etters 249 (2006) 6273www.elsevier.com/locate/epslEarth and Planetary Science LSupport of high elevation in the southern Basin and Range based onthe composition and architecture of the crust in the

Basin and Range and Colorado Plateau

Andy Frassetto a,, Hersh Gilbert a, George Zandt a, Susan Beck a, Matthew J. Fouch b

a Department of Geosciences, University of Arizona, Gould-Simpson Building #77, 1040 E 4th St., Tucson, AZ 85721, USAb Department of Geological Sciences, Arizona State University, Box 871404, Temple, AZ 85287-1404, USA

Received 17 February 2006; received in revised form 26 June 2006; accepted 27 June 2006Editor: R. D. van der HilstAbstract

To explore the nature of how the structure and physical properties of the crust vary from extended to relatively unextended domainswe present teleseismic receiver functions which measure crustal thickness, shear wavespeed structure and the Vp/Vs ratio at 12seismic stations in eastern Arizona. The crustal thickness is28 km, increases slightly eastward, and remains nearly uniform beneaththe varying elevations in the southern Basin and Range. The observed Vp/Vs ratio in the Basin and Range (1.78) exceeds the globalaverage. The southern Colorado Plateau exhibits thicker crust (40 km) and a slightly greater observed Vp/Vs (1.81). A discreteregion in the Colorado Plateau generates an unusually high Vp/Vs ratio (1.90) and contains low wavespeed zones which serve asevidence of partial melt related to Quaternary volcanism. The metamorphic core complexes in the southern Basin and Range likewiseexhibit anomalously high Vp/Vs values (1.791.87) and lack locally compensating crustal roots. Density models show that 85 kg/m3 lighter crust or 35 kg/m3 lighter mantle than that of the surrounding Basin and Range helps these metamorphic core complexesmaintain their high elevation. Compositional modeling of intrusive bodies exposed throughout the CatalinaRincon metamorphiccore complex indicates that the observed high Vp/Vs ratio and modeled low density could result from substantial amounts of aplagioclase-rich, quartz-poor rock. These Vp/Vs data are evidence that significant compositional heterogeneity of the crust can occurover a short distance and provide a clue as to how these areas that underwent significant Tertiary extension may have beenpreconditioned for orogenic collapse. 2006 Elsevier B.V. All rights reserved.Keywords: Vp/Vs ratio; receiver function; metamorphic core complex; isostasy; Basin and Range; Colorado Plateau1. Tectonic setting and broadband deployment

Southwestern North America has been tectonicallymodified to an extreme degree since the Mesozoic; fromthe late Cretaceous to early Tertiary (8040 Ma) theshallowing dip of the subducting Farallon slab shifted Corresponding author. Tel.: +1 520 730 6731; fax: +1 520 621 2672E-mail address: andyf@geo.arizona.edu (A. Frassetto).

0012-821X/$ - see front matter 2006 Elsevier B.V. All rights redoi:10.1016/j.epsl.2006.06.040.

served.magmatism inboard of the arc and promoted Laramidecrustal thickening throughout the Cordillera (see Dick-inson [1] for a summary). The subsequent orogeniccollapse in the middle Tertiary extended and exhumedmid-crustal metamorphic and igneous rocks from withinthe overthickened crust forming metamorphic corecomplexes (MCCs) in the Basin and Range [2].Continued extension and magmatism persisted duringthe transition to a transform boundary as the Farallon

mailto:andyf@geo.arizona.eduhttp://dx.doi.org/10.1016/j.epsl.2006.06.040

Fig. 1. The COARSE deployment (2003-Present) across easternArizona. All sites (excluding KATH) are near bedrock and producehigh quality broadband earthquake recordings. For additionalinformation visit http://asuarray.asu.edu/coarse/index.php.

63A. Frassetto et al. / Earth and Planetary Science Letters 249 (2006) 6273plate was subducted (see Sonder and Jones [3] for areview). Meanwhile unresolved factors allowed uplift ofthe Colorado Plateau to its present elevation withoutimparting significant internal deformation. The Ceno-zoic collapse of the Cordillera and its linkage to thecurrent structure, composition, and form of support forthe high elevations of both the Colorado Plateau and theBasin and Range remains a controversial issue that isfundamental to understanding the tectonic developmentof western North America.

Here we present a teleseismic receiver functionanalysis using data from a recent regional broadbandseismic deployment. These observations effectivelyconstrain the Poisson's ratio (or Vp/Vs) of the crust,allow us to identify layers of differing shear wavespeedand measure crustal thickness, and provide a vitalframework to interpret the evolution of the Cenozoicextensional domain in western North America. Inparticular we hope to further constrain the forces whichdrive the profound changes in elevation and topographythat are observed from the Basin and Range to theColorado Plateau (e.g., [4]). To this end we use thereceiver function method to characterize the crust andupper mantle in the region, produce constraints neces-sary to broadly explore the composition of the crust, andmodel the density distribution in the upper lithosphere toexamine the sources that compensate high elevation inthe southern Basin and Range.

The Consortium for an Arizona ReconnaissanceSeismic Experiment (COARSE) deployment of ninebroadband seismometers spans the three main tectonicprovinces of Arizona in an approximate SWNE swath(Fig. 1). Arizona State University's ASU seismograph,National Seismic Network station WUAZ, and GlobalSeismic Network station TUC are included to improvearray coverage.

2. Teleseismic receiver function analyses

To analyze the structure of the crust and uppermostmantle beneath COARSE array stations we computehigh quality teleseismic receiver functions with theiterative pulse stripping time domain deconvolutiontechnique (Fig. 2) [5]. During this process the vertical Pwave energy is deconvolved from the horizontal (radialand tangential) components, removing source effects andpreserving only near receiver PSVand PSH convertedphases which are generated by structure in thelithosphere beneath each seismic station [6,7]. Thisstudy focuses on the PSV (hereafter called Ps)converted phases of the radial component which aremost sensitive to large scale variability in the shearwavespeed structure. Frequency content of our receiverfunctions from 0.05 Hz to 2.5 Hz allows for resolution ofPs phases from layer thicknesses of greater than0.5 km. The amplitude of a Ps phase depends on thewavespeed contrast at a layer boundary. Negative Pspolarities characterize higher to lower wavespeeds withincreasing depth, while positive polarities characterizelower to higher wavespeeds (Fig. 2). We refer to theamplitude of a Ps as a percentage of its amplitude to thatof normalized initial P pulse. To analyze the highestquality results we accept receiver functions only if theysatisfy a variance reduction cutoff above 80% and retaintheir highest amplitude, the direct P, within one second ofthe zero time on the deconvolved trace. Direct P sourcesare selected from epicentral distances of 30 to 90 andyield a dataset where most events are clustered innorthern and southwestern Pacific subduction zones, aswell as in Central and South America.

