sea level and vertical motion authors of continents from dynamic earth models … › 35517 › 1...

AUTHORS

Sonja Spasojevic Seismological Labora-tory California Institute of Technology 1200East California Boulevard MC 252-21 PasadenaCalifornia present address BP America Hous-ton Texas Kisis1bpcom

Sonja Spasojevic received a BSc degree in geo-physics from the University of Belgrade Serbiain 2001 and an MSc degree in exploration geo-

Sea level and vertical motionof continents from dynamicearth models since theLate CretaceousSonja Spasojevic and Michael Gurnis

physics from the University of Houston in 2003After working for two years as a geophysicistfor BP in the deep-water Gulf of Mexico shejoined the California Institute of Technology(Caltech) in 2005 to work on research related tothe dynamics of long-term sea level changevertical motion of continents and evolution ofsedimentary basins with Michael Gurnis as ad-visor After receiving her PhD in geophysics in2010 she continued working as a staff scientistat Caltech during 2011 She recently returned toBP America working in the Brazil regional ex-ploration team

Michael Gurnis Seismological Labora-tory California Institute of Technology Pasa-dena California gurnisgpscaltechedu

Michael Gurnis received a BS degree from theUniversity of Arizona in 1982 and a PhD fromthe Australian National University in 1987 both ingeophysics Following a postdoctoral researchat the California Institute of Technology (Caltech)he was a faculty member at the University ofMichigan In 1994 he returned to Caltech wherehe is currently John E and Hazel S Smits Pro-fessor of Geophysics and director of the Seis-mological Laboratory His effort is divided be-tween linking geodynamics and the rock record(especially stratigraphy) the basic physics ofgeophysical phenomena studies of the deepinterior of the Earth and the development ofnew computational methods including theGPlates package for plate tectonic modeling

ACKNOWLEDGEMENTS

All calculations were conducted at the California

ABSTRACT

Dynamic earth models are used to better understand the im-pact of mantle dynamics on the vertical motion of continentsand regional and global sea level change since the Late Cre-taceous A hybrid approach combines inverse and forwardmodels of mantle convection and accounts for the principalcontributors to long-term sea level change the evolving distri-bution of ocean floor age dynamic topography in oceanic andcontinental regions and the geoid We infer the relative im-portance of dynamic versus other factors of sea level changedetermine time-dependent patterns of dynamic subsidence anduplift of continents and derive a sea level curve

We find that both dynamic factors and the evolving dis-tribution of sea floor age are important in controlling sea levelWe track the movement of continents over large-scale dynamictopography by consistently mapping between mantle and plateframes of reference and we find that this movement results indynamic subsidence and uplift of continents The amplitude ofdynamic topography in continental regions is larger than globalsea level in several regions and periods so that it has controlledregional sea level in North and South America and Australiasince the Late Cretaceous northern Africa and Arabia since thelate Eocene and Southeast Asia in the OligocenendashMioceneEastern and southern Africa have experienced dynamic upliftover the last 20 to 30 my whereas Siberia and Australia haveexperienced Cenozoic tilting The dominant factor controllingglobal sea level is a changing oceanic lithosphere productionthat has resulted in a large amplitude sea level fall since theLateCretaceous with dynamic topography offsetting this fall

Institute of Technology (Caltech) GeosciencesSupercomputer Facility partially supported byNational Science Foundation EAR-0521699 Thiswork was supported through Statoil the NationalScience Foundation (EAR-0810303) and theCaltech Tectonics Observatory (by the Gordon

Copyright copy2012 The American Association of Petroleum Geologists All rights reserved

Manuscript received August 12 2011 provisional acceptance November 21 2011 revised manuscriptreceived February 1 2012 final acceptance March 26 2012DOI10130603261211121

AAPG Bulletin v 96 no 11 (November 2012) pp 2037ndash2064 2037

and Betty Moore Foundation) The originalCitcomS software was obtained from the Com-putational Infrastructure for Geodynamics Thisis contribution number 169 of the Division ofGeological and Planetary Sciences and contri-bution number 10063 of the Tectonics Obser-vatory CaltechThe AAPG Editor thanks the following reviewersfor their work on this paper Steven G Henryand Peter Webb

DATASHARE 45

High-resolution versions of Figures 2 and 4ndash8with captions and appendixes 1ndash3 may beseen on the AAPG Web site (wwwaapgorgdatashare) as Datashare 45

2038 Sea Level from Dynamic Earth Models

INTRODUCTION

Continental interiors experienced extensive marine inunda-tion and sedimentation during the Cretaceous associated witha global (eustatic) sea level high (Hallam 1992) The marineinundation of North America is a clear example of the ex-tensive flooding where approximately 40 of the continentflooded resulting in more than 1 km (gt06 mi) of sedimentsbeing deposited over a horizontal scale of 1000 km (621 mi)(Bond 1976Liu andNummedal 2004)The eustatic change ontime scales of millions of years may be controlled by changes inocean basin volume caused by variations in oceanic spreadingrates or ridge lengths (Hays and Pitman 1973 Kominz 1984)However North American Cretaceous stratigraphic sequences(Sloss 1963) suggest that observed continental-scale trans-gressional and regressional cycles cannot be explained by eustasyalone (Bond 1976 Sleep 1976) and that large-scale tectonicor dynamic mechanisms are required Namely analysis of theCretaceous North America sediments suggests that the ob-served flooding would require a sea level rise of 310 m (1017 ft)resulting in accumulation of 700 m (2297 ft) of sediments(Bond 1976) whereas the observed Cretaceous isopachs (Cookand Bally 1975) are significantly thicker

Continental flooding can be controlled by dynamic topog-raphy a vertical deflection of the surface resulting frommantleflow (eg Richards and Hager 1984) Although dynamic to-pography results in large-amplitude geoid anomalies becausethese anomalies are small compared to topography (Hager1984) sea level variations should be mainly controlled bychanges in dynamic topography (Gurnis 1990) Early geody-namic studies explored the basic process of dynamic controlsof sea level developing simple two-dimensional (2-D) isovis-cous models of convection that demonstrated that slabs createbroad topographic depressions that can be filled with sediments(Mitrovica and Jarvis 1985) and that continental flooding isinfluenced by movement of continents over large-scale patternsof dynamic topography (Gurnis 1990) Early regional studieslinked mantle dynamics and vertical motions with observationsand models of North America showed that general features ofPhanerozoic strata can be explained by dynamic topographycreated by subduction and supercontinent aggregation and dis-persal (Burgess et al 1997) The extensive Cretaceous flood-ing resulting in the creation of significant hydrocarbon accu-mulations and the subsequent uplift of the Western Interiorseaway (WIS) were interpreted first with simple 2-D dynamicmodels of Farallon slab subduction (Mitrovica et al 1989) andmore recently with three-dimensional (3-D) inverse models

assimilating seismic tomography and platemotions(Liu et al 2008 Spasojevic et al 2009) Three-dimensional models applied to Australia showedthat anomalous Cretaceous vertical motions can berelated to the movement of the continent over aslab associated with Gondwanaland-Pacific sub-duction while also explaining the anomalous geo-chemistry and geophysics of the present-day seafloor south ofAustralia (Gurnis et al 1998Gurniset al 2000b) Two episodes of long-wavelengthtilting of the Russian platform during the Devo-nian to the Permian have been attributed to sep-arate episodes of subduction (Mitrovica et al 1996)whereas the excess subsidence of the Campbellplateau between 70 and 40 Ma have been relatedto the drift of New Zealand away from a dynamictopography high associated with a Ross Sea mantleupwelling (Sutherland et al 2009 Spasojevic et al2010a) Global models of mantle flow driven bypaleogeographically constrained subduction matchpatterns of flooding in the Middle OrdovicianLate Permian and Early Cretaceous (ie periods offlooding during eustatic highs) (Gurnis 1993) aswell as the Cenozoic uplift and subsidence of NorthAmericaAustralia and Indonesia (Lithgow-Bertelloniand Gurnis 1997) In addition it was demonstratedthat dynamic topography should also modify theshape (depth) of the basins of the oceans globallywhich will affect the time-dependent trends of sealevel change (Gurnis 1993 Conrad and Husson2009)

These studies demonstrate the importance ofmantle dynamics in modulating global and regionalsea levels and in some cases they have linked themodels to stratigraphic constraints but they havesome limitations Most dynamic topography mod-els were in a mantle frame with the exception ofthe regionalmodels of Australia (Gurnis et al 1998Gurnis et al 2000b DiCaprio et al 2009) andNorth America (Liu et al 2008 Spasojevic et al2009) whereas geologic observations are made inthe plate frame of reference Previous global modelsof sea level change have not accounted for long-termfactors of sea level change self-consistently Herewe present spherical dynamic earthmodels (DEMs)since the Late Cretaceous that overcome these limi-tations by assimilating plate tectonic reconstructions

into variable viscosity models of mantle convec-tionWe combine inverse and forwardmethods forsolving for mantle convection allowing estimatesof initial conditions with improved formulations fordensity viscosity and convergent plate margins Wealso use globally distributed stratigraphic data toconstrain the control of the mantle on uplift andsubsidence and regional and global sea level changeGlobal DEMs are computed in the mantle framebut the integrated plate reconstructionndashmantle con-vection system enables a linkage with geologic ob-servations in plate frames of reference We firstcompute the shape of the ocean and eustasy thataccounts for changing ocean floor age (OFA) (Muumllleret al 1997 Muumlller et al 2008b) time-dependentdynamic topography and the geoid Second wecompute marine inundation and paleoshorelineswithin continental interiors and compare themwithgeologic observations Such DEMs account for themost important factors controlling long-term sealevel change self-consistently and from them weinfer the relative importance of different factors forthe control of vertical motions and tilting of conti-nents and sea level change since theLateCretaceous

Dynamic earth models provide an improvedunderstanding of the effects of mantle dynamicson subsidence uplift and tilting of continents thathave important implications for petroleum systemsMantle dynamics affect basin subsidence and sed-imentation and in turn the deposition of sourcerock reservoirs and seals Dynamic uplift can af-fect petroleum systems through the cessation ofthe active hydrocarbon maturation or expulsion ofhydrocarbons variability in sediment supply af-fecting evolution of drainage patterns and tilting ofsedimentary sequences and traps One of the goalsof DEMs is to understand which sedimentary ba-sins have been affected by mantle dynamics withthe hope that these findings will be incorporatedinto hydrocarbon exploration workflows

METHOD

Dynamic earth models link plate reconstructionsand seismic images with forward and inverse man-tle convectionmodels inwhich the resultingmodel

Spasojevic and Gurnis 2039

predictions are compared with stratigraphic ob-servations Dynamic earth models are developedfor the periods spanning the Late Cretaceous tothe present with a goal of improving our under-standing of the history of the vertical motion ofthe surface of earth and global and regional sealevels


Plate Reconstructions

We use plate reconstructions in which plate geom-etry is represented as continuously evolving closedplate polygons using the GPlates package (Boydenet al 2011) (Figure 1) Each plate is defined as arigid plate with its own Euler pole and a set of

