Crustal Collapse, Mantle Upwelling, and Cenozoic Extension in the North American Cordillera

TECTONICS, VOL. 17, NO. 2, PAGES 311-321, APRIL 1998

Crustal collapse, mantle upwelling, and Cenozoic extension in the North American Cordillera

Mian Liu and Yunqing Shen Departmentof GeologicalSciences, University of Missouri,Columbia

Abstract. Gravitational collapse has been suggestedas the still convergingand the regionalstress field is predominantly major cause of Cenozoic extension in the North American compressional.This is illustrated by active extension in the Cordillera and many other orogenic belts. Although both centralAndes [Dalmayer and Molnar, 1981;Dewey, 1988] and crustal thickening and mantle upwelling may have contributed the Tibetan plateau [England and Houseman, 1989; Molnar to the Cordilleran extension,previous models of gravitational and Chen, 1983]. collapse have focused on the former; the cause of mantle The popularexplanation is the hypothesisof gravitational upwelling and its relationship to crustal collapse remain collapse[Dewey, 1988; Molnar and Lyon-Caen,1988]. Most obscure. Here we attempt to addressthe questionof whether orogenicbelts in the world are supportedby an Airy-type gravitationalcollapse of an overthickenedcrust could induce crustal root [Airy, 1855]. The isostatic balance of vertical major mantle upwelling and whole-lithosphere extension. forces,however, does not meanmechanical equilibrium in the Thermal-rheologicalcalculations indicate that crustalcollapse lateral directions. A thickened crust, with a higher may decouple from the mantle lithosphere, because the gravitational potential energy than the adjacent lowlands, extensional forces arising from an overthickened crust are tendsto spreador collapseunder its own weight [Artyushkov, limited to the crust, while the rheology of continental 1973; McKenzie, 1972]. Gravitational collapse is now lithosphereis intrinsically stratified. Even when the mantle believedto have playeda majorrole in both synorogenicand lithosphere is mechanically coupled to the crust, postorogenicextension in many orogenic belts [Dewey, thermomechanicalmodeling indicates that strain is localized 1988], including the Cenozoic extension in the North in the weak lower crust during crustal collapse, and no American Cordillera (Figure 1) [Coney, 1987; Harry et al., significant (<10 km) thinning of the mantle lithospheremay 1993; Livaccari, 1991; Sonder et al., 1987; Wernicke et al., be induced at the absenceof extensional forces from plate 1987]. boundaries. Crustal collapse of the Sevier-Laramide orogen Although gravitationalcollapse of an overthickenedcrust seems adequate to account for much of the mid-Tertiary is conceptually simple, its geodynamics are not well extensionin the Cordillera, including formation of many core understood. One major problem is the role of mantle complexes,but it is unlikely to have been the major causeof upwelling,which is often involvedin orogeniccollapse and the more recent basin-and-rangeextension. We suggestthat a contributesto the driving forces [Dewey, 1988; England and strong pulse of mantle upwelling in the mid-Tertiary, as Houseman, 1989]. However, becausemantle processesare indicated by the "ignimbrite flare-up," may have triggered difficult to constrain, most studies have avoided them or basin-and-rangeextension by weakening the lithosphereand included them in calculations without offering much providingexcess gravitational potential energy. The causeof discussionof their causeand relationshipto crustalcollapse mantle upwelling remains uncertain, but the continued [Sonderet al., 1987]. Various assumptionshave been made, extension and volcanism since mid-Miocene in the northern leading to divergentconclusions. Some workers, assuming Basin and Range province favor an active mantle upwelling mechanical decoupling of the crustal processesfrom the with internal convective heating. mantle lithosphere because of the weak lower crust, have predictedquick flattening of crustal welts into "pancakes" with little effect on the mantle lithosphere[Bird, 1991]; 1. Introduction others, assuminga full crust-mantlecoupling, have shown that extensionalcollapse of an overthickenedcrust can lead to Orogenic belts, the products of intensive compressional major extensionof the whole lithosphere[Govers and Wortel, tectonics, are often the sites of continental extension and 1993; Harry et al., 1993]. rifting, as first observedby Wilson [1966] in his study of the In this study we attempt to constrainthe role of crustal openingof the Atlantic Ocean. The theory of plate tectonics collapse and the relationshipbetween crustal collapse and offers no ready explanation for such extension, because mantle upwelling. The well-established history of the orogenicbelts usually form near convergentplate boundaries, Cenozoic extension in the North American Cordillera and extensionof orogenic belts happenseven when plates are providesa goodexample for this study.The questionswe hope to address include the following: (1) What is the role of gravitationalcollapse of the Sevier-Laramideorogen in the Cordilleran extension?(2) Could crustal collapse of the Copyright1998 by the AmericanGeophysical Union. Sevier-Laramide orogen have led to basin-and-range Papernumber 98TC00313. extension?(3) What is the causeof mantleupwelling under the 0278-7407/98/98TC-00313512.00 Basin and Range province?

311 312 LIU AND SHEN: CRUSTAL COLLAPSE AND CORDILLERAN EXTENSION

50 ø

40 ø

30 ø

125 ø 115 ø 105 ø Figure 1. Shadedrelief map of the North AmericanCordillera. The toothcurve shows the easternboundary of the Sevier-Laramidefold-and-thrust belts. The bold solid line indicatesthe boundaryof the Basin and Range province.The stippledareas are the locationsof major metamorphiccore complexes,and the thin solid line shows the main volcanic field of the ignimbrite flare-up during mid-Tertiary. Abbreviationsare as follows: NBR, northern Basin and Range, also called the Great Basin (GB); SRB, southernBasin and Range; SRP, Snake River Plain; and CP, Colorado Plateau.

2. Cenozoic Extension in the North American Although localized extensionmay have startedas early as Cordillera the Mesozoic [Hodgesand Walker, 1992], major extensionin the Cordillera is postorogenic,occurring after the Laramide There are numerous excellent review articles on extensional orogeny [Zobacket al., 1981]. In many placesthe inception tectonics in the North American Cordillera [Burchfiel et al., of extensionis markedby developmentof metamorphiccore 1992; Burchfiel and Davis, 1975; Coney, 1987' Eaton, 1982; complexes [Coney, 1987]. The occurrenceof many core Hamilton and Myers, 1966; Stewart, 1978' Thompson and complexesalong the core zone of the Sevier-Laramideorogen Burke, 1974; Wernicke, 1992; Zoback et al., 1981]. A where crust was significantly thickenedstrongly indicatesa summary here highlights some of the problems pertinent to causeof crustalcollapse [Coney and Harms, 1984]; however, our discussion. not all core complexes are formed in the overthickenedcore During the Mesozoic and early Cenozoic (-165-55 Ma), zone. In the southernBasin and Range province, most core western No_rthAmerica experienced a protracted phase of complexesdeveloped in the midst of a deep-seatedthrust belt crustal compressionas the oceanic Farallon plate subducted [Coney, 1980]; detachmentfaults and core complexesare also underneath North America. The crustal contraction involved a found in the Mojave desertwhere there is no clear evidencefor complex history of subduction-related deformation and a significantly thickened crust before extension [Dokka and massive plutonism along the coastal margin [Burchfiel and Ross, 1995; Glazner and Bartley, 1984]. Furthermore, the Davis, 1975]. Crustal shortening in the inland Cordillera inception of major extension in the Cordillera was occurredin a zone stretchingfrom Canadato northernMexico; diachronous,although the duration and the amount of crustal it telescoped more than 200 km of crust and caused contraction were remarkably uniform from southeastern progressivedevelopment of the fold-and-thrustbelt [Elison, British Columbia to Nevada and Utah [Elison, 1991]. In the 1991]. In the hinterland of the orogenic belt the crustal southern Canadian Cordillera, northern Washington, Idaho, thicknesswas nearly doubled to more than 50 km [Coney and and Montana, extensionbegan in early Eocene. Farther south, Harms, 1984; Parrish et al., 1988]. the inception time was largely Oligocene in the Great Basin LIU AND SHEN:CRUSTAL COLLAPSE AND CORDILLERANEXTENSION 313