Analyses of these data in the depth domain allow forthe most accurate tectonic interpretation. To accomplishthis and to produce a robust seismic characterization weeliminate one major unknown by assuming a standardVp for the entire crust. We use a wavespeed of 6.2 km/sdeveloped by Christensen and Mooney [8] in theirreview of crustal seismic observations from extensional

http:////asuarray.asu.edu/coarse/index.php

Fig. 2. Stacked Basin and Range receiver functions, grouped by Moho amplitude and Vp/Vs. Commonly observed interfaces have been identifiedwith labels. M represents the estimated Moho.

64 A. Frassetto et al. / Earth and Planetary Science Letters 249 (2006) 6273domains. Numerous refraction experiments throughoutthe Basin and RangeColorado Plateau transition zonealso measure a bulk crustal Vp of 6.20.1 km/s [912].This value also matches observations for the southernColorado Plateau [13]. A range of wavespeeds between6.16.3 km/s is probable based on the variety of geologicenvironments sampled by COARSE and the uncertaintyin the refraction results. Tests on COARSE data showthat variations in Vp have negligible effects on theobserved Vp/Vs ratio and alter crustal thicknessFig. 3. Vp/Vs-thickness relation from phase stacks of: (i) receiver function az(30090) sampling beneath the CatalinaRincon MCC.estimates by only 0.5 km for Vp changes of 0.1 km/s. Thus minor deviations from our constant Vp model donot significantly affect the overall findings.

We evaluate the primary Ps phase and its multiples byusing the thicknessVp/Vs domain phase stackingalgorithm (Fig. 3) [14]. This stacking technique allowsus to map the convergence of the direct Ps and its PpPsand PpSs+PsPs multiples for the Moho interface inorder to estimate the crustal Vp/Vs and thickness beneathour stations.With ideal convergence a point of maximumimuths (90300) sampling beneath the Tucson Basin and (ii) azimuths

Table 1Elevation, crustal thickness and Vp/Vs (with uncertainty range at95% of maximum stacked amplitude) for Basin and Range stations

Station Elevation(m)

Thickness(km)

Vp/Vs Events

ABBY 794 31 (3133) (1.741.68) 1.72 64ASU 364 24.5 (2425.5) (1.801.74) 1.77 34DSRTa, b 802 25.5 (2527.5) (1.951.90) 1.95 28KITT 1467 27.5 (25.528.5) (1.861.76) 1.79 49LEMN 2489 29 (28.530) (1.891.84) 1.86 21SQRL 3085 29 (2829.5) (1.871.84) 1.84 36TUCc 874 30 (2932) (1.751.67) 1.72 270TUCd 874 28.5 (2829) (1.891.82) 1.87 117

Tradeoff between crustal thickness and Vp/Vs ratio exhibits that thelowest allowable thickness corresponds to the greatest Vp/Vs withinthe maximum amplitude bounds. Uncertainty range does not accountfor variations in Vp.a Azimuthal heterogeneity and/or significant noise present.b 90% maximum amplitude contour used.c Azimuths (90300) sampling beneath the Tucson Basin.d Azimuths (30090) sampling beneath the CatalinaRincon MCC.

Table 2Elevation, crustal thickness and Vp/Vs (with uncertainty range at 95%of maximum stacked amplitude) for Colorado Plateau stations

Station Elevation(m)

Thickness(km)

Vp/Vs Events

KNTHb 2286 41.5 (4142) (1.781.73) 1.76 71PFNP 1759 39.5 (38.541) (1.791.73) 1.76 35RENO 1909 36.0 (34.536.5) (1.861.81) 1.82 36WUAZ 1592 43.5 (4344.5) (1.921.88) 1.90 165ZIZZ a, b 2011 43 (4243) (1.711.67) 1.68 44

Tradeoff between crustal thickness and Vp/Vs ratio exhibits that thelowest allowable thickness corresponds to the greatest Vp/Vs withinthe maximum amplitude bounds. Uncertainty range does not accountfor variations in Vp.a Azimuthal heterogeneity and/or significant noise present.b 90% maximum amplitude contour used.

65A. Frassetto et al. / Earth and Planetary Science Letters 249 (2006) 6273amplitude or bull's-eye signifies the most likelysolution. The inherent weakness in this technique, thetradeoff between crustal thickness and Vp/Vs whichleads to a diagonal smearing of the bull's-eye, producessome uncertainty in both the crustal thickness and Vp/Vssolutions. Additionally the peak amplitude of the stackedphases is often asymmetrically located within a largerzone of high amplitude. In order to constrain theconfidence in our picks we determine the width of themaximum amplitude range and accept any solutionwithin 95% (or 90% at stations DSRT, KNTH, ZIZZ) ofthe peak amplitude. This produces a range of estimatesencompassing the maximum convergence that representthe uncertainty in our results (Tables 1 and 2).

To verify these measurements we plot the moveout ofphase arrivals as a function of ray parameter [15].Moveout accounts for the differential arrival time ofdirect or converted waves plotted as a function of rayparameter. Direct Ps arrivals exhibit positive moveoutwith respect to the direct P arrival with increasing rayparameter while the ray geometry of PpPs and PpSs+PsPs multiples produces a negative moveout. Thus wecan predict and compare expected and observed move-out on receiver functions in order to verify that the arrivalof Ps phases and their later multiple reverberations areproduced at particular depths. The identification of laterphases is critical prior to interpretation of seismicstructure because it allows us to discriminate betweendirect arrivals and multiples that may be spuriouslyinterpreted. Receiver functions from each station arecorrected for moveout, stacked to improve the observedsignal, and migrated to depth using the measured Vp/Vsto determine the depth of the Moho.