Figure 1 GPlates reconstructions at 9060 and 30 Ma The colors represent thereconstructed age of the sea floor (Muumllleret al 2008a) the arrows show plate ve-locities in the moving hotspot frame ofreference The black lines with filled blacktriangles indicate the position of subduc-tion zones whereas the plain black linesindicate ridges and transform plateboundaries

evolving margins each having a different Euler pole(Gurnis et al 2012) defined in a moving hotspotreference frame (Muumlller et al 2008a)Continuouslyevolving plate margins are necessary so that slabsremain coherent as they continuously subductWedynamically close plates by continuously findingthe intersection of all margins with no gaps or over-laps between adjacent plates (Gurnis et al 2012)The continuously closed plate polygons are self-consistent with paleoage grids that give the age ofoceanic lithosphere (Muumlller et al 2008a) The timeinterval between reconstructions can be as small asneeded but here we use 1 my

Different aspects of the plate reconstructionsare assimilated into the DEMs First plate veloc-ities defined at 1-my steps and linearly interpo-lated at intermediate times are used as a surfacekinematic boundary condition on the convectionmodels Second the position and character of plateboundaries are used to define buoyancy associatedwith subduction in hybrid models Third the pa-leoage grids are used with the hybrid models aswell as for global time-dependent sea level calcula-tions Finally the plate reconstructions are used formapping outputs from themantle to plate frames ofreference using closed plate polygons and rotationpoles enabling the comparison ofmodel predictionswith stratigraphic constraints

Mantle Convection Models

Models of mantle convection are formulated usingthe finite-element codeCitcomS version 32 (Zhonget al 2000 Tan et al 2006) which solves theequations of mass momentum and energy con-servation for an incompressible Newtonian fluidwhile making the Boussinesq approximation Pre-vious mantle convection models investigating dy-namic effects on sea level are defined either as for-ward or inverse models The main limitation offorward models run from the geologic past to thepresent is the unknown past structure of the man-tle Inverse models however are run from thepresent to the past with the present-day mantlestructure constrained by seismic tomography How-ever inverse models are limited by (1) not beingable to reverse thermal diffusion to predict past

structures which is reasonable for the mantle in-terior away from boundary layers on time scalesstudied here (Conrad and Gurnis 2003 Liu andGurnis 2008) (2) the long-wavelength and low-resolution mantle structure and (3) the inabilityto realistically treat subduction zones in viscousmodels (Liu et al 2008)

Here we formulate hybrid geodynamic modelsthat combine inverse (Conrad and Gurnis 2003Liu and Gurnis 2008) with forward models toovercome their individual limitations First we sim-ply integrate backward (eg simple backward in-tegration [SBI] method of Liu and Gurnis 2008)from the present to the Late Cretaceous with ki-nematic surface velocity by reversing the directionof plate motions (Figure 1) The initial global SBItemperature field at 0 Ma is defined from seismictomographic inversions of surface and body wavesusing models SB4L18 (Masters et al 2000) orS20RTS (Ritsema et al 2004) with tomography-to-density scaling constrained by fitting the ob-served geoid (Spasojevic et al 2010b) Second wecreate a hybrid buoyancy field in the past by merg-ing the SBI with synthetic subducted slabs For eachtime step in the past we define lower mantle buoy-ancy from the SBI whereas in the upper mantle wemerge high-resolution synthetic slabs with positivebuoyant anomalies from the SBI The geometry andthermal structure of subducted slabs in the uppermantle are defined from the reconstructed subduc-tion zones from GPlates (Figure 1) age of the sub-ducted lithosphere (Muumlller et al 2008a) and rela-tions among subduction zone parameters (C Tapeet al 2011 unpublished work) Simple backwardintegration buoyancy is stripped from the upper250 km (155 mi) of the mantle as it is likely closeto neutral buoyancy (Jordan 1988Goes and van derLee 2002)

Finally the Stokes equation is solved in a for-wardmodel with this hybrid buoyancy imposing ano-slip boundary condition at the top surface toobtain the dynamic topography and geoid Dy-namic topography on the top surface is determinedfrom the total normal stress in the radial directionwhile accounting for self-gravitation (Zhong et al2008) Figure 2 shows the predicted dynamic to-pography geoid and buoyancy for a hybrid model


A four-layer mantle-viscosity structure consistingof a high-viscosity lithosphere upper mantle tran-sition zone and lower mantle is modified by atemperature-dependent viscosity which introduces


lateral viscosity variations shown to be an importantinfluence on dynamic topography and the geoid insubduction zones (Billen and Gurnis 2001 Billenet al 2003) and for present-day plate motions

Figure 2 Dynamic topography and geoid prediction for model M3 (A) Predicted geoid at 0 Ma (B) Observed geoid (CndashD) Predicteddynamic topography at 0 Ma (C) and 80 Ma (D) (EndashF) Global equatorial spherical cross section through nondimensional temperaturefield at 0 Ma (E) and 80 Ma (F) See a high-resolution version on Datashare 45 wwwaapgorgdatashare

(Stadler et al 2010) Details of the viscosity res-olution and other model parameters are given inAppendix 1 (see Datashare 45 at wwwaapgorgdatashare)

Paleogeographic Constraints and Relative SeaLevel Changes

The primary geologic constraints are a set of pub-lished global and regional inferred paleoshorelinesfrom the Late Cretaceous to the present We usetwo sets of global Cretaceous and Cenozoic shore-lines one from Smith et al (1994) determined atevery 10-my intervals from 90 to 10 Ma and asecond from Blakey (2008) defined for the LateCretaceous (90Ma)CretaceousndashTertiary boundary(65 Ma) Eocene (50 Ma) Oligocene (35 Ma) andMiocene (20 Ma) We also use a regional set ofreconstructions for (1) northern and central Africa

and Arabia (Guiraud et al 2005) in latest Albianndashearly Senonian (989ndash85 Ma) late SantonianndashMaastrichtian (84ndash65Ma) Paleocene (65ndash548Ma)earlymiddleEocene (548ndash37Ma)Oligocene (337ndash238 Ma) and Miocene (238ndash53 Ma) (2) South-east Asia (Hall andMorley 2004) in late Oligocene(284ndash23 Ma) early Miocene (23ndash16 Ma) and mid-dleMiocene (16ndash116Ma) and (3) Australia at 7768 60 44 33 20 8 and 4 Ma (Langford et al1995) with numerical ages indicating the middleof the respective reconstruction periods Divisioninto marine and subaerial units simplifies the pa-leoshoreline maps

We infer relative sea level change by differenc-ing these maps between a more recent (t1) and anearlier age (t2 t2 gt t1) Marine regression occurswhen a submerged area at t2 is exposed above thesea level at t1 implying a eustatic fall a sedimen-tation rate that outpaces accommodation space

Figure 3 Two-dimensional schematicof the algorithm used for global sea levelcalculations in dynamic earth models witht denoting time VE = vertical exaggeration


creation regionally or a relative uplift regionally(Hallam 1992) Marine transgression occurs whena subaerially exposed area at t2 is flooded at a sub-sequent period t1 which can be the result of a eu-static rise a sea level rise outpacing sedimentationregionally or regional subsidence When an area ofinterest is subaerially exposed (or flooded) at bothtimes may be also experiencing vertical motionsand sedimentation but with amplitudes and wave-lengths insufficient to result in changes of flooding

We compare maps of relative sea level changetomaps of differential dynamic topography createdby subtracting dynamic topography at time inter-val t1 from dynamic topography at t2 (t1 lt t2) Wehypothesize that a good agreement between mapsof relative sea level change and maps of differentialdynamic topography might imply that dynamictopography is one of the factors contributing tothe observed regional shoreline migration If suchagreement does not exist then shoreline migra-tion is likely controlled by factors other thanmantleconvection-induced subsidence and uplift Thisreasoningwill be limited if the hybridmodel fails topredict the correct trends in dynamic topographybecause of incorrect estimates of the mantle buoy-ancy field We attempt to overcome this limitationby finding alternative dynamic models that providea better agreement between differential dynamictopography and inferred relative sea level change

Global Sea Level Calculations

To calculate global sea level we first determineglobal topography at each time step t (Figure 3Aand B) which is subsequently inundated for pa-leoshoreline predictions (Figure 3C) The ocean-continent function is defined from oceanic age grids(Muumlller et al 2008a) such that areas with OFAsare defined as oceanic and the remaining are as-sumed continental For oceanic regions bathym-etry is nominally determined from age-depth re-lationships (Figure 3A) using a half-space thermalcooling model (Parsons and Sclater 1977) withoutflattening

d frac14 2500 m +350 mffiffi

tp eth1THORN


where d is ocean floor depth in meters and t is sea-floor age in million years We assume a half spacewithout flattening because age-depth relationshipsthat include flattening (Parsons and Sclater 1977Stein and Stein 1992) likely have a component ofdynamic topography The air-loaded topographyof ocean floor is

dal frac14rm rwrm ra

times d frac14 0689times d eth2THORN

where dal is the air-loaded depth of ocean floorand rm rw and ra are the densities of mantlewater and air respectively (Appendix 1)

For continental regions we first define air-loadedisostatic topography (Figure 3A) At present thetopography is assumed to be the sum of an isostaticand a dynamic component We determine the iso-static component at time t by subtracting the pre-dicted dynamic topography at 0 Ma from the ob-served topography (Figure 3A) and merging therotated field for each plate from 0 Ma to time tThis assumes that isostatic topography remains un-changed from 0Ma to time t a significant limitationin places because it ignores orogenesis and com-plicates global sea level predictions The final air-loaded topography at time step t is obtained bysumming air-loaded dynamic topography (Figure 3B)and isostatic topography

The air-loaded global topography is then in-undated (Figure 3C) Starting with the lowestparts of the air-loaded topography we fill depressedareas with the present-day volume of ocean waterso that the water surface conforms to the evolvinggeoid which is well predicted at present (Figure 2)(Spasojevic et al 2010b) In areas covered withwater isostatic adjustment for water loading is ap-plied and the final topography is a combinationof water-loaded areas and air-loaded continents(Figure 3C) Paleoshorelines are the intersectionof the final topography with the ocean surface(Figure 3C) at each time step We compare shore-line predictions that only account for the changingOFA with those that account for dynamic topog-raphy geoid and OFA (DYN-OFA) We infer therelative importance of dynamic factors versus oceanfloor age on sea level change based on the comparison

of these two predictions at 80Ma (Late Cretaceous)60Ma (early Cenozoic) and 30Ma (mid-Cenozoic)Results for India are excluded because they arebiased by the time-invariant isostatic assumption(because of the Himalayan orogen)

We also attempt to quantify global sea levelchange (1) because of variations in ocean age usingthe half-spacemodel without flattening (equation 1OFA-HS) half-space model with flattening (Ap-pendix 1 OFA-HSF) and global depth and heatflow (GDH)ndash1 model (Stein and Stein 1992) (Ap-pendix 1 OFA-GDH) by calculating mean oceanbathymetry using relationships 4 to 7 inAppendix 1and (2) by calculating variations in sea level forDYN-OFAwith the half-space cooling model whiletracking contributions caused by dynamic topog-raphy in oceanic regions (eg regions with definedOFAs in age grids) dynamic topography in floodedcontinental regions (eg regions that are predictedto be inundated but have no values of OFAs in agegrids) and geoid The DYN-OFA sea level esti-mates are corrected for sedimentation and oceanplateau emplacement using previously publishedtime-dependent estimates (Muumlller et al 2008a)