(the northern Basin and Range province), and was slightly a b later in the Mojave-Sonora desert region. Extension in areas near the latitude of Las Vegas did not occuruntil mid-Miocene B A P- pgz [Wernicke et al., 1987]. Wernicke et al. [1987] have pointed out that the timing of inception of extensionin the Cordillera is apparently correlated with the abundanceof associated plutonism, and numerousstudies have stressedthe importance of magmatism in core-complex formation [Armstrong and Ward, 1991;Axen et al., 1993]. Much of the magmatism may have resulted from postorogenic thermal relaxation and radioactive heating [Glazner and Bartley, 1985]; however, at least in the Canadian Cordillera where extension occurred Prn :' ' within a few million years after the orogeny, some thermal Figure 2. Conceptual model of crustal collapse. (a) Sketch perturbations from the mantle seem necessary [Liu and of a thickened crust with an Airy-type crustal root. (b) Furlong, 1993]. Lithostatic pressures along vertical profiles across the Calc-alkaline volcanism was widespreadthrough much of lowland crust (line A) and the mountain range (line B). The the early to middle Tertiary and culminated with eruption of differential pressureAp tends to causecrustal collapse;the voluminous(>35,000 km3) silicictuff in the centralGreat lateral pressure gradient in the transition zone between the Basin between 34 and 17 Ma [Best and Christiansen, 1991]. mountain range and the lowland tends to drive lateral Although eruption of mafic magma is rare, significant mantle extrusion of the ductile lower crust. Notice that Ap vanishes below the crustal welt. upwelling seems necessaryto provide heat for the extensive crustal anatexis [Hildreth, 1981] and source materials for some of the silicic tuff [Grunder, 1995; Johnson, 1991]. The The approachwe take here is to first constrainthe effectsof relationship between extensional tectonics and volcanism is controversial. Some workers find that extension was mainly crustal collapse of the Sevier-Laramide orogen, which synvolcanical [Gans et al., 1989], others [Axen et at., 1993; involves fewer uncertainties than either plate interactions or Best and Christiansen,1991; Taylor and Bartley, 1992] argue mantle processes.By isolating the effects of crustal collapse, we may reach a better understandingof the role of mantle for a poor spatial-temporalcorrelation between volcanismand extension on a provincial scale. Liu and Furlong [1994] upwelling and plate interactions. suggested that the apparently poor correlation between extension and volcanism may be partially attributed to the 3. Crustal Collapse competing effects of thermal weakening and rheological 3.1. Driving Forces and Thermal-Rheological hardening associated with intrusion and underplating of Control mantle-derived magmas. Since mid-Miocene (-17 Ma) another major phase of The dynamic instability of an overthickened crust at extension, with characteristic deep-penetrating (10-15 km) isostaticequilibrium is illustratedin Figure 2. The crustal welt block faulting and association with bimodal (basaltic- is unstablebecause at any depth above the compensationlevel rhyolitic) volcanism, has led to formation of the Basin and (taken to be at the Moho of the mountain range) the Range province, where the total extension is estimated to be lithostatic pressure under the mountain range is greater than between 50% and 300% [Hamilton and Myers, 1966; that under the surroundinglowland. This differential pressure Wernicke, 1992]. It is commonly recognized that this tends to drive gravitational collapse of the mountain range. younger phase of basin-and-rangeextension [Zoback et al., The vertical integral of the differential pressurerepresents the 1981] is fundamentally different from the earlier low-angle total extensionalforce [Lynch and Morgan, 1987]: detachment faults [Burchfiel et al., 1992; Coney, 1987], although the change between these two types of extension L was gradualin many places[Zoback et al., 1981]. l[Pt(Z)-Pr(Z)],:lZ The cause of extension in the Cordillera is a subject of -h (]) intensivestudy and debate.One view attributesthe Cordilleran tectonics to plate interactions along the western margin of where Pt (z) and Pr (z) are the lithostaticpressure at depthz North America [Atwater, 1970; Severinghaus and Atwater, under the mountain range and the lowland, respectively;h is 1990], where subduction of the Farallon plate under North the elevation of the mountain range above the reference America has been replaced by the evolving San Andreas lowland, and L is the compensationdepth. The extensional transform fault since 25-30 Ma. Sonder et al. [1987] and force F actson the sidesperpendicular to the planeof Figure2 Wernicke et al. [1987], noting that much of the Cordilleran andtherefore has dimensions of force per unit length (N m-•) extension happened when western North America was under a [Turcotteand Schubert,1982]. The valueof F, representedby compressive and transpressive tectonic regime, have the shadedarea in Figure 2b, is numerically equivalent to the emphasizedthe role of gravitationalcollapse. Because of the excess gravitational potential energy stored in a column of involvement of major mantle upwelling in the basin-and- the mountainrange relative to that in the lowland [Molnar and range extension, there are also numerous suggestionsfor a Lyon-Caen, 1988]. It is clear from Figure 2 that F is causativerole of mantle thermal perturbations[Parsons et at., dependent on the density of the crust and mantle and the 1994; Saltus and Thompson,1995; Suppeet at., 1975]. elevation of the mountainrange above the referencelowland, 314 LIU AND SHEN: CRUSTAL COLLAPSE AND CORDILLERAN EXTENSION

which may be related to crustal thicknessif a uniform crustal collapse, because the continental lithosphere has an• density and the Airy isostasycan be assumed: intrinsically stratified rheology (Figure 3). The strength envelopesof the lithosphereare defined by h= (Pm-Pc )(Ht- Hr ) (2) Pm O'yield(Z)=min(%, o a) (4) where H t and H r are the thicknessesof the thickenedand referencecrust, respectively;Pm and Pc are the mantle and where o't, and o'a are the stressdifference ((o' 1 - o3) / 2 ) crustal density, respectively.The extensionalforce defined in needed for brittle and ductile extension at depth z, (1) can be easily calculatedfrom the elevation and the density respectively. Laboratory experiments [Brace and Kohlstedt, contrastbetween the crust and the mantle [Molnar and Lyon- 1980] suggestthat ductile deformationdepends on lithology, Caen, 1988]: strain rate, and, especially,temperature:

F = pcgh H r +• (3) h+AH2 /

where g is gravitationalacceleration and Mar is the thickness where k is the strain rate; the typical value of • for of the Airy-type crustalroot: Mar= H t -H r -h. For Airy-type continentaldeformation is in the rangebetween 10 -14 s -1 and isostasy, Mar= hPc/(Pm-Pc)' The crustin the hinterlandof 10-16S-1. The parameterA is a constant,H is the activation the Cordilleran fold-and-thrust belts was thickened to -60 km enthalpy, R is the gas constant, and T is the absolute near the end of the Sevier-Laramide orogeny [Coney and temperature.For a given range of stressesthe parametern is a Harms, 1984; Parrish et al., 1988]. If we take the reference constant associated with deformation mechanism (n --3 for crust (the crust adjacentto the hinterland) to be 40 km thick, dislocation creep in the lithosphere). All results discussed thecrustal and mantle density to be 2800kg m-3 and3300 kg belowassume a graniticcrust (A=10 -8'8 MPa -n s-1, H=123 KJ m-3, respectively,we obtainan elevationof 3 km from(2); tool-1, and n=3) andan olivine-dominatedmantle lithosphere the correspondingextensional force is about4.2x1012 N m-1 (A=103'28MPa -n s-1, H=420 KJ mo1-1,and n=3) [Kirby and from (3), comparablewith typical tectonic forces associated Kronenberg, 1987; Korato et al., 1986]. Using other with ridge pushand slab pull [Bott, 1993; Forsythand Uyeda, publishedrheological parametersfor the lithospherewill not 1975]. affect the generalconclusions drawn here. Notice that the gravitationaldriving forces arising from an Brittle deformationof rocks is characterizedby slidingon overthickened crust alone are limited to the crust (Figure 2). fracturesand faults and is generallyindependent of strainrate, This simple fact has important ramifications for crustal temperature, and lithology [Byerlee, 1978]:

Yield Strength(MPa) zSP(MPa) t•t, =ktt• n (6) 0 100 200 300 400 0 100 0 wherekt is the frictionalcoefficient (taken to be 0.85) and t•n is normal stresson the fault plane. Assuming fracturesoccur in all orientations[Brace, 1972], crn can be replacedby the effective lithostatic stress(the lithostatic pressureminus the pore fluid pressure). Recent studies indicate that the brittle • 20ß [ Ductile strengthbecomes less sensitiveto pressureas depthincreases • 40 [Shimada, 1993]; the complexitiesof t•/• at higherpressures are not critical for our discussion here and therefore are not considered.The yield strengthof the lithosphereis usually 60 defined as the vertical integral of the yield stressacross the whole lithosphere [Ranalii, 1995]:

80 0 S= I O'yield (z)dz (7) Figure 3. (left) Strength envelopesof a model continental l lithosphere.The dashedprofile is for the referencelithosphere characterizedby a 30 km crust and an equilibrium geotherm where l is the baseof the referencelithosphere. We alsodefine witha surfaceheat flux of 60 mWm -2. The shaded profiles are the crustal strength as the yield stressintegrated acrossthe for 20 and 50 m.y. after an instant crustal thickening that crust. increased the crustal thickness to 50 km; reduction of the lithospheric strength is due to postkinematic thermal The "jelly sandwich"rheological structure shown in Figure relaxationand radioactiveheating and the mantlelithosphere 3 is intrinsic to the continental lithospherebecause of the being pushedto a greater depth. Notice the thick channelof temperature dependence of rheology and different ductile lower crust under orogens. (right) The differential compositions of the crust and mantle. It is also clear from pressure(Ap) in Figure 2 is replotted. Figure 3 that the ductile lower crust is particularlyweak and LIU AND SHEN: CRUSTAL COLLAPSE AND CORDILI.ERAN EXTENSION 315

thick under orogenic belts owing to heating associatedwith 1993]; this is a condition representativefor core-complex crustal thickening (thermal relaxation, radioactive heating, formation in the Cordillera [Parrish et al., 1988; Wernicke et shearheating, etc. [see Liu and Furlong, 1993]) and the fact al., 1987]. The decreaseof extensionalforce and lithospheric that the mantle material is pushed down to a hotter regime strengthwith time in Figure 4 is due to thinningof the crustal [Glazner and Bartley, 1985]. Such rheologicalstructures raise welt by crustal collapse,which in the model occursin the form the important questionof whether crustal collapse,driven by of lateralductile extrusion within the lower crust,driven by an extensionalforces arising within the crust, can lead to whole- initial lateral topographic gradient of 0.01. The inset in lithosphere extension. Figure 4 showsthe typical velocity profile. For simplicity, brittle deformation is not included; this is not critical here 3.2 Crust-Mantle Decoupling and Evolution of becausethe extensionalforces and lithosphericstrength are Crustal Collapse mainly determined by the crustal thickness and thermal The effects of crustal collapseon mantle lithospheredepend structures and are not sensitive to the details of extensional on mechanical coupling between the crust and mantle, which, processes.Since crustalcollapse is decoupledfrom the mantle in turn, is influenced by thermal structuresof the lithosphere. lithosphere, no asthenospheric upwelling is induced to Predicting thermal structures of orogens inevitably involves compensatefor the lost gravitationalpotential energy, so the poorly constrainedfactors such as shear heating, erosion, and total extensional force decreasesrapidly; within 10 million heat input from the mantle [Liu and Furlong, 1993]. To derive yearsor so it becomesless than the yield strengthof the crust, some general constraints,we considerhere two end-member and crustal collapseis expectedto stop. We found that, within cases.The first case is for a relatively cold lithospherewhere reasonable ranges of crustal thickness (40- 60 km) and the strengthof the uppermostmantle preventsit from flowing thermalstructures, the predictedlifespan of crustalcollapse is together with the lower crust. This is the assumptionin most about 5-15 million years. Such a relatively short lifespanof models of ductile flow within the lower crust [Bird, 1991]. For crustal collapse is comparable with that of core-complex crustal collapse to induce extensionin the mantle lithosphere, formation in the Cordillera [Parrish et al., 1988]. sufficient shear stressesneed to be transmitted through the The other end-member case is for a relatively hot ductile lower crust.The channelPoiseuille flow may be usedto lithospherewhere the uppermostmantle is sufficientlyweak approximateductile flows within the lower crustdriven by the to flow togetherwith the lower crust duringcrustal collapse; lateral pressuregradient induced by topographicchanges; the in this sensethe crust and mantle are fully coupled.We have shear stress exerted on the top of the mantle lithosphere is modeledductile flow within the crust and mantle lithosphere [Bird, 1991]: induced by crustal thickening at orogenic belts. Figure 5a shows the model geometry and boundary conditions. The ductile flows are driven by the lateral pressure gradient z=bPcgC•x x (8)

