Modellingof bed contours and cross-sections adjacent to planarnormal faults. J. Struct. Geol. 11, 317-328.

GROSHONG, R. H., Jr. (1989) Half-graben structures: Bal-anced models of extensional fault-bend folds. Geol. Soc.America Bull. 101, 96-105.

HAMBLIN, W. K. (1965) Origin of “reverse drag” on thedownthrown sides of normal faults. Geol. Soc. AmericaBull. 76, 1145-1164.

JACKSON, J. A. (1987) Active normal faulting and crustalextension. In: Continental Extensional Tectonics (Ed. byCOWARD, M. P., DEWEY, J. F. & HANCOCK, P. L.). Spec.Publ. Geol. Soc. London 28, 3-17.

KUSZNIR, N. J., MARSDEN, G. & EGAN, S. S. (1991) A flex-ural-cantilever simple-shear/pure-shear model of conti-nental lithosphere extension: Applications to the Jeanned’Arc basin, Grand Banks and Viking Graben, NorthSea. In: The Geometry of Normal Faults (Ed. by ROBERTS,A. M., YIELDING, G. & FREEMAN, B.) Spec. Publ. Geol. Soc.

R.W. Schlische Half-Graben Filling Models

13

London 56, 41-60.LAMBIASE, J. J. (1991) A model for tectonic control of

lacustrine stratigraphic sequences in continental riftbasins. In: Lacustrine Exploration: Case Studies andModern Analogues (Ed. by KATZ, B. ). Am. Assoc.Petrol. Geol. Memoir 50, 265-276.

LEEDER, M. R. & GAWTHORPE, R. L. (1987) Sedimentarymodels for extensional tilt-block/half-graben ba-sins. In: Continental Extensional Tectonics (Ed. byCOWARD, M. P., DEWEY, J. F. & HANCOCK, P. L.). Spec.Publ. Geol. Soc. London 28, 139-152.

MANDL, G. (1987) Tectonic deformation by rotatingparallel faults: the "bookshelf" mechanism. Tectono-physics 141, 277-316.

MILLER, E. L., GANS, P. B. & GARING, J. (1983) TheSnake Range decollement: an exhumed mid-Tertiaryductile-brittle transition. Tectonics 2, 239-263.

MORLEY, C. K. (1989) Extension, detachments, andsedimentation in continental rifts (with particularreference to East Africa). Tectonics 8, 1175-1192.

MURAOKA, H., & KAMATA, H. (1983) Displacement dis-tribution along minor fault traces. J. Struct. Geol. 5,483-495.

OLSEN, P. E. (1986) A 40-million-year lake record ofearly Mesozoic climatic forcing. Science 234, 842-848.

OLSEN, P. E. & KENT, D. V. (1990) Continental coring ofthe Newark rift basin. EOS [Trans. American Geo-phys. Union] 71, 385 & 394.

PROFFETT, J. M., Jr. (1977) Cenozoic geology of the Yer-ington district, Nevada, and implications for the na-ture and origin of Basin and Range faulting: Geol.Soc. America Bull. 88, 247-266.

SCHLISCHE, R. W. (1990a) Anatomy of the Newark ex-tensional basin: Structural geometry and tectonos-tratigraphic evolution. In: The rifting of a continent:The geology of the Newark basin (Ed. by HUSCH, J. M.,HOZIK, M. J. & OLSEN, P. E.). Geol. Soc. AmericaMemoir in Honor of Franklyn B. Van Houten (inpress).

SCHLISCHE, R. W. (1990b) Aspects of the structural andstratigraphic development of early Mesozoic rift ba-sins of eastern North America: Ph.D. thesis, ColumbiaUniversity, 579 p.

SCHLISCHE, R. W. & OLSEN, P. E. (1990) Quantitativefilling model for continental extensional basins withapplications to the early Mesozoic rifts of easternNorth America. J. Geology 98, 135-155.

SCHOLZ, C. H. (1982) Scaling laws for large earth-quakes: Consequences for physical models. Bull.Seis. Soc. America 72, 1-14.

SCLATER, J.G. & CÉLÉRIER, B. (1989) Errors in extensionmeasurements from planar faults observed on seismic re-flection lines: Basin Research 1, 217-221.

SIBSON, R. H. (1985) A note on fault reactivation. J. Struct.Geol. 7, 751-754.