RESULTS

Global Dynamic Topography

We compute six hybrid models (Table 1) for threedifferent radial viscosity profiles and two seismic to-mographymodels The reference viscosity is 1021 Pa s(Appendix 1) and an upper mantle average back-ground viscosity of 1020 Pa s is observed for allmodels Dynamic topography decreases with in-creasing lower mantle viscosity (Appendix 2 seeAAPG Datashare 45 at wwwaapgorgdatashare)The largest dynamic topography lows are located inthe regions of subduction so that at 0 Ma large

dynamic topography lows develop in the westernPacific Southeast Asia South America and theMediterranean During the Late Cretaceous dy-namic topography lows are in western North Amer-ica associated with subduction of the Farallon slaband the region of Alpine Tethys subduction Dy-namic topography highs are associated with oce-anic regions (PacificOceanNorthernAtlantic andIndian Ocean) and Africa between the Late Creta-ceous and the present (Appendix 2)

Regional Sea Level and Vertical Motions

North AmericaNorth America experienced a widespread marineinundation during the Late Cretaceous resultingin the development of the WIS (Sloss 1988) fol-lowed by an extensive uplift during the Laramideorogeny (English and Johnston 2004)Maps of LateCretaceous relative sea level (Figure 4A) show theWIS being continuously flooded its eastern edgeexperiencing marine transgression and the rest ofNorth America remaining subareal and presum-ably high (Smith et al 1994 Blakey 2008) Dy-namic earth models predict an increasing dynamictopography low in the WIS (Figure 4A) with thewestern edge being aligned with the Sevier beltModel M4 predicts well the eastern edge of theWIS whereas model M1 predicts a wider differen-tial dynamic topography low covering both theWISand eastern North America (Figure 4A) Cenozoicrelative sea level maps indicate that North Amer-ica either experienced an uplift or remained high(Figure 5A) Dynamic earth models predict ac-cumulated dynamic uplift of western North Amer-ica (Figure 5A) over the Cenozoic and dynamicsubsidence of the eastern part of the continent withthe East Coast subsiding between 100 and 400 m(328 and 1312 ft)

Table 1 Hybrid Model Summary

Model Name

M1

M2

M3

Sp

M4

asojevic and G

M5

urnis

M6

Tomography model in SBI

S20RTS

S20RTS

S20RTS

SB4L18

SB4L18

SB4L18

Average transition zone to lower mantle viscosity ratio

120

160

1100

120

160

1100

SBI = simple backward integration

2045

Figure 4 Inferred relative sea level change in the Late Cretaceous for North America (A) northern Africa and Arabia (B) Eurasia (C)South America (D) and Australia (E) The first image in panels A to E represents relative sea level inferred from paleogeography usingreconstructions at 70 and 90 Ma (Smith et al 1994) (A) late SantonianndashMaastrichtian and latest Albianndashearly Senonian (Guiraud et al2005) (B) 65 and 90 Ma (Blakey 2010) (CndashD) and 60 and 77 Ma (Langford et al 1995) (E) Relative vertical motions inferred fromdynamic models are shown with ldquoMrdquo corresponding to the specific hybrid model (Table 1) See a high-resolution version on Datashare 45wwwaapgorgdatashare


Figure 4 Continued

The predicted pattern of vertical motions(Table 2) is related to the subduction of the Far-allon slab Remnants of the Farallon slab are pres-ently imaged as a north-southndashtrending highndashseismic velocity mid-mantle anomaly under theEast Coast of North America (Masters et al 2000Grand 2002 Ritsema et al 2004) As the mantleflow and platemotions are reversed in the SBI fromthe present to the Late Cretaceous the Farallonslab rises upward and North America moves east-ward in the mantle frame of reference so that it islocated in the upper mantle under the WIS duringthe Late Cretaceous High-density Farallon slab cre-ates a dynamic topography low with an amplitudeof approximately 1 km (sim06 mi) (Appendix 2) re-sulting in 500 m (1640 ft) of Late Cretaceous dy-namic subsidence (Figure 4A) We could not matchthe observations of differential vertical motions andflooding without the introduction of a shallow-dipping slab with a dip of approximately 10deg sim-ilar to previous studies that propose a shallow to

flat-lying Farallon slab during the Late Cretaceous(Mitrovica et al 1989 Liu et al 2008 Spasojevicet al 2009) which could be potentially related tothe subduction of the Shatsky oceanic plateau con-jugate (Liu et al 2010) As North America moveswestward in the mantle reference frame from theLate Cretaceous to the present and the Farallonslab sinks into the lower mantle the dynamic to-pography lowmigrates eastward in the plate frameresulting in an overall Cenozoic subsidence of east-ern North America and a dynamic uplift in the west(Figure 5A) The Cenozoic subsidence does notcreate land subsidence because it is contempora-neous with an overall sea level fall The proposeddynamic subsidence may help explain discrepan-cies betweenmost global sea level curves and thosederived exclusively on the New Jersey coastal mar-gin (Spasojevic et al 2008) The timing of theLaramide orogeny is currently debated but it prob-ably initiated in the Late Cretaceous and the pro-posed timing of its termination ranges from 35 to

Table 2 Inferred Dynamic Subsidence (ndash) and Uplift (+) from Dynamic Earth Models

Continent

Late Cretaceous

PaleocenendashEocene

Spasojevic and Gu

OligocenendashMiocene

North America (Western Interior seaway)

ndash

+

+

North America (east)

ndash

ndash

Northern Africa

ndash

ndash

ndash

Southern and eastern Africa

+

Southeast Asia

+

ndash

ndash

Eurasia (north-central)

Siberian tilt

Siberian tilt

Alpine Tethys

ndash

ndash

ndash

South America (Western Interior)

ndash

+

South America (Amazon)

ndash

ndash

ndash

South America (Patagonia)

+

Australia

+()

ndash

ndash

rnis 2047

Figure 5 Inferred relative sea level change in the Cenozoic 0 to 60 Ma in North America (A) 0 to 40 Ma in northern Africa and Arabia(B) 0 to 60 Ma in Eurasia (C) 10 to 30 Ma in Southeast Asia (D) 0 to 20 Ma in South America (E) and 0 to 60 Ma in Australia (F) Thefirst image in panels A to E represents relative sea level inferred from paleogeography using reconstructions at 60 Ma (Smith et al 1994)(A) Miocene and early-middle Oligocene in northern Africa (Guiraud et al 2005) and 35 Ma in southern and central Africa (Blakey 2010)(B) 65 Ma (Blakey 2010) (C) middle Miocene and late Oligocene (Hall and Morley 2004) (D) 20 Ma (Blakey 2010) (E) and 60 Ma(Langford et al 1995) (F) Relative vertical motions inferred from dynamic models are shown with ldquoMrdquo corresponding to the specific hybridmodel (Table 1) CPB = central Patagonian Basin The legend is the same as for Figure 4 See a high-resolution version on Datashare 45wwwaapgorgdatashare


Figure 5 Continued

50 Ma (Bird 1998) to the Holocene (Bird 1998Liu et al 2010)We find that a continuous dynamiccontribution to the western North America (NAM)uplift has existed since the end of the Late Creta-ceous (Figure 5A) resulting from the movement ofthe Farallon slab away from this region similar to thefindings of adjoint geodynamic models (Spasojevicet al 2009)

Predicted Late Cretaceous flooding patternsin North America are closer to the observations inDYN-OFA calculations compared to estimatesmade solely from the changing age of sea floor(Figure 6 Table 3) further suggesting that the dy-namic effects of the Farallon subduction are im-portant in controlling NAM flooding and associateddeposition of thick hydrocarbon-bearing sedimen-tary sequences The predicted Late Cretaceousflooding is shifted toward the east compared tothat inferred (Figure 6) related to the assumed iso-static topography containing Laramide orogenEarly Cenozoic flooding is slightly better predictedwhen dynamic effects are considered (Figure 7)whereas the OligocenendashMiocene flooding is equallywell predicted in the OFA and DYN-OFA models(Figure 8) This suggests that the dynamic effectsprobably are a secondary factor in controlling Ce-nozoic flooding

Africa and ArabiaPaleogeographic maps (Guiraud et al 2005) showcontinuous flooding of northernmost Africa andeastern Arabia and a shifting pattern of marinetransgression and regression in the northern Africaninterior during the Late Cretaceous (Figure 4B)Dynamic earth models predict a long-wavelength

dynamic subsidence of northern and western Africaand Arabia (Figure 4B Table 2) For model M1dynamic subsidence is concentrated in three re-gions central western Africa northeastern Africaand Arabia broadly equivalent to inferred marinetransgression (Figure 4B) The dynamic subsidenceresults from the combined effect of the overallnorthward mantle flow of large positively buoyantlower mantle structures and isolated negativelybuoyant structures in the uppermost lower mantlebelow the central part of western Africa The com-plex inundation patterns in the interior of northernandwesternAfrica are not predicted by the dynamicmodels (Figures 6ndash8) This is expected because itwas proposed that African intracontinental seasdeveloped in response to the northwest-southeastndashtrending to north-southndashtrending intracontinentalrifting (Guiraud et al 2005) ignored in the currentDEMs The relationship between geologically in-ferred and predicted vertical motions indicates thatdynamic topography is a secondary factor control-ling regional sea level and that other effects such aseustatic changes caused by the changing age of seafloor andor rift tectonics (Guiraud et al 2005) area significantly more important factor

From the late Eocene to late Miocene a pro-nounced transgression in eastern Arabia is observedwhereas the eastern parts of the northern Africanmargin remained continuously flooded (Figure 5B)(Guiraud et al 2005) coincident with an overallfall in global sea level empirically suggesting sealevel controls beyond eustasy Dynamic earth mod-els show an overall dynamic subsidence in the last40my (Figure 4B Table 2) in the regions of easternnorthern Africa and Arabia The proposed sub-sidence and Cenozoic tilt of the Arabian platform


toward the northeast develop as a combined effect(1) increasing negative dynamic topography in theAlpine Tethys region from the increasing age of thereconstructed sea floor from the Late Cretaceous(Figure 1) (2) northeastward movement of Arabiatoward and over a highndashseismic velocity anomalypresently located between 300 and 800 km (186and 497mi) below easternmost Arabia and Persian


Gulf and (3) overall uplift of eastern Africa and theRed Sea caused by mantle upwelling associatedwith the African superplume (Lithgow-Bertelloniand Silver 1998 Daradich et al 2003)

Since the Eocene substantial uplift of easternand southern Africa is predicted with models thathave a smaller ratio between the transition zoneand lower mantle viscosities (Figure 4B model

Figure 6 Late Cretaceous global sea level predictions Paleogeographic reconstructions at 80 Ma (Smith et al 1994) (A) and 90 Ma (B)(Blakey 2010) showing reconstructed marine regions in light blue and land regions in brown (C) DYN-OFA flooding predictions (80 Ma)accounting for the changing age of the ocean floor dynamic topography and geoid usingmodel M3 (Table 2) (D) OFA flooding predictions(80 Ma) accounting only for the changing age of the ocean floor using the half-space model without flattening (Parsons and Sclater 1977)(E) Difference between dynamic topography in the past (80 Ma) and present-day predicted dynamic topography for model M3 The darkblue lines on panels C and D show a predicted shoreline See a high-resolution version on Datashare 45 wwwaapgorgdatashare