3. Receiver function observations andinterpretations

Although the event count at some stations is small, atleast 21 receiver functions pierce the crust and uppermantle within several kilometers of each location.Azimuthal distribution of seismicity reveals a consider-able gap in sampling towards the N and E where fewsuitable teleseismic earthquakes are generated. Despitethese hindrances we record pronounced and systematictrends in the receiver functions sampling within eachtectonic region. Within solely the Basin and Range weview interesting variations between stations located onMCC summits and those located in less extended regions.

3.1. Basin and range

Receiver functions for the Basin and Range (Fig. 2)contain similar features for most stations. At 620 kmdepth we find a broad negative Ps (labeled A in figure) of510% that is indicative of the top of slow wavespeedmaterial, perhaps due to the presence of high tempera-tures, partial melt or a shear zone. The negative arrival ismost pronounced near the MCCs and is entirely absent atABBY, the only station to sample a relatively undeformedregion, suggesting that it may result from extension. Awell-defined Moho Ps (M) arrives from depths of 24.531 km. With the exception of the especially thin crust atASU crustal thickness remains similar throughout theregion and is comparable to thicknesses estimated fromprevious refraction profiling [10,11].

Moho amplitude varies significantly. Stations locatedat low elevation (ABBY, DSRT, TUC) record a uniformly

66 A. Frassetto et al. / Earth and Planetary Science Letters 249 (2006) 6273weak Moho amplitude (10%). This small amount ofconverted energy may result from a reduced shearwavespeed and/or density contrast across the crustmantle interface. The Moho Ps beneath or near the SouthMountain, Baboquivari, CatalinaRincon and PinaleoMCCs (stations ASU, KITT, LEMN, SQRL respectively)is considerably greater (1530%) suggesting a more rapidtransition from crust to mantle or a larger impedancecontrast at the base of the crust; either slow or low densitycrust atop normal mantle or alternatively normal lowercrust atop fast or high density mantle.

The receiver functions show few converted phasesbelow the Moho Ps. There is evidence of a diffuse 5%positive Ps (B) at 4055 km beneath all stations exceptDSRT (Fig. 2). Noise may obscure this interface at DSRTand the layer is probably characteristic of the region andwould represent an abrupt increase in shear wavespeed at1520 km below the Moho. This may signify anincrease with depth from the low Pnwavespeeds of 7.677.85 km/s observed previously [1012]. We also notice a57% negative Ps (C) that appears at most stations at 6075 km and corresponds to a decrease in shear wavespeedin the upper mantle.Moveout analysis verifies that neitherthis nor the earlier arrival result from the reverberations ofshallower layers. Therefore these observed Ps conver-sions from interfaces marked by B and C bound a 2025 km thick layer of relatively dense or fast mantle thatunderlies the southern Basin and Range. Additional weakPs observed beneath the Moho result from a combinationof reverberations from the shallow negative Ps andassorted small Ps from the upper mantle.

Crustal thickness and Vp/Vs estimates (Table 1) showpronounced variability amongst Basin and Range stations.Stations KITT, LEMN and SQRL show no increase incrustal thickness, an unexpected result for being situatedatop high and large ranges, and contain anomalously highVp/Vs values. The lack of a crustal root beneath LEMNdiffers from the interpretations by Myers and Beck [16],who detect an apparent deepening of the Moho at nearbyTUC from receiver functions that sample beneath thehigher elevations. Now with sufficient data to explore theVp/Vs-thickness relation, we offer an alternative explana-tion that the apparent crustal thickening beneath theCatalinaRincon MCC may result from a significantazimuthal variation of the crustal Vp/Vs ratio on a localscale. We evaluate this shift with the phase stackingtechnique by using 270 receiver functions that sampleoutside the MCC and 171 that sample in a narrow swathbeneath the range (Fig. 3). This distribution and density ofreceiver functions reveals a strong zonation in the Vp/Vswith a low value (1.71) measured external of the MCCcompared to a much higher Vp/Vs ratio (1.87) detectedbeneath it (Table 1). This result is also consistent with thehigh value (1.86) observed at site LEMN located at thesummit of the MCC. In fact the average Vp/Vsmeasurement of all MCCs is greater (1.83) than remainingBasin and Range stations (1.73) ignoring station DSRT.Measurements at ABBYand TUC curiously exhibit equalor slightly thicker crust than the higher elevation MCCssampled at KITT, LEMN and SQRL. The thinnest crustsurveyed at ASU also exhibits a higher Vp/Vs (1.77),which may arise from its proximity to the South MountainMCC, but we currently lack sufficient data to evaluate itsVp/Vs characteristic azimuthally in the same manner asachieved at TUC. Important to note is DSRTwhich is sitedat a noisy location in the TucsonMountains. The data showan outlierVp/Vs of 1.95 for this station,which recorded foronly a short duration in a high noise environment. For thisreason we exclude this station from any overarchinganalysis and interpretation of Vp/Vs trends.

Zandt and Ammon [17] measured the global crustalaverage for Vp/Vs inMesozoicCenozoic orogenic beltsto be 1.73, below our average (1.78). However non-MCC Vp/Vs estimates of 1.73 imply that overall Basinand Range crust matches the global average while theMCC value (1.83) may result from the presence of partialmelt or a significant shift in crustal composition. Thepartial melt scenario is unlikely because of the currenttectonic quiescence of MCC systems and lack ofNeogene volcanism in the area [18]. The presence ofmafic material would support a model for sizeablemantle input during MCC formation [19]. Howeveranalysis by Christensen [20], which illustrates thedependence of Poisson's ratio on mineral assemblage,describes the significant influence that plagioclasecontent has on the Vp/Vs ratio of a rock. While thesodium plagioclase end-member (albite) has a relativelylower Vp/Vs (1.81), the calcium end-member (anor-thite) has a very highVp/Vs (1.92). Thus the increasingpresence of calcic-rich plagioclase will drive the Vp/Vsratio of a crustal rock above the average value fororogenic belts and may be responsible for the elevatedvalues observed here.