SMOOT, J. P. (1985) The closed basin hypothesis and the useof working models in facies analysis of the Newark Su-pergroup. U.S. Geol. Surv. Circular 946, 4-10.

STEIN, R. S. & BARRIENTOS, S. E. (1985) The 1983 BorahPeak, Idaho, earthquake; Geodetic evidence for deep rup-ture on a planar fault. U.S. Geol. Surv. Open File Report85-290, 459-484.

THOMPSON, G. A. (1960) Problem of late Cenozoic struc-ture of the Basin Ranges. Proc. 21st Int. Geol. Congr., Co-penhagen, 18, 62-68.

UNGER, J. D. (1988) A simple technique for analysis andmigration of seismic reflection profiles from the Meso-zoic basins of eastern North America. In: Studies of theEarly Mesozoic Basins of the Eastern United States (Ed.by FROELICH, A. J. & ROBINSON, G. R., Jr.). U.S. Geol. Surv.Bull. 1776, 229-235.

VERRALL, P. (1981) Structural interpretations with applica-tions to North Sea problems. Joint Association of Petro-leum Exploration Course Notes 3, London.

WALSH, J. J. & WATTERSON, J. (1987) Distribution of cumula-tive displacement and of seismic slip on a single normalfault surface. J. Struct. Geol. 9, 1039-1046.

WALSH, J. J. & WATTERSON, J. (1988) Analysis of the rela-tionship between displacements and dimensions of faults.J. Struct. Geol. 10, 239-247.

WALSH, J. J. & WATTERSON, J. (1989) Displacement gradientson fault surfaces. J. Struct. Geol. 11, 307-316.

WALSH, J. J. & WATTERSON, J. (1991) Geometric and kine-matic coherence and scale effects in normal fault systems.In: The Geometry of Normal Faults (Ed. by ROBERTS, A.M., YIELDING, G. & FREEMAN, B.) Spec. Publ. Geol. Soc.London 56, 193-203.

WATTERSON, J. (1986) Fault dimensions, displacements andgrowth. Pure Appl. Geophys. 124, 365-373.

WERNICKE, B. P. & AXEN, G. J. (1988) On the role of isostasyin the evolution of normal fault systems: Geology 16, 848-851.

WERNICKE, B. & BURCHFIEL, B. C. (1982) Modes of exten-sional tectonics. J. Struct. Geol. 4, 105-115.

WHITE, N.J., JACKSON, J.A. & MCKENZIE, D. P. (1986) Therelationship between the geometry of normal faults andthat of the sedimentary layers in their hanging walls. J.Struct. Geol. 8, 897-909.

WILLIAMS, G. & VANN, I. (1987) The geometry of listric nor-mal faults and deformation in their hanging walls. J.Struct. Geol. 9, 789-795.

Table 1. Stratigraphy of continental extensional basins. Summarized from Schlische & Olsen (1990) andLambiase (1990).

Basin Location Age Stratigraphic Succession

Keweenawan USA/Canada Precambrian Fluvial lacustrine fluvialMorondava Madagascar Permian-Triassic Fluvio-deltaic deep lacustrine fluvio-deltaic fluvialMombasa Kenya Permian-Triassic Lacustrine fluvio-deltaicDeep River Eastern USA Triassic Fluvial lacustrine fluvialDan River Eastern USA Triassic Fluvial deep lacustrine shallow lacustrine/fluvialRichmond Eastern USA Triassic Fluvial deep lacustrine shallow lacustrine fluvialCulpeper1 Eastern USA Triassic-Jurassic Fluvial lacustrine fluvio-lacustrine; deep lacustrineGettysburg Eastern USA Triassic-Jurassic Fluvial lacustrineNewark1 Eastern USA Triassic-Jurassic Fluvial deep lacustrine shallow lacustrine/fluvial;

deep lacustrine shallow lacustrineHartford/Deerfield Eastern USA Triassic-Jurassic Fluvial deep lacustrine shallow lacustrine fluvialFundy1 Eastern Canada Triassic-Jurassic Fluvial lacustrine shallow lacustrine;

deep lacustrine shallow lacustrineReconcavo Brazil Cretaceous Deep lacustrine deltaic fluvialWest African1 Gabon/Angola Cretaceous Fluvial deep lacustrine deltaic;

shallow lacustrine deltaicSudan1 Southern Sudan Cretaceous Fluvio-lacustrine deep lacustrine fluvial;

lacustrine deltaic fluvialCentral Sumatra Sumatra Paleogene Shallow lacustrine deep lacustrine fluvio-deltaic1Indicates basin that experienced multiple tectonic cycles (dual-cycle basins of Lambiase 1990)

Table 2. Basins in which progressively younger synrift strata onlap pre-rift rocks of the hanging wallblock. Summarized from Schlische & Olsen (1990).