Figure 7 Early Cenozoic global sea level predictions Paleogeographic reconstructions at 60 Ma (A) (Smith et al 1994) and 65 Ma (B)(Blakey 2010) showing reconstructed marine regions in light blue and land regions in brown (C) DYN-OFA flooding predictions (60 Ma)accounting for the changing age of the ocean floor dynamic topography and the geoid and using model M3 (Table 2) (D) OFA floodingpredictions (60 Ma) accounting only for the changing age of the ocean floor using the half-space model without flattening (Parsons andSclater 1977) (E) Difference between dynamic topography in the past (60 Ma) and present-day predicted dynamic topography for model M3The dark blue lines on panels C and D show a predicted shoreline See a high-resolution version on Datashare 45 wwwaapgorgdatashare

Table 3 Relative Importance of Factors Controlling Regional Sea Level

Continent

Late Cretaceous


Spasojevic


North America

Dynamic + OFAdagger

OFA + Dynamic

OFA (dynamic secondary)

Northern Africa India and South America

Eurasia

OFA

OFA

OFA

Australia

Dynamic

Dynamic

Dynamic

The interpretation is complicated by unknown paleotopography andor difference in paleogeographic reconstructionsSubduction-related dynamic subsidence might be important in the Alpine Tethys regiondaggerOFA = ocean floor age

and Gurnis 2051

M1 Appendix 3 see AAPG Datashare 45 at wwwaapgorgdatashare) related to the upwelling oflow-density material in the lower mantle belowAfrica (eg Lithgow-Bertelloni and Silver 1998Gurnis et al 2000a) The uplift of eastern Africaoccurs continuously from the late Eocene to thepresent (Appendix 3) and the uplift in southernAfrica mostly develops in the last 20 my The com-puted southward progression of the African uplift


is consistent with geologic studies that suggestthat southern Africa experienced Miocene uplift(Partridge and Maud 1987) whereas the uplift andrifting in easternAfrica initiated earlier between 20and 30 Ma (Burke 1996)

The flooding for the entire period since the LateCretaceous is underpredicted using both OFA andDYN-OFA calculations (Figures 6ndash8) which wethink is at least partially a result of significant change

Figure 8 Global sea level predictions (30 Ma) Paleogeographic reconstructions at 30 Ma (A) (Smith et al 1994) and 35 Ma (B) (Blakey2010) showing reconstructed marine regions in light blue and land regions in brown (C) DYN-OFA flooding predictions (30 Ma) accountingfor the changing age of the sea floor dynamic topography and the geoid and using model M3 (Table 2) (D) OFA flooding predictions(30 Ma) accounting only for the changing age of the sea floor using the half-space model without flattening (Parsons and Sclater 1977)(E) Difference between dynamic topography in the past (30 Ma) and present-day predicted dynamic topography for model M3 The darkblue lines on panels C and D show a predicted shoreline See a high-resolution version on Datashare 45 wwwaapgorgdatashare

in continental isostatic topography which is not ac-counted for by ourmethod because the overall upliftof Africa occurs at least since the Eocene (Bond1979 Gurnis et al 2000a) The onset of basin in-version and folding related to the collision of African-Arabian and Eurasian plates initiated in the lateSantonian affecting both northern African-Arabianplate margin and central African rifts (Guiraud andBosworth 1997 Guiraud et al 2005) A briefcompressional phase around the end of the Cre-taceous and a major plate-scale compression in theearly late Eocene (Guiraud et al 2005) have bothstrongly affected the development of topographyin the region As the flooding predictions are com-plicated by unknown isostatic changes we do notfurther consider the relative importance of sea levelchange factors for northern Africa (Table 3)

EurasiaIn western Siberia and the Alpine Tethys region atransgression in the Cretaceous (Figure 4C) fol-lowed by an overall regression in the Cenozoic(Figure 5C) was observed equivalent to expectedglobal sea level trends Cretaceous differential dy-namic topography (Figure 4C) ranges from sub-sidence on the order of 100 to 200 m (328ndash656 ft)(model M1) to neutral values (model M6) over alarge part of Eurasia suggesting an absence of or alittle dynamic control on vertical motions (Table 2)Similar to the observed trends the Alpine Tethysregion is predicted to have high-amplitude dynamicsubsidence (Figure 4C) related to the increasingage of subducted oceanic lithosphere (Figure 1) Theregion of Southeast and eastern Asia is predicted toexperience overall uplift related to a decreasing ageof the subducted oceanic lithosphere in the east andthe closing of the easternmost part of the Tethys Seaby early Cenozoic in the south (Figure 1) For theeastern Asia margin the eventual total consumptionof the Izanagi plate by 60 Ma is observed (Figure 1)(Whittaker et al 2011) resulting in a progressivesubduction of younger sea floor During the Ce-nozoic models predict subsidence in easternand Southeast Asia and the Alpine Tethys region(Figure 5C) related to the increasing age of sub-ducted oceanic lithosphere (Figure 1) In terms ofpredicted sea level OFA models predict well the

observed patterns of flooding (Figures 6ndash8 Table 3)and the introduction of dynamic effects in DYN-OFA models does not strongly perturb the inun-dation This leads us to suggest that mantle dy-namics is secondary in controlling platform flooding(Table 3) with a possible exception of the AlpineTethys region

However more recent geomorphic indicatorsmight indicate a present-day influence of mantledynamics in Siberia Analysis of regional incisionand lateral shifts of rivers in theWest Siberian Basinindicate that it is currently tilting over a long wave-length (Allen and Davies 2007) The tilting isgenerally down to the east away from the Uralsand central Asian mountains toward the Siberiancraton Significant drainage shifts exist in the riversystems of the Yenisei Ob and Irtysh caused bythis long-wavelength tilt of theWest Siberian Basinand surroundings (Allen andDavies 2007) Subtlefaulting of the basement located under the thickbasin fill of Mesozoic and lower Cenozoic sedi-ments has been proposed as a cause of this long-wavelength deformation (Allen andDavies 2007)speculated to potentially be a far-field effect of theIndia-Eurasia collision (Allen and Davies 2007)The hybrid models predict overall increasing dy-namic subsidence in the general west-to-east di-rection (Figure 9) for all models with the totalamplitude dependent on the tomography and vis-cosity The gradient of differential dynamic topog-raphy has the same direction as the inferred direc-tion of tilting from geomorphic analysis (Allen andDavies 2007) (Figure 9C) and we propose that itis related to the dynamic effects of mantle convec-tion Two thermal anomalies in the lower mantle(Figure 9C) cause the increasing dynamic subsi-dence and tilting of the Western Siberian Basinfrom west to east A warm upwelling structure islocated on the western edge of the basin causingit to be dynamically higher than the eastern edgeof the basin which is associated with a large coldmantle downwelling located in the lower mantle

Southeast AsiaThe Sundaland core was subaerially exposed dur-ing the lateOligocene but was significantly floodedby the middle Miocene (Hall and Morley 2004)


(Figure 5D) Dynamic earth models predict anoverall dynamic subsidencewithin the southern partof Sundaland and Borneo (Figure 5D Table 2) A


large fastndashseismic velocity anomaly exists belowSoutheast Asia between 400 and 1200 km (249and 746 mi) (Masters et al 2000 Ritsema et al

Figure 9 Inferred tilting of theWest Siberian Basin based ondifferential dynamic tomographyDifferential dynamic topographyfor models M1 and M3 in theperiod 0 to 30 Ma (A) and 0 to10 Ma (B) (C) East-west crosssection (location shown on panelA) showing dynamic topographytrends for models M1 to M6(bottom panel) thermal buoy-ancy field (middle panel) sur-facendashtondashcore-mantle boundary(CMB) and drainage-shift azimuthdirections (Allen and Davies2007) (top panel azimuth isshown in 60deg bins with respectto north)

2004) As the mantle flow is reversed in SBI thisanomaly rises upward with a large part locatedwithin the upper mantle in the Eocene resultingin predicted dynamic subsidence Integrated for-ward dynamic earth models suggest that South-east Asia experienced an event similar to a mantleavalanche when a significant part of subductingslabs penetrated into the lower mantle resultingin a significantly increased dynamic topographylow potentially causing observed marine transgres-sion The predicted amplitudes of subsidence aredependent on tomography and viscosity and haveamplitudes up to 500 m (1640 ft) within 1500 km(932 mi) of the subduction zones (Figure 5D)whereas higher subsidence amplitudes are predictedalong the subduction zone itself The wavelength ofpredicted dynamic subsidence is much larger thanthat associated with the anomalous Sundaland ba-sins These basins have relatively small widths andpostrift sedimentary sequences with thicknessesup to 12 km (7mi) (Hall andMorley 2004Morleyand Westaway 2006) so that processes beyondsimple rifting must be invoked to explain their for-mation The amplitude of predicted dynamic sub-sidence in the Sundaland interior (Figure 5D) has asimilar magnitude as the dynamic subsidence esti-mated based on borehole subsidence in SoutheastAsia (Wheeler and White 2002) suggesting thatdynamic topography might have had an importantinfluence on subsidence and overall flooding duringthe Neogene affecting the evolution of hydrocar-bon reservoirs in the region The zone of predicteddynamic subsidence is relatively well correlatedwiththe zone of inferred Sundaland flooding Howeverwe predict that Borneo and Java subsided oppositeto their inferred trends which suggests that verti-calmotions are controlled bymore regional tectonicsin this region We propose that dynamic subsidencebetween 30 and 10Mamight be contributing to theoverall platform drowning between the late Oligo-cene and middle Miocene

South AmericaIn South America a north-southndashtrending inte-rior seaway developed between 90 and 50 Ma(Figure 4D) with a relative stability of the inun-dated region in the early Cenozoic (Blakey 2010)

Dynamic earth models predict a region of LateCretaceous dynamic subsidence in western SouthAmerica with amplitudes from 100 to 500 m(328ndash1640 ft) (Figure 4D Table 2) and a wave-length larger than one inferred paleogeographicallyAlthough the development of the South AmericaInterior seaway has been previously related toflexural loading by the Andes (Hoorn et al 1995)we postulate that a component resulting fromincreasing dynamic topography may also exist(Figure 6D)

Global Cenozoic paleogeographic reconstruc-tions imply relatively isolated regions of verticalmotions in the north and northeast (Figure 5E)with most of South America being stable How-ever more regionally focused studies indicate thatcentral Patagonia has been experiencing Neogeneuplift (Guillaume et al 2009) and that the in-terior Pebas Sea developed in the Miocene (Hoornet al 1995 Shepard et al 2010) (Figure 5E)Dynamic earth models indicate the overall Neo-gene uplift of Patagonia concentrated in the cen-tral Patagonian Basin in the last 20my (Figure 5ETable 2) From the Cretaceous to Cenozoic sub-sidence migrated eastward in the plate frame andis concentrated in the northern part of central SouthAmerica in the last 20my formodelM1(Figure 5E)The continuous dynamic uplift of the Andes andwestern South America emerge from the modelsfor the last 20 my (Figure 5E Table 2)

The observed pattern of the eastward migra-tion of subsidence from the Late Cretaceous to thepresent is related to the geometry of the subductingFarallon (Nazca slab) Currently the slab is imagedas a continuous structure in seismic tomography upto mid-mantle depths under the whole of SouthAmerica As the plate motions and mantle flow arereversed in an SBI from 0 Ma to the Late Creta-ceous the high-density material moves upwardand it is located under western South America inthe Late Cretaceous which causes subsidence Themodeled pattern of vertical motions associated withthe subducting slab below South America is equiv-alent to the scenario described for North America