4.0 where g is the gravitationalacceleration, c (equal to cos 0 (Moho slope)) is a small geometriccorrection factor, and dh/dx is the topographicgradient. The parameterb is roughly half the thickness of the ductile channel for Newtonian fluids 3.0 but is only aboutone fourth for power law fluids [Bird, 1991]. 55 For an upper bound stress estimation, take b = 10 km, a 0.0 topographicgradient of 0.01, and a flat Moho, we find •=2.8 Velocity(mm/yr) 0.1 MPa. Similar results can be obtained from a Couette flow 2.0 approximation,assuming the lower crust flows by the shearof the sliding upper crust [Hopper and Buck, 1996]. Given the typical lithosphericstrength (>1012 N m-1) [Lynchand 1.0 S1 Morgan, 1987], such shear stressesare unlikely to cause significant deformation in the mantle lithosphere. In this Crustal Collapse ---• case, crustal collapsemay be mechanicallydecoupled from the mantle lithosphere, as assumedin previous models [Bird, 0.0 1991; Block and Royden, 1990]. 0 5 10 15 Crustal collapsethat decouplesfrom the mantle lithosphere is expected to be short-lived. This is illustrated in Figure 4, Time (m.y.) which compares the total extensional force with the yield strength of a model lithosphere determined in a one- Figure 4. Evolution of the total extensionalforce (F) and dimensional thermomechanical model, assuming complete the yield strengthof the lithosphere(Si) andCrust (Sc) during crustal collapse that is mechanically decoupled from the mechanical decoupling between the crust and mantle mantle lithosphere.The inset shows one typical velocity lithosphere.In the model the crust with an initial thicknessof profile of ductile extrusion in the crust (brittle extension is 35 km is thickened instantly to 55 km by overthrustingof a not included in the model). Time is after the initiation of 20-km crustal sheet. Crustal collapseis allowed to occur when crustal collapse.Crustal collapseis expectedto stop within temperatureat the Moho reaches650øC by postkinematic --10 million years when the extensionalforce is insufficient thermal relaxation and radioactive heating [Liu and Furlong, to overcomethe crustal yield strength. 316 LIU AND SHEN: CRUSTAL COLLAPSE AND CORDILLERAN EXTENSION

crustalflows, we chosea rigid upperboundary and a free lower Surface T = 0øC •x. boundary so that the lateral extrusionof crustal material is all balancedby upwelling of the mantle material under the crustal BrittleCrustu-v=0 welt. As shown in Figure 5b, the lateral flow is channelizedin the lower crust and occurs mainly in the transition zone ThickenedCrust Reference Crust between the thickened crust and the adjacent lowland, where the lateral pressuregradient is the highest.The vertical flow is Mantle induced by the lateral flow, as required by volume • Lithosphere • conservation.Figure 6 shows the maximum flow rate and the accumulative thinning of the mantle lithosphere under the crustal welt for one experiment that started with a thermal structurecharacterized by a surfaceheat flux of 80 mW m-2. Baseof Lithosphere u = ohv / c)y = 0 T = 1300øC The sharp drop of flow rate with time in Figure 6 is due to Distance decrease of the lateral pressuregradient as the ductile flow reducesthe elevation gradient; the flow essentially stops after b 10-20 million years. The total amount of thinning of the mantle lithosphere is about 4 km. Within the reasonable range of model parameters,including the amount of crustal thickening (20-30 km) and the initial crustal temperature 25- (characterizedby an equilibriumsurface heat flux between60 and90 mW m-2),the totalamount of thinningof the mantle lithosphereis less than 10 km. In other words, even when the mantle is fully coupledto the crust,crustal collapse alone does not induce significant thinning of the mantle lithosphereand upwelling of the asthenosphere.The short lifespan of crustal 75- collapseis consistentwith formationof somecore complexes in the Cordillera [Parrish et al., 1988]; however, crustal collapse of the Sevier-Laramide orogen seems unlikely to have been the major causeof basin-and-rangeextension. 0.0 0.5 1.0 1.5 Distance (100 km) 4. Mantle Upwelling and Basin-and-Range Extension Figure 5. (a) Model geometry and boundary conditionsfor ductile deformation driven by the lateral pressuregradient There is little doubt that basin-and-range extension has between the crustal welt and the lowland. Only the right half involved major mantle upwelling [Stewart, 1978]. Seismic of the model is solved becauseof the symmetry of the model. studies indicate that the lithospherein the Great Basin is as The topographicgradient, not shownin Figure 5, is calculated thin as -65 km [Benz et al., 1990; Smith et al., 1989]. An assuming local isostasy. The numerical mesh used for the abnormallyhot mantle under the Basin and Range provinceis calculationsis 51x51. (b) Snapshotof the velocity field. The also indicatedby the gravity [Eaton et al., 1978], high surface thick shaded line indicates the initial Moho; the thin solid feat flux (-90 mW m-2) [Lachenbruchand Sass,1978], and line is the Moho at 12 m.y. The mantle lithosphereis fully coupledto the crust,as shownby the continuousvelocity field across the crust and mantle. 10.0

between the thickened crust and the lowland. Assuming .=• constant crustal and mantle density and local isostasy, this 7.5 7.5 lateral pressure gradient is related to the change of crustal thickness: 5.0 5.0 Lithosphericthinnin cgP Pc) dH (9) 2.5 2.5• '•xx=P cg (1- p m dx .,• Flow rate where P is the pressure,x is the horizontaldistance, and H is 0.0 o.o the crustal thickness.To isolate the effects of crustal collapse, 0 lO 20 30 40 50 we impose no lateral displacement (i.e., no shortening or Time (m.y.) stretching) at the right-side boundary, which is also taken to be a far-distanceboundary (calculated to a distanceof 300 km) Figure 6. The maximum flow rate and the accumulative so its influence on the velocity field near the crustal welt thinning of the mantle lithosphere during a model crustal (Figure 5b) is proven negligible. To maximize the amountof collapse that is mechanically coupled to the mantle thinning of the mantle lithosphere that may be induced by lithosphere.Time is after the initiation of crustal collapse. LIU AND SHEN:CRUSTAL COLLAPSE AND CORDILLERANEXTENSION 317 seismic attenuationin the upper mantle under the Basin and a Range [Rornanowicz, 1979; Smith et al., 1989]. The high 16 elevation of the Basin and Range province (~1.5 km) is largely compensatedby thermally induced mass deficiency in the mantle, although some of the mass deficiency may be 12 related to compositional heterogeneities[Humphreys and Dueker, 1994]. Although projecting the present lithospheric structureback into the geologicalpast is difficult, significant mantle upwelling since mid-Miocene is clearly indicated by the widespreadbimodal volcanism associatedwith basin-and- range extension [Lipman, 1980]. There was probably strong mantle upwelling just before basin-and-rangeextension, as indicatedby the voluminousmid-Tertiary (34-17 Ma) volcanic eruptionin the Great Basin. This so called ignimbriteflare-up, recordedby morethan 35,000 km 3 of silicicvolcanic ash flow depositedover a region of >71,000 km2 [Best and Christiansen, 1991], requires significant upwelling of the 40 80 120 asthenosphereto supply both the parental magmasfor some of the volcanic tuff [Grunder, 1995; Johnson, 1991] and heat Depth to the Upwelled Mantle (km) for the extensive crustal anatexis [Hildreth, 1981; Liu, 1996]. Such a mantle upwelling could have played a major role in driving basin-and-range extension. The mechanics of continental rifting induced by mantle upwelling have been b intensively studied [Crough, 1978; Sengor and Burke, 1978] and will not be discussed. Here we examine the general gravitational instability of the lithosphere over an upwelled Yield Strength(MPa) AP(MPa) asthenosphere.Figure 7 shows that an elevated lithosphere 0 50 100 0 50 isostatically supported by a buoyancy asthenosphere is dynamically unstable and tends to collapse. The situation is similar to that of an overthickenedcrust (see Figure 2) with two major differences:(1) With asthenosphericupwelling the gravitational extensional forces are distributed across the • 50 whole lithosphere, and (2) heat advected by the upwelling