Basin Location AgeRichmond Eastern USA TriassicNewark Eastern USA Triassic-JurassicAtlantis Offshore eastern USA Triassic-Jurassic?Long Island Offshore eastern USA Early MesozoicNantucket Offshore eastern USA Early MesozoicFundy Eastern Canada Triassic-JurassicHopedale Labrador Margin MesozoicSaglek Labrador Margin MesozoicNorth Viking North Sea MesozoicTanganyika East Africa CenozoicDixie Valley Western USA CenozoicNorthern Fallon Western USA CenozoicDiamond Valley Western USA CenozoicRailroad Valley Western USA CenozoicGreat Salt Lake Western USA Cenozoic

.

i i

F' F' F F

F'F'

Area A

Area A

Area A

Area B

Area B

Area B

A(1) (2)

h h

B

c c'

d d'

c

c'

d d'TA

a

a'

b b'

a a'

A

b b'

L

R

C

Fault displacement (km)(horizontal component)

Cum

ulat

ive

basi

n vo

lum

e (x

10

km )3

3

01 2 3 4 5

1

2

3

4

5

Fault growth I

Fault growth II

Domino

Detachment

6

7

8

D

d d

FIG. 1: Three models of extensional basin development. (A) Linked fault system model involves two endmembers: (1) listric fault-subhorizontal detachment and (2) planar kink fault geometry. In both instanceshorizontal displacement (h) on the detachment fault creates a potential void between the hanging wall andfootwall, which is erased by the collapse of the hanging wall along vertical faults in (1) and antithetic faultsdipping at 45° in (2). The deformation is area balanced. Adapted from Gibbs (1983) and Groshong (1989). (B)Domino fault block model in which both the faults and the intervening fault blocks rotate during extension. i isthe initial dip angle of the faults; is the dip after extension; is the dip of a horizon that was horizontal beforeextension; F' is the initial fault spacing; F is the fault spacing after extension. Adapted from Wernicke &Burchfiel (1982). (C) Essential elements of the fault growth model (modified from Gibson et al. 1989). The ruled"ellipse" is the map view of a normal fault in which displacement is greatest at the fault center and decreases tozero at the ends. Contours represent the elevation change (positive for dotted contours, negative for solidcontours) of the originally horizontal free surface. Note that the footwall uplift is smaller than the hanging wallsubsidence. L is the length of the fault, R is the radius of the fault (L/2), T is fault motion toward the reader, A isaway. (D) Graph of cumulative basin volume vs. horizontal component of fault displacement for the modelspresented in this paper. The change in the rate of increase in basin volume is zero for the detachment faultmodel, negative for the domino model, and positive for the two fault growth models.

.

log L (L=length of fault in cm)0 2 4 6 8 10

-2

0

2

4

6

8

log

Dm

(D

m =

mea

n d

isp

lace

men

t in

cm

)

Thrust earthquakes (Scholz 1982)

Normal faults in lake sediments(Muroaka & Kamata 1983)

Canadian Rockies (Elliot 1976)

British coalfield normal faults(Walsh & Watterson 1987)

=10-1

=10

=10

-2

-3-4

-5=1

0=1

0

L D1/2

L D

C

(Walsh & Watterson 1988)

(Cowie & Scholz 1990)

100102030

0.4

-0.4

-1.2

0

-0.8

Distance from fault (km)

Ele

vati

on

ch

ang

e (m

)