Late Cretaceous and Paleocene flooding areoverpredicted caused by the relatively low iso-static topography of South America in the flooded


regions (Figures 6 7) This might imply that theisostatic topography of SouthAmerica has changedsince the Late Cretaceous and that the continenthas experienced overall subsidence This putativesubsidence could be attributed to active subduc-tion which creates a progressive dynamic topog-raphy low in South America (Appendix 2) Flood-ing at 30 Ma is best predicted by OFA models(Figure 8) whereas the inclusion of dynamic ef-fects in DYN-OFA models overpredicts floodingHowever if more regional reconstructions (egHoorn 1996) including those that include theinterior Pebas Sea are considered the DYN-OFApredictions yield a better match to the observa-tions The dynamics of Farallon (more recentlyNazca) plate subductionmight have controlled theMiocene switch in the direction of the AmazonRiver at 11 Ma (Shepard et al 2010) A similaroutcome emerges from our DEMs but more de-tailed stratigraphic data and paleogeographic in-formation are needed to clarify the link

AustraliaAustralia was significantly more subaerial duringthe Cretaceous than now remaining high from theLate Cretaceous until early Cenozoic (Figure 4E)The models predict an overall neutral to positivedynamic topography change from 80 to 60 Ma(Figure 4E) Modeled dynamic uplift is consistentwith the movement of Australia away from a sub-ducted Gondwanaland slab as it was previouslyproposed (Gurnis et al 1998 Gurnis et al 2000b)Australia experienced an overall marine transgres-sion over a large part of the Cenozoic (Figure 5F)Dynamic models predict an overall dynamic sub-sidence over the Cenozoic (Figure 5F Table 2) in-creasing from south-southwest to north-northeastcausing an overall tilt of Australia The Cenozoictilt has been inferred from a detailed study of pa-leoshorelines (Sandiford 2007 DiCaprio et al2009) and was also previously attributed to thedynamic topography and geoid (Sandiford 2007)and dynamic topography low resulting from de-veloping subduction zones toward the north andeast (DiCaprio et al 2009 DiCaprio et al 2010)Dynamic earth models support the findings of


DiCaprio et al (2009) and DiCaprio et al (2010)showing that dynamic subsidence is caused bythe development of subduction zones in the sur-rounding regions In terms of prediction of sealevel in the Late Cretaceous and early Cenozoic(Figures 6 7) DYN-OFAmodels yield predictionsmore similar to the observations suggesting thatdynamics are important in controlling the regionalsea level in Australia (Table 3)

Global Sea Level Estimates

With the volume of ocean water assumed constantfor over millions of years global sea level on geo-logic time scales results from a change of oceanbathymetry caused by variations in ocean floor agedynamic topography in oceanic and flooded con-tinental areas emplacement of oceanic plateausand sedimentation The predicted global sea levelby DEMs shows an overall decrease from the LateCretaceous to the present with a maximum at80 Ma (Figure 10) The predicted sea level fallamplitude from changing ocean floor age rangessignificantly between 304 and 528 m (997 and1732 ft) for different cooling models (Figure 10A)a half space without flattening (OFA-HS) givesthe largest prediction whereas global depth andheat flowmodel (GDH) gives the smallest The sealevel decrease fromOFA is caused by the deepeningof bathymetry from the aging sea floor from the LateCretaceous to the present day (Figure 1) as dis-cussed previously (eg Kominz 1984 Kominzet al 2008 Muumlller et al 2008a) In isolationchanging dynamic topography in oceanic regionscauses a sea level rise since the Late Cretaceouswith amplitudes between 100 and 200m (328 and656 ft) (Figure 10A) The contribution of dynamictopography in the flooded continental areas is muchsmaller than that in the oceanic regions havingmaximum amplitudes from ndash10 to 80 m (ndash33ndash262 ft) (Figure 10A) of sea level rise The ampli-tudes of dynamic topography contributions are thelargest at 70 Ma and they significantly decreaseduring the Cenozoic (Figure 10A) The dynamictopography contribution to global sea level in oce-anic regions is mostly driven by positive dynamic

topography from low-densitymantle upwellings Aslow-density anomalies rise from the LateCretaceousto the present the amplitude of dynamic topogra-phy increases resulting in sea level rise which offsetsoverall sea level fall associated with the increasingage of the sea floor (Figure 10A) The contribution ofhigh-density downwelling anomalies in the oceanic

regions to the overall sea level change is less signifi-cant because the regions of dynamic topographylows in the oceans are much more spatially limited(Appendix 2) The contribution of dynamic topog-raphy in the flooded continental regions is mostlydriven by dynamic topography lows which have anoverall increasing mean amplitude (Figure 10A)

Figure 10 Global sea levelpredictions (A) Predictions ac-counting for the changing age ofthe sea floor using half-space(OFA-HS) half-space model withflattening (OFA-HSF) and globaldepth and heat flow (GDH)ndash1model (OFA-GDH) and DYN-OFA models accounting for thechanging age of the sea floordynamic topography and geoidThe blue area shows the range ofsea level values estimated fromDYN-OFA models The gray or-ange and red areas show therange of dynamic topography inoceanic areas dynamic topog-raphy in flooded continentalareas and geoid contribution toglobal sea level respectively(B) Final predictions of sea levelfrom DYN-OFA models (bluearea DYN-OFA-C) accounting forcorrections resulting from sedi-mentations (yellow area Sed-Corr) and emplacement of largeigneous provinces (orange areaLIPSCorr) (C) Comparison be-tween the final prediction of sealevel from DYN-OFA-C models(blue area) and previously pub-lished sea level curves DynTopo = dynamic topographyDyn Topo (Cont) = dynamictopography (continents) Sed-Corr = sediment correctionLIPSCorr = correction for largeigneous provinces emplacement

Spas

ojevic and Gurnis 2057

The contribution of the geoid to global sea level ismuch smaller than that from dynamic topography(Figure 10A) because the ratio between geoid andtopography is small (Figure 2)

Therefore the sea level estimates for DYN-OFAmodels before accounting for sedimentationand oceanic plateau emplacement havemaximumamplitudes at 80 Ma between 330 and 365 m(1083 and 1198 ft) (Figure 10A) for the six DEMsThe predicted range of dynamic sea level for DYN-OFAmodels lies between the predictions for the twocooling models that include flattening (Figure 10AOFA-HSF OFA-GDH) This is not unexpectedif dynamic topography is the ultimate cause for theflattening of the old sea floor empirically includedin the age-to-depth relationships forOFA-HSF andOFA-GDH Sedimentation and emplacement oflarge igneous provinces result in an overall sea levelrise from the Late Cretaceous to the present (Muumllleret al 2008b) Maximum amplitudes of sea levelchange caused by sedimentation and igneous prov-inces are much smaller at 60 and 27 m (197 and89 ft) respectively so that the combined effect ofall of the processes still results in an overall fall witha maximum amplitude of up to 286 m (938 ft)since the Late Cretaceous (Figure 10B DYN-OFA-C) The DYN-OFA-C global sea level is within therange of those previously proposed (Figure 10C)

DISCUSSION

We use DEMs to better understand regional andglobal sea levels from the Late CretaceousWe findthat oceanic regions on average have a positive dy-namic topography which increased from the LateCretaceous to the present because of the overall riseof low-density upwelling anomalies resulting in asea level rise with the estimated amplitude rangingbetween 100 and 200m (328 and 656 ft) in the last90 my (Figure 10) Conrad and Husson (2009)use forward convection models to estimate Ceno-zoic sea level change caused by a sea floor dynamictopography of 90 plusmn 20 m (295 plusmn 66 ft) which issimilar to the estimated amplitudes in DEMs of90 to 140m (295ndash459 ft) of sea level change in thelast 60myWe also find that dynamic topography in


flooded continental areas has contributed to sea levelchange by introducing up to 80 m (262 ft) of sealevel rise since the Late Cretaceous (Figure 10)Conrad and Husson (2009) also find that dynamictopography in continental areas has an effect on sealevel inwhich the rate of change ismore than threetimes smaller than the one caused by oceanic dy-namic topography The sea level rise caused bydynamic topography along with the rise resultingfrom the emplacement of oceanic plateaus and sed-imentation partially offsets the larger amplitude sealevel fall because of increases in OFA (Figure 10)The overall fall of the sea level caused by the changein OFA and spreading-ridge crustal production aremostly driven by changes in the Pacific Ocean asthe largest drop in the global crustal production ofapproximately 50 occured between 65 and 60Macaused by the subduction of the Izanagi-Pacificspreading ridge (Muumlller et al 2008b) (Figure 1)and followed by a subsequent ridge subductionbeneath the Americas (Muumlller et al 2008b)

When calculating sea level amplitudes we pre-fer to use the half-space cooling model so as not todouble count the effect of sea floor flattening (Kidoand Seno 1994) which can arise when dynamictopography is considered from a mantle flow model(Muumlller et al 2008b) However if we implementedcooling models that include flattening (eg OFA-HSF and OFAndashGDH) the final sea level wouldhave the same trend butwith a significantly smalleramplitude (Figure 10A) We predict the total globalsea level fall from the Late Cretaceous with a max-imum amplitude of up to 286 m (938 ft) reachedat 80Ma LateCretaceous sea levels are on the highend of previous estimates (Figure 10C) (Watts andSteckler 1979 Kominz 1984 Haq et al 1987Haq and Al-Qahtani 2005 Miller et al 2005Kominz et al 2008Muumlller et al 2008b)with thevalues being similar to the ones inferredbyHaq andAl-Qahtani (2005) During the Cenozoic the pro-posed sea levels were generally lower than the onesderived using the global correlation of stratigraphicsequences (Watts and Steckler 1979 Haq and Al-Qahtani 2005) but higher than the ones derivedbased on the backstripping of sedimentary sectionsat five boreholes located on the New Jersey coastalplain (Miller et al 2005) and within the range of

other studies (Figure 10C) Our study demonstratesthe importance of accounting for the effects of dy-namic topography when calculating global sealevel but the actual amplitudes should be moreconstrained whenmore comprehensive stratigraphicand geologic constraints are used in the future

Whereas positive dynamic topography associ-ated with mantle upwellings represents the maindynamic factor controlling global sea level thedynamic topography associatedwith subducting orpreviously subducted slabs is also an importantfactor that controls the pattern of continental flood-ing and relative vertical motion and impacts sev-eral regions with important hydrocarbon accumu-lations We find that the Farallon slab in NorthAmerica creates a dynamic topography low thatcontrols Cretaceous flooding and subsidence inthe WIS similar to previous inverse models (Liuet al 2008 Spasojevic et al 2009) As NorthAmericamoves westward and the Farallon slab sinksinto the lower mantle during Cenozoic the West-ern Interior experiences a dynamic uplift and east-ern North America experiences an overall subsi-dence The Farallon andNazca plates continuouslysubduct in South America contributing to an in-terior flooding developed from 90 to 50 Ma andpotentially contributing to aMiocene switch in thedirection of drainage from a predominantly north-ward direction to a formation of the eastward-draining Amazon River (Shepard et al 2010) andaffecting the evolution of petroleum systems of theAtlanticmargin of SouthAmerica Subduction in theAlpine Tethys region causes an overall subsidenceand flooding of the Mediterranean region and a Ce-nozoic dynamic subsidence and flooding of northernAfrica and Arabia A mantle-avalanchendashtype eventin Southeast Asia between the late Oligocene andthe Miocene resulted in the dynamic subsidenceand potential drowning of Sundaland Develop-ment of subduction zones north and east fromAustralia contributes to an overall tilting towardnorth-northeast that potentially explains discrepan-cies betweenAustralian and global sea levels duringthe Cenozoic as previously found (DiCaprio et al2009 DiCaprio et al 2010)