These findings bring into consideration an alternativeinterpretation for post-Laramide tectonics; that the style ofcollapse of the Cordillera may have been controlled byspatial variations in the composition of the crust, asevidenced today by this divide of Vp/Vs estimatesbetween these different geologic environments. Alsoimportant and possibly connected is the absence ofcompensating roots beneath MCCs in the southern BasinandRange. This finding suggests that the traditional linearrelation between elevation and crustal thickness is invalidhere and that these edifices are supported via other means.

67A. Frassetto et al. / Earth and Planetary Science Letters 249 (2006) 62733.2. Colorado plateau and Arizona transition zone

Receiver functions in the Colorado Plateau andTransition Zone are characterized by substantially greaterseismic heterogeneity than the Basin and Range. Weobserve numerous Ps phases other than the Moho thatpossess high amplitude throughout the crust and uppermantle (Fig. 4). In addition all receiver functions appear tocontain a direct P that is offset from 0 km. This is probablythe result of the estimated 23 km of seismically slowsedimentary rocks that cap most of the Colorado Plateauand are exposed in the Grand Canyon. The base of thislayer creates a large near surface Ps that interferes withand slightly shifts the direct P arrival.

Crustal thickness ranges from 3643.5 km through-out the region (Table 2). Observed Ps from theMoho (M)are similar throughout the region (10% of the direct P)and generally smaller than in the Basin and Range withan exception at RENO (20%) (Fig. 4). Additional directPs phases disclose substantial stratification and seismicheterogeneity in the crust. Stations KNTH, WUAZ andZIZZ contain a large (1015%) mid-crustal arrival (B)from 1525 km that is the bottom of a slow wavespeedlayer. At deeper levels in the crust WUAZ contains a Ps(C) above the Moho at 34 km. Stations KNTH and ZIZZboth exhibit a sub-Moho Ps (D) at 4854 km whichillustrates a further increase in wavespeeds. TheFig. 4. Stacked Colorado Plateau receiver functions, grouped by structural trehave been identified with labels. M represents the estimated Moho.COCORP and PACE controlled source seismic profilingexperiments identify an interface in this depth range andinterpret it as possible deepMoho beneath sections of theColorado Plateau [13,21]. Locations where we observethe unusual converted phase appear to be limited tovolcanic centers and this feature may result frommagmatic migration through the lower crust and uppermantle. The complex pattern of several arrivals near, andjust below, the Moho at these stations suggests that atlocalized regions in the Colorado Plateau the crustmantle transition may be layered or more gradational(Fig. 4). A previous study of upper mantle wavespeedmeasures a well-constrained Pn of 8.12 km/s beneaththe Plateau. Thus a crustmantle boundary with multiplelayers would minimize a potentially large impedancecontrast between crust and relatively fast mantle andproduce a similar receiver function response to ourobservations. Receiver functions at KNTH and WUAZdistinctly express two converted phases with consistentarrival times at all azimuths, casting doubt on thepossibility that this intriguing feature results fromstacking azimuths which sample the Moho at differentdepths. Careful examination of the phase stacked plots atKNTH,WUAZ and ZIZZ reveal that despite larger areasof high amplitude generated by the complicated crustalstructure, the convergence peak interpreted as Mcontains amplitudes that are at least 5% greater thannds and proximity to volcanic centers. Commonly observed interfaces

Fig. 5. Relationship of elevation and crustal thickness for COARSEstations. A least squares trend-line is fit to all stations with the exclusionof those atopMCCs. Data from TUC is divided in the same manner as inFig. 3.

68 A. Frassetto et al. / Earth and Planetary Science Letters 249 (2006) 6273the next highest peak. Furthermore, these peaks provideestimates of crustal thickness close to that found at PFNPwhich possesses a single peak.

In the uppermost mantle a 722% negative Ps (E)appears at 5365 km beneath the survey area. Itsmagnitude is strongest beneath KNTH, WUAZ andZIZZ. This intriguing feature may signal a partial meltzone in the upper mantle which feeds recent volcanismwitnessed along the southern margin of the Plateau.Additionally a negative Ps (A) at 1020 kmwith 1530%amplitude (Fig. 4) everywhere except RENO appears tobe a PpSs+PsPs reverberation from the sediment cap. Atstations WUAZ and ZIZZ however the amplitude isnearly 30% and moveout analysis does not conclusivelyprove it to be reverberated energy. Its presence as a first-order structure may be explained by a zone of slowwavespeed created by partial melt in the crust. These largenegative arrivals are similar to ones associated with theAltiplanoPuna magma body in South America [22] andthe Socorro magma body in the Rio Grande Rift [23].

Phase stacking and moveout analyses indicate anaverage Vp/Vs ratio of 1.81 for Colorado Plateaustations, noticeably higher than the non-MCC relatedBasin and Range and global average Vp/Vs values andpossibly indicative of greater mafic material or partialmelt within the crust (Table 2). Separate results fromGilbert et al. [24] also support this observation. WUAZpossesses an unusually high Vp/Vs value (1.90) inaddition to the large negative Ps following the Mohowhich suggests a slow wavespeed region in the uppermantle. When viewed in proximity to the nearbyvolcanic centers it provides possible evidence for thepresence of partially molten crust and upper mantle nearthe currently active San Francisco Volcanic Field [25].Azimuthal variability of the Ps phases displays reducedamplitudes within a zone near the most recent volcanismat Sunset Crater. Melt dominated material in the mid- toupper-crust may be responsible for the reduction inconverted energy. These findings coincide with thelocation of substantial volumes of Quaternary volcanismviewed along the Colorado Plateau rim and in theArizona transition zone [18,26]. Unlike WUAZ, ZIZZdoes not have a definitive and station-wide increase inVp/Vs ratio that indicates melt, but instead showssubstantial noise and azimuthal heterogeneity whichmakes a reasonable Vp/Vs value difficult to constrain.Like DSRT we have excluded the Vp/Vs ratio at ZIZZfrom broader interpretations.