B

10 km0

2

-2

Footwall cutoff

Hanging wall cutoff

Cross fault

km

A

FIG. 2: (A) Horizon separation diagram for the Tiim Phonolite on the Saimo fault, Kenya rift. Note that theseparation is greatest at the center of the fault and decreases toward either end. Modified from Chapman et al.(1978). (B) Elevation change resulting from the 1983 Borah Peak Earthquake on the Lost River fault, Idaho.Note that this geometry closely matches that of a half-graben. Modified from Stein & Barrientos (1985). (C)Log-log plot of fault length vs. mean displacement for dip-slip faults and two suggested scaling laws.Compilation courtesy of M.H. Anders and C.H. Scholz.

h = 10 km3

A

B

C

h = 5 km

D

0 5 10 15

km

A = 50 km

1

2

V = 3000 km31

1

h = 5 km2

22A = 50 km

32

V = 3000 km

V = 6000 km3

3

1

h' = 2.5 km1

A' = 25 km 2

V' = 1500 km1

3

h' =7.5 km

A' = 75 km

V' = 4500 km

2

2

2

2

3

A' = 150 km3

2

V' = 9000 km3

3

h" = 5 km

A" = 50 km

V" = 3000 km

1

1

1

2

3

h " = 5 km

A" = 50 km

V" = 3000 km

2

22

23

h" = 2 km

A" = 20 km

V" = 1200 km

3

3

3

2

3

a = 50 km

v = 3000 km1

1

2

3

a = 50 km

v = 3000 km2

2

2

3

a = 55.5 km

v = 3330 km3

3

2

3

a' = 25 km

v' = 1500 km1

1

2

3

a' = 75 km

v' = 4500 km

2

2

2

3

a' = 100 km

v' = 6000 km3

3

2

3

a" = 40 km

v" = 2400 km1

1

2

3

a" = 40 km

v" = 2400 km

2

32

2

a" = 40 km

v" = 2400 km

2

33

3

onlap

onlap

pinchout

onlap

onlap

T = T' = T" = 0 M.yr.

T = 5 M.yr.

T = 10 M.yr.

T = 15 M.yr.

T' = 5 M.yr.

T' = 10 M.yr.

T' = 15 M.yr.

T" = 5 M.yr.

T" = 10 M.yr.

T" = 15 M.yr.

pinchout

pinchout

pinchout

pinchout

d = 10 km

A = 100 km3

2

h' = 15 km3

FIG. 3: Filling sequences for half-graben governed by the linked faultsystem model involving a listricborder fault. For all cases theconfiguration illustrated in (A) is thestate before extension;the hangingwall collapsed along antithetic faultsdipping at 60°. The along-strikedimension of the basin is 60 km. Alsoshown are the cross-sectional area (A)and equivalent volume (V) of theincremental capacity of the basin toreceive sediment after each incrementof fault displacenent (h) as well as thecross-sectional area (a) and volume (v)of sediment deposited. The availablevolume of sediments is 3330 km3/5M.yr. for the sequence shown in (B),6000 km3/5 M.yr. for the sequenceshown in (C) and 2400 km3/5 M.yr.for the sequence shown in (D). Fluvialsediments are stippled; lacustrinesediments are shaded gray.

h = 5 km

A = 50 km

V = 3000 km

a = 50 km

v = 3000 km

A = 50 kmV = 3000 km

h = 5 km

a = 50 km

v = 3000 km

A =150 kmV = 6000 km

h = 15 km

a = 62.9 kmv = 3780 km3

3

2

3

3

3

3

2

3

2

2

2

2

32

2

2

3

1

1

2

3

1

1

1

2

3

T = 0 M.yr.

T = 5 M.yr.

T = 10 M.yr.

T = 15 M.yr.

0 5 10 15

km

pinchout

pinchout

d = 10 km

FIG. 4: Filling sequence for a half-graben governed by the linked fault system modelinvolving a planar border fault. The hanging wall collapsed along antithetic faultsdipping at 60°. The along-strike dimension of the basin is 60 km. Also shown are thecross-sectional area (A) and equivalent volume (V) of the incremental capacity of thebasin to receive sediment after each increment of fault displacement (h) as well as thecross-sectional area (a) and volume (v) of sediment deposited. The available volume ofsediments is 3780 km3/5 M.yr. Fluvial sediments are stippled; lacustrine sediments areshaded gray.