a• 100. a b

B A P=pgz 150,

I Figure 8. (a) Total gravitational driving force (F) and 1 lithospheric strength (S) as functions of mantle upwelling, oho_ calculatedassuming 15 m.y. after an instant mantle upwelling I to various depths, with temperature of the upwelled mantle kept constant at 1300øC. The initial thickness of the A B lithosphere is 150 km. (b) The strength envelope (left) and the differential pressure(right) for mantle upwelling to 60 km Upwelleld• I depth. The numbersshow the values of the total driving force andyield strength ( in 1012N m-l). Asthenosphere•I asthenospheremay significantly weaken the lithosphere.The • Z total extensional force and the lithospheric strength depend mainly on the amountof asthenosphericupwelling (Figure 8). Figure 7. Conceptualmodel of gravitationalcollapse of the The results in Figure 8 are derived with a model lithosphere lithospheredue to mantleupwelling. (a) Structureof the model that is initially 150 km thick; the driving force and the lithosphere. The uplifted topography is isostatically supported by thermal buoyancy forces in the upwelled lithosphericstrength are calculatedfor 15 million years after asthenosphere.The depth of isostaticcompensation is at the an instant asthenosphericupwelling, comparableto the time base of the reference lithosphere.(b) Pressureprofiles across interval between the peak mid-Tertiary volcanism and basin- the reference lithosphere(line A) and the thinned lithosphere and-rangeextension [Best and Christiansen,1991]. Figure8a (line B). indicates that asthenosphericupwelling to depths shallower 318 LIU AND SHEN: CRUSTAL COLLAPSE AND CORDII.LERAN EXTENSION

than 70 km depth would provide sufficient extensionalforces dimensionaladvection model; the changeof elevationin the to cause whole-lithospheric extension. Figure 8b shows the extensionalregion was calculated assuminglocal isostasy. vertical distribution of the extensionalforces (shown by the The extensionalforce and the lithosphericstrength were then differential pressure between the thinned and the reference calculatedby integratingthe differentialpressure and the yield lithosphere) and lithospheric strength for asthenospheric strength across the lithosphere according to (1) and (7). upwelling to 60 km depth. Although uncertaintiesof mantle Figure 9 suggeststhat lithosphericextension may not last for compositions under the Basin and Range province and the more than 10 million years for case B, because thermal exact amount of basaltic magma involved in the ignimbrite buoyancy forces in the upwelled mantle quickly diminish flare-up make it difficult to place tight bounds on the through conductive cooling during extension. The total magnitudeof mantle upwelling during the peak mid-Tertiary amount of extension(vertically averagedhorizontal strain) is volcanism, asthenosphericupwelling to around 60 km depth less than 30% (equal to 10-•5 s-• x 10 m.y.) in caseB. is not unreasonable. Decompressional partial melting of Conversely,more than 20 million years of extensioncan be typical mantle materials would require mantle upwelling to expectedfor caseA, with -80% extension.These values vary less than 50 km depth [Liu and Furlong, 1992; McKenzie and mainly with the amount of mantle upwelling and the initial Bickle, 1988], and a significant amount of basaltic magmas thicknessof the crustin the extensionalzone and the adjacent was likely involved in the eruption of voluminous mid- reference lithosphere. Initial thermal structures of the Tertiary volcanic rocks in the Great Basin [Feeley and lithosphereare not critical here, becausethey are quickly Grunder, 1991; Grunder, 1995; Johnson, 1991]. Such a mantle overprintedby heat advectedby the upwellingasthenosphere. upwelling would be sufficient to trigger basin-and-range The initial extensionalforce would be higher for a greater extension. contrast of crustal thickness between the extensional zone and The causeof mantle upwelling under the Basin and Range the adjacentlowland. In any case the predictedextension for remainsuncertain. Our resultsargue against crustal collapse of active and passivemantle upwelling is significantlydifferent. the Sevier-Laramideorogen being the major cause.Harry et al. The continued extension and volcanism in the Basin and [1993] have shown that gravitational collapseof a thickened Range province since mid-Miocene and the large total strain crust could lead to significant mantle upwelling and (50% to 300%) across the Cordillera are more consistentwith lithospheric extension, but that may be attributed to the an active mantle upwelling. constant lateral stretching imposed in their model. Other causesinclude delaminationof the mantle lithosphere[Bird, 5. Discussion 1979], convective thinning [Houseman et al., 1981], mantle upwelling in a slablesswindow [Dickinsonand Snyder, 1979] Gravitational collapse of an overthickened crust is or slab gap [Severinghausand Atwater, 1990] associatedwith conceptuallysimple and geologically observable,a major the migration of the Mendocino triple junction, subduction- reason for gravitational collapse to have gained much induced mantle upwelling in a back arc setting [Stewart, 1978], and a mantle plume [Parsons et al., 1994; Saltus and Thompson, 1995; Suppe et al., 1975]. Unfortunately, many of these processesare difficult to test. We may group the 3.0 proposed mantle processes into two major categories according to thermal evolution within the upwelling asthenosphere:(1) passive mantle upwelling, where mantle upwells adiabatically and then cools by conduction;and (2) 2.t) active mantle upwelling, where mantle upwelling is adiabatic or superadiabatic, such as in a mantle plume, and the temperaturein the upwelled mantle is maintained by some kind of convective flows. The thermal histories between these two types of mantle upwelling are significantly different so that some general constraintsmay be derived by comparing the predicted stability of the lithosphere with the history of basin-and-rangeextension in the Cordillera. Figure 9 shows the predicted total extensional force and 0o0 yield strengthof a model lithospherewith an active (case A) 0 10 20 and a passive(case B) mantle upwelling. Both casesstarted with an instantmantle upwelling to 60 km depth.The crustis Time (m.y.) initially 40 km thick in the extensional zone and is 35 km thick in the referencelowland. In case A, temperaturewas Figure 9. Evolution of the total extensionalforce (F) and lithosphericstrength (S) during lithosphericextension with maintainedto be 1300øC at 60 km depthwith an adiabatic an active(case A) anda passive(case B) mantleupwelling. In geothermwithin the upwelled asthenosphere,while in case B both cases, extension starts 15 m.y. after an instantaneous conductivecooling of the upwelled mantle was allowed after upwelling of the asthenosphereto 60 km depth. During its initial ascension.Assuming pure shear extensionof the lithosphericextension the upwelledasthenosphere is kept at lithospherewith a uniform strain rate of 10-15s-l, the 1300øC at 60 km depthin caseA but coolsconductively in transient thermal structure was calculated using a one- case B. LIU AND SHEN: CRUSTAL COLLAPSE AND CORDILLERAN EXTENSION 319