Although crustal thickness, isostatic support andmeansof uplift for the Colorado Plateau are thoroughly debatedand the subject of numerous studies (e.g., [2732]) nomodel manages to unify the numerous seismic, gravityand heat flow observations for the region. With ourdeployment we provide new seismic observations (Fig. 4)for the Colorado Plateau which help clarify our under-standing of the region. The range of crustal thicknessestimates (3643.5 km) generally matches those fromprevious geophysical investigations that found crust of3542 km in nearby regions [13,33]. Both the uppermantle low velocity zone [34] and layered crust [35]previously seen across the central and western ColoradoPlateau are present in measurements here. The weakMoho Ps and large negative upper mantle Ps indicatingrelatively slow upper mantle wavespeed beneath the studyarea concurs with the calculation by Roy et al. [30] for alayer of lower density upper mantle residing beneath thePlateau's eastern edge. All COARSE stations demonstratethick and stratified crust beneath the Colorado Plateau.The exact cause for this zone of substantial shearwavespeed variation is unknown, but the layering andhigh average Vp/Vs ratio may be linked to the hypothesisof crustal growth via magmatic addition [29].

4. Support of high elevation in the Basin and Range

The complex and surprising nature of seismic observa-tions for the southern Basin and Range warrants furtherconsideration and is the primary focus here. Estimateddepth for the Moho beneath the MCCs clearly deviatesfrom the traditional linear relationship between crustalthickness and elevation (Fig. 5). An Isostatic ResidualGravity Map by Simpson et al. [36] shows variation of the

69A. Frassetto et al. / Earth and Planetary Science Letters 249 (2006) 6273calculated isostatic anomaly due to the presence or absenceof an Airy root. An anomaly of zero mgal serves asevidence that the region is balanced by a root necessary toaccount for the observed elevation and Bouguer gravityanomaly. Negative or positive excursions result from acorresponding mass deficit or surplus created by thecorrection. Stations ABBY, DSRTand TUC coincide withanomalies of approximately zero mgal while KITT,LEMN and SQRL exist within 2030 mgal lows. Theselow values demonstrate that the isostatic correction is toolarge and when compared with receiver function datasupport our observation that none of the MCCs surveyedby COARSE are supported by a crustal root.

4.1. Density models

Even geodynamic models that use a flexural responseto support MCC elevation still require a local thickeningof the crust approaching 10 km to compensate for theobserved elevation [37]. We investigate the possibilitythat lower density crust may allow for high elevationwithout a corresponding crustal root. To test thishypothesis we use the ISOBAL isostasy modelingprogram [Clem Chase, personal communication]. Thissoftware balances a user specified reference columnagainst a model with density or layer thickness allowed tovary, and also calculates a 1D Bouguer anomaly for eachcolumn. For this analysis we test three crustal referencemodels and use them to calculate the correspondingcrustal density of the CatalinaRinconMCC (Fig. 6). Forthe MCC, we hold the seismically determined crustalthickness (29 km), the approximate average elevation(1.8 km), and the estimated density of the mantle(3250 kg/m3) fixed. We then balance the two columnsand adjust the crustal density until it matches the observeddifference between the Bouguer gravity anomalies eachlocation [38]. This final step is needed to account for theFig. 6. Three isostaticmodel pairs for density structure. Crust andmantle are deno(A) BuckskinRawhideMCC, (B) Basin and Range in southeastern Arizona, anat the CatalinaRincon MCC. The mantle value indicates the necessary densitydensity in each case.linearly trending, long wavelength decrease in theBouguer gravity field across southern Arizona [33].

Two of the reference cases (Fig. 6A,B) are based on thedetailed seismic profiling and gravity modeling of theBuckskinRawhide MCC and the surrounding Basin andRange [9,39]. The third reference case (Fig. 6C) uses ourseismic constraints for relatively unextended crust (ABBY)and an average density for the crust in the region [ClemChase, personal communication]. All three solutions fortheMCC require a density decrease of 3% (27202784 kg/m3) relative to the standard Basin and Range crust(2810 kg/m3), the BuckskinRawhide MCC (2850 kg/m3), and the relatively unextended case (2830 kg/m3).

We consider differences in density distribution to be amajor factor for the high elevation of the CatalinaRincon and other MCCs in our study area. Relativelylight crust beneath these features satisfies the observedgravity low while presenting a stable environment tomaintain high elevation. Lower density crust atopnormal density mantle would also produce the largeseismic impedance contrast observed at the Moho onreceiver functions. The alternative, a 1% less densemantle of 32063233 kg/m3 (noted in yellow, Fig. 6),would presumably generate a higher magnitude andlonger wavelength gravity low that is unobserved anddiminish the significant impedance contrast at theMoho.However a recent study by Abers et al. [40] indicates thatlighter density mantle may provide a significant sourceof support for young MCCs and this finding prevents usfrom ruling out a mantle contribution in this case.

4.2. Compositional models

Comparing modeled densities to the compositions ofMCC rocks helps to validate our model findings as wellas include the independently constrained Vp/Vs esti-mates as we seek to determine the composition andtedwith different shades. The left columnof eachpair is the reference case:d (C) Station ABBY. The crustal value in the crust is the calculated densityto produce the same observed gravity if the crust is kept as the reference

Table 4Model compositions and calculated seismic properties for deepintrusions beneath the CatalinaRincon metamorphic core complex

Mineralweight %

Model 1 Model 2 Model 3

Quartz 20 10 5K-Feldspar 5 0 0Anorthite 20 30 35Albite 30 35 35Biotite 20 20 20Hornblende 5 5 5 (kg/m3) 2726 2737 2741Vp (km/s) 6.05 6.15 6.21Vs (km/s) 3.41 3.39 3.37Vp/Vs 1.77 1.82 1.84

The Vp/Vs ratio changes show the significant effect of compositionalvariations on the seismic response of a material.