70°

A

B

D

C

a = 44.6 km12

v = 2680 km13

a = 31.7 km

v = 1900 km2

2

2

3

v = 3310 km

a = 55.1 km

3

32

3

a' = 25.2 km

v' = 1510 km1

1

2

3

a' = 34.2 km

v' = 2050 km2

22

3

a' = 34.2 km32

v' = 2050 km33

a" = 35.2 km12

v" = 2110 km13

a" = 35.2 km

v" = 2110 km2

2

2

3

a" = 32.6 km

v" = 1960 km3

32

3

1

= 1.1

A = 44.6 km

V = 2680 km1

12

3

= 1.22

A = 31.7 km2

2

V = 1900 km23

3

= 1.4

A = 57.3 km

V = 3440 km

3

32

3

' = 1.051

A' = 25.2 km12

V' = 1512 km13

' = 1.152

A' = 34.2 km22

2V' = 2050 km3

' = 1.43

A' = 57.3 km32

V' = 3440 km33

" = 1.11

A" = 44.6 km12

V" = 2680 km13

" = 1.22

A" = 31.7 km22

2V" = 1900 km3

" = 1.33

A" = 28.4 km32

V" = 1700 km33

onlap

onlap

fanning wedge

fanning wedge

0 5 10

kmT = T' = T" = 0 M.yr.

T = 5 M.yr.

T = 10 M.yr.

T = 15 M.yr.

T' = 5 M.yr.

T' = 10 M.yr.

T' = 15 M.yr.

T" = 5 M.yr.

T" = 10 M.yr.

T" = 15 M.yr.

pinchout

pinchout

F' = 20 km FIG. 5: Filling sequences for half-graben governed by the dominofault block model. For all casesthe configuration illustrated in(A) is the state before extension,with an initial fault spacing of 20km. The along-strike dimensionof the basin is 60 km. Also shownare the cross-sectional area (A)and equivalent volume (V) of theincremental capacity of the basinto receive sediment after eachincrement of exten-sion ( ) aswell as the cross-sectional area (a)and volume (v) of sedimentdeposited. The available volumeof sediments is 3310 km3/5 M.yrfor the sequence shown in (B),2050 km3/5 M.yr for thesequence shown in (C) and 2110km3/5 M.yr for the sequenceshown in (D). Fluvial sedimentsare stippled; lacus-trinesediments are shaded gray.

.

d4

c4

b4

a4

T =

20

M.y

r.L

= 1

37.4

km

D h

w =

700

0 m

c.i.

= 1

600

m

a1

T =

5 M

.yr.

L =

34.

4 km

D h

w =

438

mc.

i. =

200

m

a2

b2

T =

10

M.y

r.L

= 6

8.8k

mD

hw =

175

0 m

c.i.

= 4

00 m

a3

b3

c3

T=

15

M.y

r.L

= 1

03 k

mD

hw =

393

8 m

c.i.

= 8

00 m

a5

b5

c5

d5

e5T

= 2

5 M

.yr.

L =

172

km

D h

w =

10,

937

mc.

i. =

200

0 m

17.2 km

0

500

1000

200

400

800

0

100

200

400

0

100

200

0

100

0

(exc

ept a

s no

ted)

(exc

ept a

s no

ted)

(exc

ept a

s no

ted)

(exc

ept a

s no

ted)

(exc

ept a

s no

ted)

010

20

km

515

25

100

FIG

. 6: (

A) F

ault

gro

wth

mod

el I

of e

xten

sion

al b

asin

dev

elop

men

t. T

he r

uled

elli

pses

are

the

map

vie

ws

of th

e ba

sin-

boun

din

g no

rmal

faul

t, an

d th

eco

ntou

rs a

re fo

r th

e d

epth

of t

he p

re-r

ift—

sed

imen

t con

tact

. Not

e th

at th

e ba

sin

incr

ease

s in

siz

e bo

th a

long

-str

ike

and

tran

sver

sely

dur

ing

faul

td

ispl

acem

ent,

the

rate

of w

hich

incr

ease

s th

roug

h ti

me.