popularity in recent years. Although mantle upwelling is Range province brings us a step closer to the causeof basin- often involved [Dewey, 1988; England and Houseman,1989], and-range extension. Localized mantle upwelling under the its relationship with crustal collapse is ambiguous. It has Cordillera may have startedin the Eoceneor even earlier. Liu been shown in some models [Govers and Wortel, 1993; Harry and Furlong [ 1993] find that an increasedmantle heat flux was et al., 1993] and implied in other studies [Livaccari, 1991; needed to account for the high crustal temperature and Molnar and Chen, 1983] that extensional collapse of the plutonism associated with the Eocene crustal extension and overthickenedcrust of the Sevier-Laramide orogen could have core-complex formation in the southwestern Canadian led to basin-and-rangeextension. Our results suggestthat the Cordillera. The mantle upwelling could have been causedby effects of crustal collapse are more limited than previously delamination of the mantle lithosphere [Bird, 1979] or thought. Crustal collapseof the Sevier-Laramideorogen could convective thinning [Houseman et al., 1981]; in either case, have accounted for much of the localized, short-lived mid- basin-and-rangeextension can be regardedas having a similar Tertiary (>mid-Miocene) extension in the Cordillera, origin to the mid-Tertiary extension, both resulting from the including formation of many metamorphiccore complexes; dynamic instability of a thickened lithosphere[Sonder et al., however, it is unlikely to have been the major causeof basin- 1987]. However, the models of delamination or convective and-range extension. A different cause for basin-and-range thinning are difficult to test; conversely, the southward extension is consistent with its fundamental differences with migration of Tertiary volcanism in the Great Basin may be the detachment faults associated with core-complex more easily explained as a result of the retreating Farallon formations [Coney, 1987; Zoback et al., 1981]. plate [Best and Christiansen, 1991; Lipman, 1980]. In any The role of crustal collapse is limited because the case, the intensive mid-Tertiary ignimbrite flare-up indicatesa extensional forces arising from an overthickened crust are strong pulse of mantle upwelling, which may have triggered limited to the crust, while the rheology of continental basin-and-range extension by providing the excess lithosphere is intrinsically stratified. Although any gravitational potential energy and by thermally weakening calculations of the lithospheric rheology inevitably involve the lithosphere [Liu, 1996]. uncertainties with extrapolating the laboratory-determined While our discussionhas been focused on crustal collapse rheologic parameters, there are abundant geological and and mantle upwelling, there is no doubt that the evolving geophysical observations indicating the existence of a tectonic setting in western North America has played an locally weak, ductile crustand mechanicaldecoupling between important role in the Cordilleran extension. The crustal the upper crust and the mantle lithosphere(see Kirby and collapse of the Sevier-Laramide orogen was probably Kronenberg [1987] for a review). Crust-mantledecoupling is triggered by the drop of compressionalstresses at the end of thought to be important in thin-skinned thrusting [Ranalli the Laramide orogeny as a result of the reducedconvergent rate and Murphy, 1987] and crustal deformation at convergent between the Farallon and the North American plates [Coney, orogens[Royden, 1996]. The casefor crust-mantledecoupling 1987]; the basin-and-rangeextension is closely related to the is especially strong for crustal collapse of orogenic belts, change of plate geometry at the western margin of North where the ductile lower crust is particularly thick and weak. America, where convergencebetween the Farallon and North Mechanical decoupling may also occur between finer-scale American plates has been gradually replaced by the San rheological layers resulting from magmatism and Andreastransform fault [Dickinson and Snyder, 1979; Zoback compositional heterogeneities in the crust [Lister and et al., 1981]. This change of tectonic setting may have Baldwin, 1993], allowing crustal collapse to occur at much facilitated basin-and-range extension by reducing the lower deviatoric stressesthan those required to overcome the compressional stresses; however, it generates no major yield strength of the whole lithosphere.This may help to extensional forces [Sonder et al., 1986]. If the major forces explain the diffusive crustalextension in the Great Basin that driving basin-and-rangeextension are the thermal buoyancy spansa greater spaceand time than the few well-developed forces in the upwelling mantle, our results indicate that some metamorphiccore complexes[Axen et al., 1993]. kinds of convective flows within the upwelling mantle were Crustal collapse that decouples mechanically from the necessaryto sustain the gravitational potential energy. mantle lithospherewould causethrusting near the marginsof orogenic belts, as required by the volume conservation. 6. Conclusions Thrusting concurrentwith extensionis observedin the Andes 1. Gravitationalcollapse of an overthickenedcrust may be and the Tibet Plateau [Burchfiel et al., 1992; Molnar and largely decoupled from the mantle lithosphere; crustal Lyon-Caen, 1988] and in some Neogene extensionalbasins in collapse alone cannot induce major mantle upwelling and the Mediterranean [Platt and Vissers, 1989]. Evidence for whole-lithosphereextension. Most crustal collapse is short- thrusting at the margins of the Cordilleran hinterland coeval lived (10-15 million years) and may have concurrentthrusting with core-complex extension is scarce; this may be partly at the margins of orogens. Crustal collapse of the Sevier- attributed to the difficulties in interpreting faults from Laramide orogen is consistentwith the mid-Tertiary extension complicatedtectonic overprinting in the Cordillera [Coney, and formation of metamorphic core complexes in the 1980]. Concurrent stretching of the mantle lithosphere by Cordillera; however, it is unlikely to have been the major processesrelated to plate interactions [Dokka and Ross, causeof basin-and-rangeextension. 1995] would also mitigate the volume problem created by 2. The strong pulse of mantle upwelling indicatedby the crustal collapse. ignimbrite flare-up may have triggered basin-and-range Eliminating crustal collapse of the Sevier-Laramideorogen extension by weakening the lithosphere and providing the as the direct cause of mantle upwelling under the Basin and excessgravitational potential energy. The causeof the mantle 320 L1U AND SHEN: CRUSTAL COLLAPSEAND CORDILLERAN EXTENSION

upwelling is uncertain, but the continuous volcanism and Acknowledgments.This work was supportedby the NSF grant extension since mid-Miocene in the Basin and Range EAR-9506460and the ACS/PRFgrant 27925-G2 administrated by the AmericanSociety of Chemistry.We thankClem Chasefor helpful province favor an active mantle upwelling with internal discussionand P. Coney, L. Royden, J. Platt, D. Scholl, and an convective heating. anonymousreviewer for their helpfulreview.