In the Cenozoic we find that mantle upwell-ing anomalies have had a relatively limited effect

on the pattern of the large-scale vertical motion ofcontinents and regional sea level for most conti-nents except Africa The African superplume hascaused an uplift in eastern and southern Africasince the Eocene and it contributes to the Ceno-zoic tilting down to north-northeast of ArabiaMid- and lower mantle upwelling located to thewest of the Siberian Basin and adjacent to the eastSiberian lower mantle downwelling resulting inthe west-east tilting of Siberia that potentiallyresults in the shifting of drainage patterns in thisregion

Sea level predictions are affected by some un-derlying assumptions (1) time-invariant isostatictopography (2) the neglect of continental marginextension and (3) constant ocean volume used insea level calculations First since paleoelevation rep-resents one of the most difficult variables to re-construct from the geologic record (Garzione et al2008) and detailed maps of the past global topog-raphy have yet to be assembled we use present-day topography rotated with plate motions andcorrected for dynamic topography As crustal col-umns experienced significant changes resulting fromorogeny such as the Laramide Andean AlpineandHimalayan orogenies (Spencer 1974 Saleeby2003 English and Johnston 2004 Ghosh et al2006) the quality of sea level predictions is sig-nificantly affected in orogenic regions Second be-cause continental margin extension has been esti-mated to result in approximately 20 m (sim66 ft)of sea level change since the beginning of Pangeabreakup (Kirschner et al 2010) significantly smallerthan the overall sea level change (Figure 10) itsneglect is reasonable Third we assume constantwater volume as we examine sea level changes attime scales greater than 1mywhich is longer thanthe ocean volume fluctuations from glacial cyclesoccurring on time scales of hundreds of thousandsof years or shorter (Kominz 2001)

The main limitations of the current method arethe use of a single average vertical viscosity struc-ture and seismic velocity-to-density scaling thetreatment of subduction zones the use of one pa-leogeographic reconstruction and the limited as-similation of stratigraphic data A significant un-certainty remains regarding the viscosity profile of


the mantle with suggestions that significant dif-ferences might exist between viscosity structuresin different regions (eg Liu et al 2008 Schellartet al 2009 Spasojevic et al 2010a) or that themantle on average might have the same verticalviscosity profile (eg van der Meer et al 2010)A single seismic velocity-to-density scaling devisedfrom the fit to global geoid (Spasojevic et al 2010)is used for all models although it was suggestedthat the 3-D velocity-to-density scalingmodel whichmight differ for upwelling and downwelling anom-alies (Forte 2007 Simmons et al 2009) might bemore appropriate As both viscosity structure andvelocity-to-density scaling influence dynamic to-pography and in turn sea level estimates in thefuture it would be important to test the sensitivityof regional and global sea levels to the variations ofthese parameters Definition of subduction zonesin the models can be improved by introducingmuch higher resolution variable viscosity slabsand surrounding mantle wedges possibly imple-menting methods at ultra-high resolution (1 km[06mi] where needed) (Stadler et al 2010) andintroducing a more realistic treatment of diffusionin an adjoint solution to the equations instead ofthe current SBI (Liu and Gurnis 2008) The im-portance of strong viscosity variations has beendemonstrated in the regional models of the Tonga-Kermadec and Aleutian subduction zones (Billenet al 2003 Billen and Gurnis 2001) which showthat significant lateral viscosity variations are re-quired between upper mantle slabs and adjacentmantle wedges to match the geoid dynamic to-pography and stress patterns in subduction zoneswhereas strong slabs also give rise to a better matchto observedpresent-day platemotions (Stadler et al2010) It has been shown that the use of differentreconstructions in the west Pacific has substantiallyaffected sea level estimates since theCenozoic (Xuet al 2006) in the future one can test alternativepaleogeographic reconstructions and their effect onpatterns of vertical motions and relative and globalsea levels Inclusion of more stratigraphic data in-cluding more comprehensive borehole data sedi-ment isopach maps and detailed regional paleo-shoreline reconstructions might impose betterconstraints on the history of the vertical motion of


continents and enable us to infer mantle viscositystructure on a finer scale

CONCLUSIONS

We present the results of DEMs that combine in-verse and forward methods for solving for mantleconvection and assimilate plate tectonic reconstruc-tions into mantle convection models with verticallyand laterally varying viscosity We self-consistentlyaccount for the changing age of the sea floor dy-namic topography and the geoid with the goal tobetter understand the verticalmotion of continentsand the regional and global sea levels We find thatboth the changing age of the ocean floor and dy-namic topography represent important factors con-trolling sea level Predicted sea level shows an overallfall from the Late Cretaceous to the present with amaximum amplitude of 286 m (938 ft) at 80 MaThe largest contribution to the sea level fall comesfrom the reduced oceanic crust production from theLate Cretaceous to the present inferred from oceanage grids The dynamic topography in oceans andflooded continental regions result in sea level risewith maximum amplitudes varying between 100and 200m (328 and 656 ft) and ndash10 and 80m (ndash33and 262 ft) respectively reducing the overall am-plitude of sea level fall inferred from age grids Themovement of continents with respect to large-scaledynamic topography causes the dynamic subsi-dence and uplift of continents controls regional sealevel and affects the hydrocarbon basins for NorthAmerica and South America from the Late Cre-taceous to the present Australia during the Ceno-zoic andNorthAfrica andArabia since theEoceneEastern and South Africa uplift dynamically in thelast 20 to 30 my whereas Siberia and Australiaexperience Cenozoic tilting

REFERENCES CITED

Allen M B and C E Davies 2007 Unstable Asia Activedeformation of Siberia revealed by drainage shifts BasinResearch v 19 p 379ndash392 doi101111j1365-2117200700331x

Billen M I and M Gurnis 2001 A low-viscosity wedge in

subduction zones Earth and Planetary Science Lettersv 193 no 1ndash2 p 227ndash236 doi101016S0012-821X(01)00482-4

Billen M I M Gurnis and M Simons 2003 Multiscaledynamics of the Tonga-Kermadec subduction zoneGeophysical Journal International v 153 no 2 p 359ndash388 doi101046j1365-246X200301915x

Bird P 1998 Kinematic history of the Laramide orogeny inlatitudes 35 to 49degN western United States Tectonicsv 17 no 5 p 780ndash801 doi10102998TC02698

Blakey R 2008 Gondwana paleogeography from assemblyto breakup A 500-my odyssey in C R Fielding T DFrank and J L Isbell eds Resolving the late Paleozoicice age in time and space p 1ndash28

Blakey R 2010 Global paleogeographic maps httpjanuccnauedu~rcb7globaltext2html (accessed March2010)

Bond G 1976 Evidence for continental subsidence in NorthAmerica during Late Cretaceous global submergenceGeology v 4 no 9 p 557ndash560 doi1011300091-7613(1976)4lt557EFCSINgt20CO2

Bond G C 1979 Evidence for some uplifts of large magni-tude in continental platforms Tectonophysics v 61no 1ndash3 p 285ndash305 doi1010160040-1951(79)90302-0

Boyden J A R D Muller M Gurnis T H Torsvik J AClark M Turner H Ivey-Law R J Watson and J SCannon 2011 Next-generation plate-tectonic recon-structions using GPlates in G R Keller and C Barueds Geoinformatics Cyberinfrastructure for the solidearth sciences United Kingdom Cambridge UniversityPress p 95ndash113

Burgess P M M Gurnis and L Moresi 1997 Formation ofsequences in the cratonic interior of North America byinteraction between mantle eustatic and stratigraphicprocesses Geological Society of America Bulletin v 109no 12 p 1515ndash1535 doi1011300016-7606(1997)109lt1515FOSITCgt23CO2

Burke K 1996 The African plate South African Journal ofGeology v 99 p 341ndash409

Conrad C P and M Gurnis 2003 Seismic tomographysurface uplift and the breakup of Gondwanaland Inte-grating mantle convection backwards in time Geochem-istry Geophysics Geosystems v 4 no 3 p 1031doi1010292001GC000299

Conrad C P and L Husson 2009 Influence of dynamictopography on sea level and its rate of change Litho-sphere v 1 no 2 p 110ndash120 doi101130L321

Cook T D and A W Bally 1975 Stratigraphic atlas ofNorth andCentral America Princeton New Jersey Prince-ton University Press 272 p

Daradich A J X Mitrovica R N Pysklywec S D Willettand A M Forte 2003 Mantle flow dynamic topographyand rift-flank uplift of Arabia Geology v 31 no 10p 901ndash904 doi101130G196611

DiCaprio L M Gurnis and R D Muller 2009 Long-wavelength tilting of the Australian continent since theLate Cretaceous Earth and Planetary Science Lettersv 278 no 3ndash4 p 175ndash185 doi101016jepsl200811030

DiCaprio L M Gurnis R D Muller and E Tan 2010Mantle dynamics of continent-wide Cenozoic subsidenceand tilting of Australia Lithosphere v 3 p 311ndash316doi101130L1401

English J M and S T Johnston 2004 The Laramide orog-eny What were the driving forces International Geol-ogy Review v 46 no 9 p 833ndash838 doi1027470020-6814469833

Forte A M 2007 Constraints on seismic models from otherdisciplines Implications for mantle dynamics and com-position in B Romanowicz and A M Dziewonski edsTreatise of geophysics Amsterdam Elsevier p 805ndash854

Garzione C N G D Hoke J C Libarkin S Withers BMacFadden J Eiler P Ghosh and A Mulch 2008Rise of the Andes Science v 320 no 5881 p 1304ndash1307 doi101126science1148615

Ghosh P C N Garzione and J M Eiler 2006 Rapid upliftof the Altiplano revealed through C13-O18 bonds in pa-leosol carbonates Science v 311 no 5760 p 511ndash515doi101126science1119365

Goes S and S van der Lee 2002 Thermal structure of theNorth American uppermost mantle inferred from seismictomography Journal of Geophysical Research v 107no B3 p 2050 doi1010292000JB000049

Grand S P 2002 Mantle shear-wave tomography and thefate of subducted slabs Philosophical Transactions ofthe Royal Society of London 360 p 2475ndash2491

Guillaume B J Martinod L Husson M Roddaz and RRiquelme 2009 Neogene uplift of central eastern Pata-gonia Dynamic response to active spreading ridge sub-duction Tectonics v 28 p TC2009 doi1010292008TC002324

Guiraud R and W Bosworth 1997 Senonian basin inver-sion and rejuvenation of rifting in Africa and ArabiaSynthesis and implications to plate-scale tectonics Tec-tonophysics v 282 no 1ndash4 p 39ndash82 doi101016S0040-1951(97)00212-6

Guiraud R W Bosworth J Thierry and A Delplanque2005 Phanerozoic geological evolution of northernand central Africa An overview Journal of AfricanEarth Sciences v 43 no 1ndash3 p 83ndash143 doi101016jjafrearsci200507017

Gurnis M 1990 Bounds on global dynamic topography fromPhanerozoic flooding of continental platforms Naturev 344 no 6268 p 754ndash756 doi101038344754a0

Gurnis M 1993 Phanerozoic marine inundation of conti-nents driven by dynamic topography above subductingslabs Nature v 364 no 6438 p 589ndash593 doi101038364589a0