Dhw

is th

e m

axim

um a

mou

nt o

f han

ging

wal

l dis

plac

emen

t. T

he s

had

ing

ind

icat

es th

e m

ap-v

iew

exte

nt o

f str

ata

dep

osit

ed a

ssum

ing

that

the

avai

labl

e vo

lum

e of

sed

imen

t was

360

km

3 /5

M.y

r. T

he b

oxes

sho

w th

e re

gion

s ov

er w

hich

vol

umes

wer

ees

tim

ated

.

a1

a2

a3

a4

a5

b3

b4

b5

c3

c4

c5

d4

d5

e5

0 10 15

km

b2

a = 1.0 km

a = 8.7 km

a = 7.2 km

a = 6.8 km

a = 6.5 kma = 5.7 km

a = 5.6 km

p1

p2

a = 1.8 km

a = 3.3 km

a = 3.7 km

onlap

pinchout

pinchout

onlap

onlappinchout

onlap

onlap

onlap

2

5

a = 3.6 kma = 4.0 km

a = 3.9 km

p1

p2

a = 0.4 kma = 0.8 km

a = 1.0 km

p1

p2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

2

FIG. 6: (B) Cross sections drawn through the medial lines of the boxes shown in (A). In all cases the cross-sectionalarea (a) of the youngest synrift stratal wedge is indicated. When multiplied by the width of the sampling box andsummed for all boxes (and multiplied by 2), this value approximates the volume of sediment deposited after eachtime increment. This value should be 360 km3 for lacustrine deposits. Fluvial synrift sediments are stippled;lacustrine synrift sediments are shaded gray. The " 5-series" cross sections also show two stratal wedges (no pattern)deposited after fault displacement had stopped; their cross-sectional areas are indicated by Ap1 and Ap2 for the firstand second wedges, respectively.

.

d5

c5

b5

a5

T =

25

M.y

r.L

= 1

75 k

mD

hw =

875

0 m

c.i.

= 2

000

m

e5

0

500

1000

d4

c4

b4

a4

T =

20

M.y

r.L

= 1

40 k

mD

hw =

700

0 m

c.i.

= 1

600

m

a1

T =

5 M

.yr.

L =

35

kmD

hw =

175

0 m

c.i.

= 2

00 m

a2

b2

T =

10

M.y

r.L

= 7

0 km

D h

w =

350

0 m

c.i.

= 4

00 m

a3

b3

c3

T=

15

M.y

r.L

= 1

05 k

mD

hw =

525

0 m

c.i.

= 8

00 m

010

20

km

515

25

200

17.5 km

400

800

00

200 10

0

400

0

200 10

00

100

(exc

ept a

s no

ted)

(exc

ept a

s no

ted)

(exc

ept a

s no

ted)

(exc

ept a

s no

ted)

(exc

ept a

s no

ted)

FIG

. 7: (

A) F

ault

gro

wth

mod

el II

of e

xten

sion

al b

asin

dev

elop

men

t. T

he r

uled

elli

pses

are

the

map

vie

ws

of th

e ba

sin-

boun

din

g no

rmal

faul

t, an

d th

eco

ntou

rs a

re fo

r th

e d

epth

of t

he p

re-r

ift—

sed

imen

t con

tact

. Not

e th

at th

e ba

sin

incr

ease

s in

siz

e bo

th a

long

-str

ike

and

tran

sver

sely

dur

ing

faul

td

ispl

acem

ent,

whi

ch is

uni

form

thro

ugh

tim

e. D

hw is

the

max

imum

am

ount

of h

angi

ng w

all d

ispl

acem

ent.

The

sha

din

g in

dic

ates

the

map

-vie

w e

xten

t of

stra

ta d

epos

ited

ass

umin

g th

at th

e av

aila

ble

volu

me

of s

edim

ent w

as 6

50 k

m3 /

5 M

.yr.

The

box

es s

how

the

regi

ons

over

whi

ch v

olum

es w

ere

esti

mat

ed.

0 5 10 15

km

a = 7.1 km

pinchoutb3

a = 8.5 km

onlapb4

onlap

a = 7.5 km

b5

c3

onlap

a = 2.1 km

c4

onlapc5

a = 4.5 km

d4

d5

e5

a = 2.3 km

a1

onlap

a = 15.5 km

a2

pinchout

a = 11.5 km

a3

pinchout

a = 7.9 km

a4

pinchout

a = 6.6 km

a5

b2a = 3.1 km onlap

2

2

2

2

2

2

2

2

2

2

2

FIG. 7: (B) Cross sections drawn through the medial lines of the boxes shown in (A). In all cases the cross-sectional area(a) of the youngest stratal wedge is indicated. Fluvial sediments are stippled; lacustrine sediments are shaded gray.