Table 3Average compositions and calculated seismic properties for theOracleRuin granite (Yo), Leatherwood suite and equivalent (Kg),Wilderness suite (Tw), and Catalina granite and equivalent (Tg)intrusions in the CatalinaRincon metamorphic core complex

MineralWeight %

Yo Kg Tw Tg

Quartz 33.0 17.8 28.4 34.5K-Feldspar 28.0 7.3 29.2 35.3Anorthite 14.0 17.5 6.5 6.4Albite 14.0 27.9 28.1 16.6Biotite 9.0 18.6 3.3 4.4Muscovite 0.0 0.0 2.4 0.6Hornblende 0.0 5.2 0.0 0.2Other 0.0 4.9 0.9 1.2 (kg/m3) 2664 2738 2650 2644Vp (km/s) 6.04 6.12 6.06 6.03Vs (km/s) 3.51 3.45 3.50 3.52Vp/Vs 1.72 1.77 1.73 1.71

Compositions from Keith et al. [38].

70 A. Frassetto et al. / Earth and Planetary Science Letters 249 (2006) 6273means of support of the MCCs studied here. We applythe method of Hacker and Abers [41] to estimate rockcomposition near the area sampled by seismic measure-ments. Mapping and petrologic analysis of the CatalinaRincon MCC show that the area is underlain by fourdistinct suites of intrusive bodies (Table 3); the region-wide Oracle granite (1.4 Ga), the more locally exposedLeatherwood suite plutons (7565 Ma), Wildernessgranite (5044 Ma), and Catalina granite andequivalents (3025 Ma) (see Keith et al. [42] for asummary). We compute average composition for eachintrusion and model their seismic properties for mid-crustal conditions of 0.6 GPa and 400 C. The resultsagree with the general results of Christensen [20]; thatplagioclase content, along with variability of quartz andmafic minerals, significantly affects the seismic proper-ties of a rock. The Leatherwood suite intrusions,corresponding temporally to significant late Cretaceousvolcanic deposits exposed in the nearby TucsonMountains [43], have a 34% increase in density andVp/Vs ratio relative to the basement granitoids andsucceeding intrusive episodes. The modeled Vp/Vsalong with the seismically determined observations ofabnormally high Vp/Vs at all MCCs surveyed heresuggest that rocks with similar properties to theLeatherwood suite intrusions may constitute a substan-tial volume of the crust beneath the CatalinaRinconMCC and possibly other MCCs in southern Arizona.

Although themodeledVp/Vs ratio for the Leatherwoodsuite (1.77) exceeds that of the Oracle granite basement(1.72), Wilderness granite (1.73) and Catalina granite(1.71) it still fails to reproduce the value observed fromreceiver functions. As the higher Vp/Vs value within theCatalinaRincon crustal block results from the prevalenceof the anorthite and biotite-rich Leatherwood suitematerial, presumably similar, and possibly more plagio-clase and mafic mineral-rich material exists at deeperlevels. To test this hypothesis we create three models fordeep crustal composition (Table 4), designed to explore theeffects of adjusting specific mineral content at the samemid-crustal PT conditions (0.6 GPa, 400 C). Buildingupon the initial Leatherwood suite composition we test amaterial with 20% quartz, 5% K-feldspar, 30% anorthite,20% albite, 20% biotite and 5% hornblende and using theHacker and Abers [41] calculation arrive at a Vp/Vs ratioof 1.77 and a density of 2726 kg/m3 which is still lowerthan our observed values for TUC/LEMN, KITT andSQRL. To determine a composition with a larger Vp/Vsvalue we eliminate K-feldspar and hold hornblende andbiotite constant while decreasing quartz and increasingplagioclase. Thismaterial, approximate in composition to aquartz diorite, maintains a relatively low density (2741 kg/m3) but has an increased Vp/Vs value (1.84). This finalmodel has a density increase of 0.5% while the Vp/Vs hasincreased by 4%, illustrating that compositional changescan often influence Vp/Vs more noticeably than density.These results support the findings of Christensen [20] andillustrate the important connection between compositionthat varies by relatively minor amounts and large changesin the material Vp/Vs. The calculated Vp/Vs and densityalso match closely the seismically constrained Vp/Vs andisostatically modeled density (Figs. 3 and 6).

The constraints from our seismic observations andisostatic modeling allow us to estimate the lower crustalcomposition of the CatalinaRincon MCC and, bysimilarity of seismic observations, the possible crustalcomposition for other MCCs in this study. There areseveral important points regarding thesemodel results and

71A. Frassetto et al. / Earth and Planetary Science Letters 249 (2006) 6273their implications. The estimated Vp of the crust for anyactual or modeled composition does not exceed 6.21 km/s. This value is near the assumed 6.2 km/s wavespeedused for depth conversion of the receiver functions, andprovides confidence in the crustal thickness and Vp/Vsestimates. Secondly, to produce the observed Vp/Vsvalues of the bulk crust at TUC/LEMN and SQRL, theintrusive bodies in the mid- to lower crust require a Vp/Vsratio above 1.90, higher than our models produce. It isalso unknown how deformation and metamorphism havecontributed to the composition of the lower crust. Finally,these results appear to form a paradox. The crustal columnof CatalinaRincon MCC has a high Vp/Vs ratio yet alsois light enough to maintain its high elevation. The densityVp/Vs trend calls for either a neutral or positivecorrelation for the compositions examined (Tables 3and 4). Thus, density of the region surrounding MCCsmay be as low or lower to coincide with lower observedVp/Vs, though Moho Ps amplitude on receiver functionsdoes not explicitly support this possibility. Because of thisinferred density heterogeneity, the broken plate flexuralmodel byHolt et al. [37] is especially appealing to explainthis observation due to the fact that it decouples theuplifted CatalinaRincon crustal block from the sur-rounding Basin and Range. A component of buoyancy inthe mantle beneath MCCs in this region may also providea source of buoyancy that cannot be explained by crustaldensity differences.