References

Airy, G.B., On the computationof the effectof the Coney, P.J., and T.A. Harms, Cordilleran Harry, D.L., D.S. Sawyer, and W.P. Leeman,The attractionof mountainmasses, as disturbingthe metamorphic core complexes: Cenozoic mechanics of continental extension in western apparent astronomical latitude of stations in extensional relics of Mesozoic compression, North America:Implications for the magmatic geodetic surveys, Philos. Trans. R. Soc. Geology, 12, 550-554, 1984. and structural evolution of the Great Basin, London, Ser. B, 145, 101-104, 1855. Crough,S.T., Thermalorigin of mid-platehotspot Earth Planet. Sci. Lett., 117, 59-71, 1993. Armstrong, R.L., and P. Ward, Evolving swells, Geophys.J. R. Astron. Soc., 55, 451- Hildreth,W., Gradientsin silicicmagma chambers: geographicpatterns of Cenozoicmagmatism in 469, 1978. Implicationsfor lithosphericmagmatism, J. the North AmericanCordillera: The temporaland Dalmayer, B., and P. Molnar, Parallel thrust and Geophys.Res., 86, 10,153-10,192, 1981. spatial association of magmatism and normal faulting in Peru and constrainson the Hodges, K.V., and J.D. Walker, Extension in the metamorphiccore complexes,J. Geophys.Res., state of stress, Earth Planet. Sci. Lett., 55, 473- Cretaceous Sevier orogen, North American 96, 13,201-13,224, 1991. 481, 1981. Cordillera, Geol. Soc.Am. Bull., 104, 560-590, Artyushkov, E.V., Stressesin the lithosphere Dewey, J.F., Extensional collapse of orogens, 1992. causedby crustalthickness inhomogeneities, J. Tectonics, 7, 1123-1139, 1988. Hopper, J.R., and W.R. Buck, The effects of lower Geophys.Res., 78, 7675-7690, 1973. Dickinson,W.R., and W.S. Snyder,Geometry of crustflow on continentalextension and passive Atwater, T., Implicationsof plate tectonicsfor the subducted slabs related to the San Andreas margin formation, J. Geophys. Res., 101, Cenozoic tectonic evolution of western North transform, J. Geol., 87, 609-627, 1979. 20,175-20,194, 1996. America, Geol. Soc. Am. Bull., 81, 3513-3536, Dokka, R.K., and T.M. Ross, Collapse of Houseman,G.A., D.P. McKenzie, and P. Molnar, 1970. southwestern North America and the evolution Convectiveinstability of a thickenedboundary Axen, G.J., W. Taylor, and J.M. Bartley, Space- of earlyMiocene detachment faults, metamorphic layer and its relevancefor the thermal evolution time patternsand tectonic controlsof Tertiary core complexes,the Sierra Nevada,orocline, and of continentalconvergent belts, J. Geophys. extensionand magmatismin the Great Basin of the San Andreasfault system,Geology, 23, Res., 86, 6115-6132, 1981. the western United States, Geol. Soc. Am. Bull., 1075-1078, 1995. Humphreys,E.D., and K.G. Dueker,Physical state 105, 56-72, 1993. Eaton,G.P., The Basin and RangeProvince: Origin of the westernU.S. uppermantle, J. Geophys. Benz, H.M., R.B. Smith, and W.D. Mooney, and tectonic significance, Annu. Rev. Earth Res., 99, 9635-9650, 1994. Crustal structure of the northwestern Basin and Planet. Sci., 10, 409-440, 1982. Johnson,C.M., Large-scale crustal formation and Range province from the 1986 PASSCAL Eaton,G.P., R.R. Wahl, H.J. Prostka,D.R. Mahey, lithospheremodification beneath Middle to Late seismic experiment, J. Geophys. Res., 95, and M.D. Kleinkopf, Regional gravity and Cenozoic calderasand volcanic fields, western 21,823-21,842, 1990. tectonicpatterns: Their relationto late Cenozoic North America, J. Geophys.Res., 96, 13,485- Best, M.G., and E.H. Christiansen, Limited epeirogenyand lateral spreadingin the western 13,508, 1991. extensionduring peak Tertiary volcanism,Great Cordillera,in CenozoicTectonics and Regional Kirby, S.H., and A.K. Kronenberg,Rheology of Basin of Nevada and Utah, J. Geophys.Res., Geophysicsof the Western Cordillera, edited the lithosphere:Selected topics, Rev. Geophys., 96, 13,509-13,528, 1991. by R.B. Smith and G.P. Eaton, Mem. Geol. Soc. 25, 1219-1244, 1987. Bird, P., Continental delamination and the Am., 152, 51-91, 1978. Korato, S.I., M.S. Paterson,and J.D. Fitzgerald, ColoradoPlateau, J. Geophys.Res., 84, 7561- Elison, M.W., Intracontinental contraction in Rheology of synthetic olivine aggregates: 7571, 1979. western North America: Continuity and Influenceof grain size and water,J. Geophys. Bird, P., Lateral extension of lower crust from under episodicity, Geol. Soc. Am. Bull., 103, 1226- Res., 91, 8151-8176, 1986. high topography in the isostatic limit, J. 1238, 1991. Lachenbruch, A.H., and J.H. Sass, Models of an Geophys.Res., 96, 10,275-10,286, 1991. England, P.C., and G.A. Houseman, Extension extending lithosphere and heat flow in the Block, L., and L.H. Royden, Core complex during continental convergencewith special Basin and Range Province, in Cenozoic geometriesand regional scale flow in the lower referenceto the Tibetan plateau,J. Geophys. Tectonicsand Regional Geophysicsof the crust, Tectonics, 9, 557-567, 1990. Res., 94, 17,561-17,597, 1989. Western Cordillera, edited by R.B. Smith and Bott, M.H.P., Modeling the plate-driven Feeley, T.C., and A.L. Grunder, Mantle G.P. Eaton, Mem. Geol. Soc. Am., 152, 209- mechanism, J. Geol. Soc. London, 150, 941- contributionto the evolutionof Middle Tertiary 250, 1978. 951, 1993. silicic magmatism during early stages of Lipman, P.W., Cenozoic volcanism in the western Brace, W.F., Laboratorystudies of stick-slip, and extension:The Egan Range volcanic complex, United States: Implication for continental their application to earthquakes, east-central Nevada, Contrib. Mineral. Petrol., tectonics, in Studies in Geophysics: Tectonophysics,14, 189-200, 1972. 106, 154-169, 1991. ContinentalTectonics, pp. 161-174,Nat. Acad. Brace, W.F., and D.L. Kohlstedt, Limits on Forsyth, D., and S. Uyeda, On the relative Sci., Washington,D. C., 1980. lithospheric stress imposed by laboratory importance of the driving forces of plate Lister, G.S., and S.L. Baldwin, Plutonism and the experiments,J. Geophys.Res., 85, 6248-6252, motion, Geophys.J. R. Astron. Soc., 43, 163- originof metamorphiccore complexes, Geology, 1980. 200, 1975. 21, 607-614, 1993. Burchfiel, B.C., and G.A. Davis, Nature and Gans, P.B., G.A. Mahood, and E. Schermer, Liu, M., Dynamic interactions between crustal controls of cordilleran orogenesis, western Synextensionalmagmatism in the Basin and shortening,extension, and magmatismin the United States: Extensions of an earlier Range Province:A case studyfrom the eastern North American Cordillera, Pure Appl. synthesis,Am. J. Sci., 275, 363-396, 1975. Great Basin, Spec. Pap. Geol. Soc. Am., 233, Geophys., 146, 447-467, 1996. Burchfiel, B.C., D.S. Cowan, and G.A. Davis, 58 pp., 1989. Liu, M., and K.P. Furlong,Cenozoic volcanism in Tectonicoverview of the Cordilleranorogen in Glazner, A.F., and J.M. Bartley, Timing and the California Coast Ranges: Numerical westernUnited States,in The Geologyof North tectonic setting of Tertiary low-angle normal solutions, J. Geophys. Res., 97, 4941-4957, America, vol. G-3, The Cordilleran Orogen: faulting and associated magmatism in the 1992. Conterminous U.S., edited by B.C. Burchfiel, southwestern United States, Tectonics, 3, 385- Liu, M., and K.P. Furlong,Crustal shortening and M.L. Zoback, and P.W. Lipman, pp. 407-480, 396, 1984. Eocene extension in the southeastern Canadian Geol. Soc. of Am., Boulder, Colo., 1992. Glazner, A.F., and J.M. Bartley, Evolution of Cordillera: Some thermal and mechanical Byerlee, J.D., Friction of rocks, Pure Appl. lithosphericstrength after thrusting,Geology, considerations, Tectonics, 12, 776-786, 1993. Geophys., 116, 615-626, 1978. 13, 42-45, 1985. Liu, M., and K.P. Furlong, Intrusion and Coney, P.J., Cordilleran metamorphic core Govers, R., and M.J.R. Wortel, Initiation of underplating of mafic magmas: thermal- complexes: An overview, Mere. Geol. Soc. Am., asymmetricextension in continentallithosphere, rheologicaleffects and implications for Tertiary 153, 7-31, 1980. Tectonophysics,223, 75-96, 1993. tectonomagmatism in the North American Coney, P.J., The regional tectonic setting and Grunder, A.L., Material and thermal roles of basalt Cordillera, Tectonophysics, 273, 175-187, possible causes of Cenozoic extension in the in crustalmagmatism: Case studyfrom eastern 1994. North American Cordillera, in Continental Nevada, Geology, 23, 952-956, 1995. Livaccari, R.F., Role of crustalthickening and ExtensionalTectonics, edited by M.P. Coward, Hamilton,W., and W.B. Myers, Cenozoictectonics extensionalcollapse in the tectonicevolution of J.F. Dewey, and P.L. Hancock,Geol. Soc.Spec. of the westernUnited States,Rev. Geophys.,4., the Sevier-Laramideorogeny, western United Publ., 28, 177-186, 1987. 509-549, 1966. States,Geology, 19, 1104-1107, 1991. LIU AND SHEN:CRUSTAL COLLAPSE AND CORDILLERANEXTENSION 321