Gurnis M R Muumlller and L Moresi 1998 Cretaceous ver-tical motion of Australia and the Australian-Antarcticdiscordance Science v 279 no 5356 p 1499ndash1504doi101126science27953561499

Gurnis M J Mitrovica J Ritsema and H van Heijst2000a Constraining mantle density structure using geo-logical evidence of surface uplift rates The case of theAfrican superplumeGeochemistryGeophysicsGeosys-tems v 1 no 7 p 1020 doi1010291999GC000035

Gurnis M L Moresi and R D Muller 2000b Models of


mantle convection incorporating plate tectonics The Aus-tralian region since the Cretaceous in M A RichardsR Gordon and R van der Hilst eds The history anddynamics of global plate motions Washington DCAmerican Geophysical Union p 211ndash238

Gurnis M M Turner S Zahirovic L DiCaprio S SpasojevicR D Muumlller J Boyden M Seton V C Manea andD J Bower Plate tectonic reconstructions with conti-nuously closing plates Computers and Geosciences 38p 35ndash42

Hager B H 1984 Subducted slabs and the geoid Con-straints on mantle rheology and flow Journal of Geo-physical Research v 89 no B7 p 6003ndash6015 doi101029JB089iB07p06003

Hall R and C KMorley 2004 Sundaland basins in P CliftW Kuhnt P Wang and D Hayes eds Continent-ocean interactions within East Asianmarginal seas Amer-ican Geophysical Union Washington DC AmericanGeophysical Union Geophysical Monograph Series 149p 55ndash85

Hallam A 1992 Phanerozoic sea level changes New YorkColumbia University Press 266 p

Haq B U and A M Al-Qahtani 2005 Phanerozoic cyclesof sea level change on the Arabian platform GeoArabiav 10 no 2 p 127ndash160

Haq B U J Hardenbol and P R Vail 1987 Chronology offluctuating sea levels since the Triassic Science v 235no 4793 p 1156ndash1167

Hays J D and W C Pitman 1973 Lithospheric plate mo-tions sea level changes and climatic and ecological con-sequences Nature v 246 no 5427 p 18ndash22 doi101038246018a0

Hoorn C 1996 Miocene deposits in the Amazonian fore-land basin Science v 273 no 5271 p 122ndash123 doi101126science2735271122

Hoorn C J Guerrero G A Sarmiento and M A Lorente1995 Andean tectonics as a cause for changing drainagepatterns in Miocene northern South America Geologyv 23 no 3 p 237ndash240 doi1011300091-7613(1995)023lt0237ATAACFgt23CO2

Jordan T H 1988 Structure and formation of the continen-tal tectosphere Journal of Petrology Special Volume e1doi101093petrologySpecial_Volume111

Kido M and T Seno 1994 Dynamic topography com-pared with residual depth anomalies in oceans and impli-cations for age-depth curves Geophysical Research Let-ters v 21 no 8 p 717ndash720 doi10102994GL00305

Kirschner J P M A Kominz and K E Mwakanyamale2010 Quantifying extension of passive margins Impli-cations for sea level change Tectonics v 29 p TC4005

Kominz M A 1984 Oceanic ridge volumes and sea levelchange An error analysis in J Schlee ed Interregionalunconformities and hydrocarbon accumulation AAPGMemoir v 36 p 109ndash127

Kominz M A 2001 Paleooceanography Variations in sealevel (MS 255) in J Steele S Thorpe and K Turekianeds Encyclopedia of ocean sciences San Diego Cali-fornia Academic Press p 2605ndash2613

Kominz M A J V Browning K G Miller P J SugarmanS Mizintseva and C R Scotese 2008 Late Cretaceous


to Miocene sea level estimates from the New Jersey andDelaware coastal-plain coreholes An error analysis Ba-sin Research v 20 no 2 p 211ndash226 doi101111j1365-2117200800354x

Langford R P G E Wilford E M Truswell and A R Isern1995 Paleogeographic atlas of Australia Canberra Aus-tralia Cainozoic Australian Geological Survey Organiza-tion 38 p

Lithgow-Bertelloni C andM Gurnis 1997 Cenozoic subsi-dence and uplift of continents from time-varying dynamictopography Geology v 25 no 8 p 735ndash738 doi1011300091-7613(1997)025lt0735CSAUOCgt23CO2

Lithgow-Bertelloni C and P G Silver 1998 Dynamic to-pography plate driving forces and the African super-swell Nature v 395 no 6699 p 269ndash272 doi10103826212

Liu L and M Gurnis 2008 Simultaneous inversion ofmantle properties and initial conditions using an adjointof mantle convection Journal of Geophysical Researchv 113 p B08405 doi1010292008JB005594

Liu L S Spasojeviaelig and M Gurnis 2008 ReconstructingFarallon plate subduction beneath North America backto the Late Cretaceous Science v 322 p 934ndash938doi101126science1162921

Liu LMGurnisM Seton J Saleeby R DMuller and JMJackson 2010 The role of oceanic plateau subduction inthe Laramide orogeny Nature Geoscience v 3 p 353ndash357 doi101038ngeo829

Liu S F and D Nummedal 2004 Late Cretaceous subsi-dence in Wyoming Quantifying the dynamic compo-nent Geology v 32 no 5 p 397ndash400 doi101130G203181

Masters G G Laske H Bolton and A Dziewonski 2000The relative behavior of shear velocity bulk sound speedand compressional velocity in the mantle Implicationsfor chemical and thermal structure in S Karato et aleds Seismology andmineral physics AmericanGeophys-ical Union Geophysical Monograph Series p 63ndash88

Miller K G M A Kominz J V Browning J D WrightG S Mountain M E Katz P J Sugarman B SCramer N Christie-Blick and S F Pekar 2005 ThePhanerozoic record of global sea level change Sciencev 310 no 5752 p 1293ndash1298 doi101126science1116412

Mitrovica J X and G T Jarvis 1985 Surface deflectionsdue to transient subduction in a convecting mantle Tec-tonophysics v 120 no 3ndash4 p 211ndash237 doi1010160040-1951(85)90052-6

Mitrovica J X C Beaumont and G T Jarvis 1989 Tiltingof continental interiors by the dynamical effects of sub-duction Tectonics v 8 no 5 p 1079ndash1094 doi101029TC008i005p01079

Mitrovica J X R N Pysklywec C Beaumont and A Rutty1996 The Devonian to Permian sedimentation of theRussian platform An example of subduction-controlledlong-wavelength tilting of continents Journal of Geody-namics v 22 no 1ndash2 p 79ndash96 doi1010160264-3707(96)00008-7

Morley C K and R Westaway 2006 Subsidence in the

super-deep Pattani and Malay basins of Southeast AsiaA coupled model incorporating lower crustal flow inresponse to postrift sediment loading Basin Researchv 18 no 1 p 51ndash84 doi101111j1365-2117200600285x

Muumlller R D W R Roest J Y Royer L M Gahagan andJ G Sclater 1997 Digital isochrons of the worldrsquos oceanfloor Journal of Geophysical Research v 102 no B2p 3211ndash3214 doi10102996JB01781

Muumlller R D M Sdrolias C Gaina andW R Roest 2008aAge spreading rates and spreading asymmetry of theworldrsquos ocean crust Geochemistry Geophysics Geosys-tems v 9 p Q04006 doi1010292007GC001743

Muumlller R D M Sdrolias C Gaina B Steinberger and CHeine 2008b Long-term sea level fluctuations driven byocean basin dynamics Science v 319 no 5868 p 1357ndash1362 doi101126science1151540

Parsons B and J G Sclater 1977 Analysis of variation ofocean-flow bathymetry and heat-flow with age Journal ofGeophysical Research v 82 no 5 p 803ndash827 doi101029JB082i005p00803

Partridge T C and R R Maud 1987 Geomorphic evolu-tion of southern Africa since the Mesozoic South Afri-can Journal of Geology v 90 p 179ndash208

Richards M A and B H Hager 1984 Geoid anomalies in adynamic Earth Journal of Geophysical Research v 89no B7 p 5987ndash6002 doi101029JB089iB07p05987

Ritsema J H J van Heijst and J H Woodhouse 2004Global transition zone tomography Journal of Geo-physical Research v 109 p B02302 doi1010292003JB002610

Saleeby J 2003 Segmentation of the Laramide slab Evi-dence from the southern Sierra Nevada region Geologi-cal Society of America Bulletin v 115 no 6 p 655ndash668doi1011300016-7606(2003)115lt0655SOTLSFgt20CO2

Sandiford M 2007 The tilting continent A new constrainton the dynamic topographic field from Australia Earthand Planetary Science Letters v 261 no 1ndash2 p 152ndash163 doi101016jepsl200706023

Schellart W P B L N Kennett W Spakman and MAmaru 2009 Plate reconstructions and tomography re-veal a fossil lower mantle slab below the Tasman SeaEarth and Planetary Science Letters v 278 no 3ndash4p 143ndash151 doi101016jepsl200811004

Shepard G E R D Muller L Liu and M Gurnis 2010MioceneAmazonRiver drainage reversal caused by plate-mantle dynamics Nature Geoscience v 3 p 870ndash875doi101038ngeo1017

Simmons N A A M Forte and S P Grand 2009 Jointseismic geodynamic and mineral physical constraintson three-dimensional mantle heterogeneity Implica-tions for the relative importance of thermal versus com-positional heterogeneity Geophysical Journal Interna-tional v 177 no 3 p 1284ndash1304 doi101111j1365-246X200904133x

Sleep N H 1976 Platform subsidence mechanisms and eu-static sea level changes Tectonophysics v 36 no 1ndash3p 45ndash56 doi1010160040-1951(76)90005-6

Sloss L L 1963 Sequences in the cratonic interior of North

America Geological Society of America Bulletin v 74no 2 p 93ndash114 doi1011300016-7606(1963)74[93SITCIO]20CO2

Sloss L L 1988 Tectonic evolution of the craton in Phanero-zoic time in L L Sloss ed Sedimentary cover NorthAmerican craton USA Boulder Colorado GeologicalSociety of America Geology of North America v Dndash2p 22ndash51

Smith A G D G Smith and B M Funnel 1994 Atlas ofMesozoic and Cenozoic coastlines New York Cam-bridge University Press 112 p

Spasojevic S L Liu M Gurnis and R D Muumlller 2008The case for dynamic subsidence of the United States EastCoast since the Eocene Geophysical Research Lettersv 35 p L08305 doi1010292008GL033511

Spasojevic S L Liu and M Gurnis 2009 Adjoint modelsof mantle convection with seismic plate motion and strat-igraphic constraints North America since the Late Creta-ceous Geochemistry Geophysics Geosystems v 10p Q05W02 doi1010292008GC002345

Spasojevic S M Gurnis and R Sutherland 2010a Infer-ring mantle properties with an evolving dynamic modelof theAntarctica NewZealand region from the Late Cre-taceous Journal of Geophysical ResearchmdashSolid Earthv 115 p B05402 doi1010292009JB006612

Spasojevic S M Gurnis and R Sutherland 2010b Mantleupwellings above slab graveyards linked to the globalgeoid lows Nature Geoscience v 3 p 435ndash438 doi101038ngeo855

Spencer A M ed 1974 MesozoicndashCenozoic orogenicbelts Data for orogenic studies Edinburgh GermanyScottish Academic Press 809 p