Another geodynamic model by Block and Royden [44]questions the applicability of Airy isostasy in MCCformation. Their model proposes that flowing lower crustin detachment systems moves laterally and vertically to fillvoids created during extension, thus maintaining a flatMoho geometry while emplacing domes of lower crustmaterial. Our seismic observations of a subhorizontalMoho support this hypothesis and flowing lower crust mayprovide a mechanism to create the Vp/Vs anomalies bytransporting higher Vp/Vs material underneath a MCCfrom the surrounding regions. However, this processwouldpresumably raise the density as well. Thus the applicabilityof this model relies on a balance between emplacing asufficient amount of lower crust with a composition thatwould raise Vp/Vs, but not increase the density beyond ourrange of estimates. In this situation, crustal flow wouldhave likely occurred long after emplacement of the highVp/Vs Laramide age plutons. The tectonic practicality ofsuch a model is beyond the scope of this study.

5. Concluding remarks

These new receiver function observations from theColorado Plateau and Basin and Range illustrate the fun-damental differences in crustal thickness, structure andcomposition between these tectonic provinces. Despitetheir geographic proximity it is clear from seismic obser-vations that late Mesozoic and Cenozoic tectonic activityhas imparted a pronounced difference in the seismiccharacter of these regions. The upper lithosphere of theColorado Plateau possesses greater crustal thickness,increased crustal layering, higher than average Vp/Vsvalues, areas of gradational Moho, and possible evidenceof partialmelt at both deep and shallow levels (Fig. 4, Table2). The Basin andRange has thinner crust, lacks significantlayering and contains observed Vp/Vs values that suggestan average crustal composition (Fig. 2, Table 1), except atMCCs where the Vp/Vs ratio is much higher thansurrounding areas. Through extension the southern Basinand Range has maintained a crust of uniform thicknessdespite major differences in elevation and inferred throughthe Vp/Vs values, composition. The large variance in theVp/Vs ratio differs from previous observations that suggestsimilar crustal compositions within the Basin and Rangeand Colorado Plateau (e.g., [13]).

Both seismic and gravity measurements of the MCCsimply the absence of a supporting root. Seismicobservations also detect a sharp Moho and significantlyhigher Vp/Vs ratio than the surrounding Basin and Range.Modeling of crustal density for the CatalinaRinconMCC results in a lower density crust than referencemodels based on the seismically constrained thicknesses.Our findings show that MCCs in the southern Basin andRange retain a substantial portion of their elevation due toa locally buoyant column of crust and possibly uppermantle. Results of compositional modeling (Table 4)allow us to speculate that the crust of the CatalinaRinconMCC and its neighboring MCCs consists primarily of aplagioclase end-member anorthite-rich and quartz-poormaterial that may relate to the emplacement of theLeatherwood suite plutons during a significant Laramide-age magmatic episode. It also appears that magmatic ordeformational processes during MCC formation mayhave sharply demarcated the crustupper mantletransition as seen fromhigh amplitudeMohoPs convertedphases on receiver functions (Fig. 2).

It is important to consider that the anomalous Basinand Range Vp/Vs values are only related to areas thathave undergone significant deformation and exhumationduring the mid-Tertiary extension. We assert that Vp/Vsanomalies like those observed beneath the CatalinaRincon MCC and others signify substantial variability incomposition of the crust, and that these differences mayhave altered the distribution of strength in the crust andas a result primed certain areas for extension andexhumation during the process of orogenic collapse.

72 A. Frassetto et al. / Earth and Planetary Science Letters 249 (2006) 6273Acknowledgements

A special thanks goes to the COARSE team at bothArizona State and University of Arizona for theirdedicated work during the installations, service runs,and database maintenance. We are grateful to numerousproperty owners, the Bureau of Land Management,National Forest Service, and National Park Service forgranting us permission to use their land. We especiallyacknowledge the help of Gayle Zizzo of the Departmentof Geosciences, Bob Peterson of the University ofArizona Steward Observatory, John Dunlap and ClaudePlymate at the Kitt Peak Observatory, John Ratje andJohn Waack at the Mt. Graham binocular telescope, andBob and Kathy Minitti. We thank Drs. Mihai Ducea,Clem Chase, Roy Johnson, Jon Patchett, RandyRichardson, Trey Wagner, and Tom Owens for illumi-nating discussions concerning this study. We usedsoftware written by Chuck Ammon, Arda Ozacar andClem Chase for the data analysis. Global SeismographicNetwork (GSN) is a cooperative scientific facilityoperated jointly by the Incorporated Research Institu-tions for Seismology (IRIS), the United States Geolog-ical Survey (USGS), and the National ScienceFoundation (NSF). The facilities of the IRIS DataManagement System, and specifically the IRIS DataManagement Center, were used for access to waveformand metadata required in this study. The IRIS DMS isfunded through the National Science Foundation andspecifically the GEO Directorate through the Instrumen-tation and Facilities Program of the National ScienceFoundation under Cooperative Agreement EAR-0004370. This work was made possible by fundingfrom the National Science Foundation through theGraduate Research Fellowship Program, the IRISUndergraduate Internship Program, and the Universityof South Carolina.References

[1] W.R. Dickinson, Cenozoic plate tectonic setting of theCordilleran region in the United States, in: J.M. Armentrout,M.R. Cole, H. TerBest (Eds.), Cenozoic Paleogeography of theWestern United States, Pacific Section, . Soc. Econ. Pal. Min.,1979, pp. 113.

[2] P.J. Coney, T.A. Harms, Cordilleran metamorphic core com-plexes: Cenozoic extensional relics of Mesozoic compression,Geology 12 (1984) 550554.

[3] L.J. Sonder, C.H. Jones, Western United States extension: howthe west was widened, Annu. Rev. Earth Planet. Sci. (1999)2741727462.

[4] G. Zandt, S.C. Myers, T.C. Wallace, Crust and mantle structureacross the Basin and RangeColorado Plateau boundary at 37Nlatitude and implications for Cenozoic extensional mechanism,J. Geophys. Res. 100 (1995) 1052910548.

[5] J. Ligorria, C.J. Ammon, Iterative deconvolution of teleseismicseismograms and receiver function estimation, Bull. Seismol.Soc. Am. 89 (1999) 13951400.

[6] C.A. Langston, Structure under Mount Rainier, Washington,inferred from teleseismic body waves, J. Geophys. Res. 84(1979) 47494762.