Lynch, H.D., and P. Morgan, The tensile strength Saltus, R.W., and G.A. Thompson, Why is it Taylor, W.J., and J.M. Bartley, Prevolcanic of the lithosphere and the localization of downhill from Tonopah to Las Vegas?:A case extensional Seaman breakaway fault and its extension, in Continental Extensional for mantle plume supportof the high northern geological implications for eastern Nevada and Tectonics, edited by M.P. Coward, J.F. Dewey Basin and Range, Tectonics, 14, 1235-1244, western Utah, Geol. Soc. Am. Bull., 104, 255- and P.L. Hancock, Geol. Soc. Spec. Publ., 28, 1995. 266, 1992. 53-66, 1987. Sengor,A.M.C., and K. Burke, Relative timing of Thompson, G.A., and D.B. Burke, Regional McKenzie, D.P., Active tectonics of the rifting and volcanism on earth and its tectonic Geophysicsof the Basin and Range province, Mediterranean region, Geophys. J. R. Astron. implications,Geophys. Res. Lett., 5, 419-421, Annu. Rev. Earth Planet. Sci., 2, 213-238, Soc., 30, 109-185, 1972. 1978. 1974. McKenzie, D., and M.J. Bickle, The volume and Severinghaus, J., and T. Atwater, Cenozoic Turcotte, D.L., and G. Schubert, Geodynamics: compositionof melt generatedby extensionof geometryand thermal state of the subducting Applications of Continuum Physics to the lithosphere,J. Petrol., 29, 625-679, 1988. slabs beneath western North America, in Basin Geological Problems, 450 pp., John Wiley, Molnar, P., and W.-P. Chen, Focal depthsand fault and Range Extensional Tectonics Near the New York, 1982. plane solutions of earthquakes under the Latitude of Las Vegas, Nevada, edited by B.P. Wernicke, B.P., Cenozoic extensional tectonics of Tibetan plateau, J. Geophys. Res., 88, 1180- Wernicke, Mere. Geol. Soc. Am., 176, 1-22, the U.S. Cordillera, in The Geology of North 1196, 1983. 1990. America, vol. G-3, The Cordilleran Orogen: Molnar, P., and H. Lyon-Caen, Some simple Shimada, M., Lithosphere strength inferred from Conterminous United States, edited by B.C. physical aspects of the support, structure,and fracture strength of rocks at high confining Burchfiel,M.L. Zoback,and P.W. Lipman, pp. evolution of mountain belts, Spec. Pap. Geol. pressures and temperatures, Tectonophysics, 553-581, Geol. Soc. of Am., Boulder. Colo., Soc. Am., 218, 179-207, 1988. 217, 55-64, 1993. 1992. Parrish, R.R., S.D. Cart, and D.L. Parkinson, Smith, R.B., W.C. Nagy, K.A. Julander, J.J. Wernicke, B., R.L. Christiansen, P.C. England, Eocene extensional tectonics and Viveiro, C.A. Baker, and D.G. Gants, and L.J. Sonder,Tectonomagmatic evolution of geochronologyof the southern Omineca belt, Geophysical and tectonic framework of the Cenozoic extension in the North American British Columbia and Washington, Tectonics, eastern Basin and Range-Colorado Plateau- Cordillera, in Continental Extensional 7, 181-212, 1988. Rocky Mountain transition, in Geophysical Tectonics,edited by M.P. Coward,J.F. Dewey, Parsons,T., G.A. Thompson,and N. Sleep, Mantle Framework of the Continental United States, and P.L. Hancock, Geol. Soc. Spec. Publ., 28, plume influence on the Neogene uplift and edited by L.C. Pakiser and W.D. Mooney, 203-221, 1987. extension of the U.S. western Cordillera?, Mere. Geol. Soc. Am., 172, 205-233, 1989. Wilson, J.T., Did the Atlantic close and then re- Geology, 22, 83-86, 1994. Sonder, L.J., P.C. England, and G.A. Houseman, open?, Nature, 211, 676-681, 1966. Platt, J.P., and R.L.M. Vissers, Extensional Continuum calculations of continental Zoback, M.D., R.E. Anderson, and G.A. collapseof thickenedcontinental lithosphere: A deformation in transcurrent environments, J. Thompson, Cenozoic evolution of the state of working hypothesis for the Alboran Sea and Geophys. Res., 91, 4797-4818, 1986. stressand style of tectonismof the Basin and Gibraltar arc, Geology, 17, 540-543, 1989. Sonder, L.J., P.C. England, B. Wernicke, and R.L. Ranges Province of the western United States, Ranalli, G., Rheology of the Earth, 413 pp., Christiansen, A physical model for Cenozoic Philos. Trans. R. Soc. London, Ser. A, 300, Chapmanand Hall, New York, 1995. extension of western North America, in 407-434, 1981. Ranalii, G., and D.C. Murphy, Rheological Continental Extensional Tectonics, edited by stratificationof the lithosphere,Tectonophysics, M.P. Coward, J.F. Dewey, and P.L. Hancock, 132, 281-295, 1987. Geol. Soc. Spec. Publ., 28, 187-201, 1987. M. Liu and Y. Shen, Dept. of Geological Romanowicz, B.A., Seismic structureof the upper Stewart,J.H., Basin and Range structurein western Sciences, University of Missouri, 101 Geology mantle beneath the United States by three- North America: a review, in Cenozoic Tectonics Building, Columbia, MO 65211. (e-mail: dimensional inversion of body wave arrival and Regional Geophysics of the Western mian@ geosc.missouri.edu, times, Geophys. J. Royal Astron. Soc., 57, Cordillera, edited by R.B. Smith and G.L. [email protected]) 479-506, 1979. Eaton, Mere. Geol. Soc. Am., 152, 1-31,1978. Royden, L., Coupling and decouplingof crust and Suppe, J., C. Powell, and R. Berry, Regional mantle in convergentorogens: Implications for topography, seismicity, quaternary volcanism, (received October 16, 1997; strainpartitioning in the crust,J. Geophys.Res., and the present-day tectonics of the western revised January 6, 1998; I01, 17,679-17,705, 1996. United States, Am. J. Sci., 275, 397-436, 1975. acceptedJanuary 28, 1998.)