Stadler G M Gurnis C Burstedde L C Wilcox L Alisicand O Ghattas 2010 The dynamics of plate tectonicsand mantle flow From local to global scales Sciencev 329 no 5995 p 1033ndash1038 doi101126science1191223

Stein C and S Stein 1992 A model for the global varia-tion in oceanic depth and heat flow with lithosphericage Nature v 359 no 6391 p 123ndash129 doi101038359123a0

Sutherland R S Spasojevic and M Gurnis 2009 Mantleupwelling afterGondwana subduction deathmay explainanomalous topography of West Antarctica and subsi-dence history of eastern New Zealand Geology v 38no 2 p 155ndash158 doi101130G306131

Tan E E Choi P Thoutireddy M Gurnis and M Aivazis2006 GeoFramework Coupling multiple models of man-tle convection within a computational framework Geo-chemistry Geophysics Geosystems v 7 p Q06001doi1010292005GC001155

van der Meer D G W Spakman D J J van HinsbergenM L Amaru and T H Torsvik 2010 Towards absoluteplate motions constrained by lower mantle slab remnantsNature Geoscience v 3 no 1 p 36ndash40 doi101038ngeo708

Watts A B and M S Steckler 1979 Subsidence and eus-tasy at the continental margin of eastern North Americain M Talwani W Hay and W B F Ryan eds Deepdrilling results in the Atlantic Ocean Continental margins


and paleoenvironment American Geophysical UnionMaurice Ewing Series v 3 p 218ndash234

Wheeler P and N White 2002 Measuring dynamic topog-raphy An analysis of Southeast Asia Tectonics v 21no 5 p 1040 doi1010292001TC900023

Whittaker JM RDMuumlller andMGurnis 2011Develop-ment of the Australian-Antarctic depth anomaly Geo-chemistry Geophysics Geosystems v 11 p Q11006doi1010292010GC003276

Xu X Q C Lithgow-Bertelloni and C P Conrad 2006Global reconstructions of Cenozoic sea floor ages Impli-cations for bathymetry and sea level Earth and Planetary


Science Letters v 243 no 3ndash4 p 552ndash564 doi101016jepsl200601010

Zhong S J M T Zuber L Moresi and M Gurnis 2000Role of temperature-dependent viscosity and surfaceplates in spherical shellmodels of mantle convection Jour-nal of Geophysical Research v 105 no B5 p 11063ndash11082 doi1010292000JB900003

Zhong S J A McNamara E Tan L Moresi andMGurnis2008 A benchmark study on mantle convection in a 3-Dspherical shell using CitcomS Geochemistry Geophys-ics Geosystems v 9 no 10 p Q10017 doi1010292008GC002048

and Betty Moore Foundation) The originalCitcomS software was obtained from the Com-putational Infrastructure for Geodynamics Thisis contribution number 169 of the Division ofGeological and Planetary Sciences and contri-bution number 10063 of the Tectonics Obser-vatory CaltechThe AAPG Editor thanks the following reviewersfor their work on this paper Steven G Henryand Peter Webb

DATASHARE 45

High-resolution versions of Figures 2 and 4ndash8with captions and appendixes 1ndash3 may beseen on the AAPG Web site (wwwaapgorgdatashare) as Datashare 45


INTRODUCTION

Continental interiors experienced extensive marine inunda-tion and sedimentation during the Cretaceous associated witha global (eustatic) sea level high (Hallam 1992) The marineinundation of North America is a clear example of the ex-tensive flooding where approximately 40 of the continentflooded resulting in more than 1 km (gt06 mi) of sedimentsbeing deposited over a horizontal scale of 1000 km (621 mi)(Bond 1976Liu andNummedal 2004)The eustatic change ontime scales of millions of years may be controlled by changes inocean basin volume caused by variations in oceanic spreadingrates or ridge lengths (Hays and Pitman 1973 Kominz 1984)However North American Cretaceous stratigraphic sequences(Sloss 1963) suggest that observed continental-scale trans-gressional and regressional cycles cannot be explained by eustasyalone (Bond 1976 Sleep 1976) and that large-scale tectonicor dynamic mechanisms are required Namely analysis of theCretaceous North America sediments suggests that the ob-served flooding would require a sea level rise of 310 m (1017 ft)resulting in accumulation of 700 m (2297 ft) of sediments(Bond 1976) whereas the observed Cretaceous isopachs (Cookand Bally 1975) are significantly thicker

Continental flooding can be controlled by dynamic topog-raphy a vertical deflection of the surface resulting frommantleflow (eg Richards and Hager 1984) Although dynamic to-pography results in large-amplitude geoid anomalies becausethese anomalies are small compared to topography (Hager1984) sea level variations should be mainly controlled bychanges in dynamic topography (Gurnis 1990) Early geody-namic studies explored the basic process of dynamic controlsof sea level developing simple two-dimensional (2-D) isovis-cous models of convection that demonstrated that slabs createbroad topographic depressions that can be filled with sediments(Mitrovica and Jarvis 1985) and that continental flooding isinfluenced by movement of continents over large-scale patternsof dynamic topography (Gurnis 1990) Early regional studieslinked mantle dynamics and vertical motions with observationsand models of North America showed that general features ofPhanerozoic strata can be explained by dynamic topographycreated by subduction and supercontinent aggregation and dis-persal (Burgess et al 1997) The extensive Cretaceous flood-ing resulting in the creation of significant hydrocarbon accu-mulations and the subsequent uplift of the Western Interiorseaway (WIS) were interpreted first with simple 2-D dynamicmodels of Farallon slab subduction (Mitrovica et al 1989) andmore recently with three-dimensional (3-D) inverse models





METHOD
































tp eth1THORN










RESULTS







Model Name

M1

M2

M3

Sp

M4

asojevic and G

M5

urnis

M6


S20RTS

S20RTS

S20RTS

SB4L18

SB4L18

SB4L18


120

160

1100

120

160

1100


2045



Figure 4 Continued




Continent

Late Cretaceous


Spasojevic and Gu



ndash

+

+


ndash

ndash

Northern Africa

ndash

ndash

ndash


+

Southeast Asia

+

ndash

ndash


Siberian tilt

Siberian tilt

Alpine Tethys

ndash

ndash

ndash


ndash

+


ndash

ndash

ndash


+

Australia

+()

ndash

ndash

rnis 2047



Figure 5 Continued














Continent

Late Cretaceous


Spasojevic


North America

Dynamic + OFAdagger

OFA + Dynamic



Eurasia

OFA

OFA

OFA

Australia

Dynamic

Dynamic

Dynamic


and Gurnis 2051
































Spas




DISCUSSION















CONCLUSIONS


REFERENCES CITED











































































































METHOD
































tp eth1THORN










RESULTS







Model Name

M1

M2

M3

Sp

M4

asojevic and G

M5

urnis

M6


S20RTS

S20RTS

S20RTS

SB4L18

SB4L18

SB4L18


120

160

1100

120

160

1100


2045



Figure 4 Continued




Continent

Late Cretaceous


Spasojevic and Gu



ndash

+

+


ndash

ndash

Northern Africa

ndash

ndash

ndash


+

Southeast Asia

+

ndash

ndash


Siberian tilt

Siberian tilt

Alpine Tethys

ndash

ndash

ndash


ndash

+


ndash

ndash

ndash


+

Australia

+()

ndash

ndash

rnis 2047



Figure 5 Continued














Continent

Late Cretaceous


Spasojevic


North America

Dynamic + OFAdagger

OFA + Dynamic



Eurasia

OFA

OFA

OFA

Australia

Dynamic

Dynamic

Dynamic


and Gurnis 2051
































Spas




DISCUSSION















CONCLUSIONS


REFERENCES CITED




































































































































tp eth1THORN










RESULTS







Model Name

M1

M2

M3

Sp

M4

asojevic and G

M5

urnis

M6


S20RTS

S20RTS

S20RTS

SB4L18

SB4L18

SB4L18


120

160

1100

120

160

1100


2045



Figure 4 Continued




Continent

Late Cretaceous


Spasojevic and Gu



ndash

+

+


ndash

ndash

Northern Africa

ndash

ndash

ndash


+

Southeast Asia

+

ndash

ndash


Siberian tilt

Siberian tilt

Alpine Tethys

ndash

ndash

ndash


ndash

+


ndash

ndash

ndash


+

Australia

+()

ndash

ndash

rnis 2047



Figure 5 Continued














Continent

Late Cretaceous


Spasojevic


North America

Dynamic + OFAdagger

OFA + Dynamic



Eurasia

OFA

OFA

OFA

Australia

Dynamic

Dynamic

Dynamic


and Gurnis 2051
































Spas




DISCUSSION















CONCLUSIONS


REFERENCES CITED









































































































RESULTS







Model Name

M1

M2

M3

Sp

M4

asojevic and G

M5

urnis

M6


S20RTS

S20RTS

S20RTS

SB4L18

SB4L18

SB4L18


120

160

1100

120

160

1100


2045



Figure 4 Continued




Continent

Late Cretaceous


Spasojevic and Gu



ndash

+

+


ndash

ndash

Northern Africa

ndash

ndash

ndash


+

Southeast Asia

+

ndash

ndash


Siberian tilt

Siberian tilt

Alpine Tethys

ndash

ndash

ndash


ndash

+


ndash

ndash

ndash


+

Australia

+()

ndash

ndash

rnis 2047



Figure 5 Continued














Continent

Late Cretaceous


Spasojevic


North America

Dynamic + OFAdagger

OFA + Dynamic



Eurasia

OFA

OFA

OFA

Australia

Dynamic

Dynamic

Dynamic


and Gurnis 2051
































Spas




DISCUSSION















CONCLUSIONS


REFERENCES CITED









































































































Figure 4 Continued




Continent

Late Cretaceous


Spasojevic and Gu



ndash

+

+


ndash

ndash

Northern Africa

ndash

ndash

ndash


+

Southeast Asia

+

ndash

ndash


Siberian tilt

Siberian tilt

Alpine Tethys

ndash

ndash

ndash


ndash

+


ndash

ndash

ndash


+

Australia

+()

ndash

ndash

rnis 2047



Figure 5 Continued














Continent

Late Cretaceous


Spasojevic


North America

Dynamic + OFAdagger

OFA + Dynamic



Eurasia

OFA

OFA

OFA

Australia

Dynamic

Dynamic

Dynamic


and Gurnis 2051
































Spas




DISCUSSION















CONCLUSIONS


REFERENCES CITED









































































































Figure 5 Continued














Continent

Late Cretaceous


Spasojevic


North America

Dynamic + OFAdagger

OFA + Dynamic



Eurasia

OFA

OFA

OFA

Australia

Dynamic

Dynamic

Dynamic


and Gurnis 2051
































Spas




DISCUSSION















CONCLUSIONS


REFERENCES CITED














































































































Continent

Late Cretaceous


Spasojevic


North America

Dynamic + OFAdagger

OFA + Dynamic



Eurasia

OFA

OFA

OFA

Australia

Dynamic

Dynamic

Dynamic


and Gurnis 2051
































Spas




DISCUSSION















CONCLUSIONS


REFERENCES CITED






































































































































Spas




DISCUSSION















CONCLUSIONS


REFERENCES CITED









































































































DISCUSSION















CONCLUSIONS


REFERENCES CITED

















































































































CONCLUSIONS


REFERENCES CITED







































































































sea level and vertical motion authors of continents from dynamic earth models … › 35517 › 1...

Documents