[7] T.J. Owens, S.R. Taylor, G. Zandt, Isolation and enhancement of theresponse of local seismic structure from teleseismic P-waveforms,Lawrence Livermore Laboratory Internal Report, 1983 33 pp.

[8] N.I. Christensen, W.D. Mooney, Seismic velocity structure andcomposition of the continental crust: a global view, J. Geophys.Res. 100 (1995) 97619788.

[9] J. McCarthy, S.P. Larkin, G. Fuis, R.W. Simpson, K.A.Howard, Anatomy of a metamorphic core complex: seismicrefraction/wide-angle reflection profiling in southeastern Cali-fornia and western Arizona, J. Geophys. Res. 96 (1991)1225912291.

[10] D.H. Warren, A seismic refraction survey of crustal structure incentral Arizona, Geol. Soc. Amer. Bull. 80 (1969) 257282.

[11] Y.A. Sinno, G.R. Keller, M.L. Sbar, A crustal seismic refractionstudy in west-central Arizona, J. Geophys. Res. 86 (1981)50235038.

[12] D.M.Gish,G.R.Keller,M.L. Sbar,A refraction study of deep crustalstructure of the Basin and RangeColorado Plateau transition zonein eastern Arizona, J. Geophys. Res. 86 (1981) 60296038.

[13] T. Parsons, J. McCarthy, W.M. Kohler, C.J. Ammon, H.M. Benz,J.A. Hole, Crustal structure of the Colorado Plateau, Arizona:application of new long-offset seismic data analysis techniques,J. Geophys. Res. 101 (1996) 1117311194.

[14] L. Zhu, H. Kanamori, Moho depth variation in southernCalifornia from teleseismic receiver functions, J. Geophys. Res.105 (2000) 29692980.

[15] H. Gurrola, J.B. Minster, T.J. Owens, The use of velocityspectrum for stacking receiver functions and imaging uppermantle discontinuities, Geophys. J. Int. 117 (1994) 427440.

[16] S.C. Myers, S.L. Beck, Evidence for a local crustal root beneaththe Santa Catalina metamorphic core complex, Arizona, Geology22 (1994) 223226.

[17] G. Zandt, C.J. Ammon, Continental crust composition con-strained by measurements of crustal Poisson's ratio, Nature 374(1995) 152154.

[18] D.J. Lynch, Neogene Volcanism in ArizonaThe recognizablevolcanoes, in: J.P. Reynolds, J.P. Jenney (Eds.), Geol. Soc. Digest,vol. 17, Geologic evolution of Arizona, Tucson, Arizona, 1989,pp. 681700.

[19] G.S. Lister, S.L. Baldwin, Plutonism and the origin ofmetamorphic core complexes, Geology 21 (1993) 607610.

[20] N.I. Christensen, Poisson's ratio and crustal seismology,J. Geophys. Res. 101 (1996) 31393156.

[21] E.C. Hauser, J. Lundy, COCORP Deep reflections: Moho at50 km (16 s) beneath the Colorado Plateau, J. Geophys. Res. 94(1989) 70717081.

[22] J. Chmielowski, G. Zandt, C. Haberland, The Central AndeanAltiplanoPuna magma body, Geophys. Res. Lett. 26 (1999)783786.

[23] K.E. Sheetz, J.W. Schlue, Inferences for the Socorro magmabody from teleseismic receiver functions, Geophys. Res. Lett. 19(1992) 18671870.

[24] H. Gilbert, A. Velasco, G. Zandt, Preservation of Proterozoicterranes within the Colorado Plateau: a source of strength? First

73A. Frassetto et al. / Earth and Planetary Science Letters 249 (2006) 6273Annual Earthscope National Meeting, Santa Ana Pueblo, NM,2005.

[25] K.L Tanaka, E.M Shoemaker, G.E. Ulrich, E.W. Wolfe,Migration of volcanism in the San Francisco volcanic field,Arizona, Geol. Soc. Amer. Bull. 97 (1986) 129144.

[26] T.C. Moyer, L.D. Nealey, Compositional regional of variationslate Tertiary bimodal rhyolite lavas across the Basin and Range/Colorado Plateau boundary in western Arizona, J. Geophys. Res.94 (1989) 77997816.

[27] N. Beghoul, M. Barazangi, Mapping high Pn velocity beneaththe Colorado Plateau constrains uplift models, J. Geophys. Res.94 (1989) 70837104.

[28] C.G. Chase, J.C. Libarkin, A. Sussman, Colorado Plateau: geoidand means of isostatic support, Int. Geol. Rev. 44 (2002) 575587.

[29] P. Morgan, C.A. Swanberg, On the Cenozoic uplift and tectonicstability of the Colorado Plateau, J. Geodyn. 3 (1985) 3963.

[30] M. Roy, J.K. McCarthy, J. Selverstone, Upper mantle structurebeneath the eastern Colorado Plateau and Rio Grande rift revealedbyBouguer gravity, seismic velocities, and xenolith data,Geochem.Geophys. Geosyst. 6 (2005), doi:10.1029/2005GC001008.

[31] J. Spencer, Uplift of the Colorado Plateau due to lithosphereattenuation during Laramide low-angle subduction, J. Geophys.Res. 101 (1996) 1359513609.

[32] M. West, J. Ni, W.S. Baldridge, D. Wilson, R. Aster, W. Gao, S.Grand, Crust and upper mantle shear wave structure of the southwestUnited States: implications for rifting and support of high elevation,J. Geophys. Res. 109 (2004), doi:10.1029/2003JB002575.

[33] J.D. Hendricks, J.B. Plescia, A review of the regional geophysicsof the Arizona transition zone, J. Geophys. Res. 96 (1991)1235112373.

[34] L.A. Lastowka, A.F. Sheehan, J.M. Schneider, Seismic evidencefor partial delamination model for Colorado Plateau uplift,Geophys. Res. Lett. 28 (2001) 13191322.

[35] H.J. Gilbert, A.F. Sheehan, Images of crustal variations in theintermountain west, J. Geophys. Res. 109 (2004), doi:10.1029/2003JB002730.

[36] R.W. Simpson, R.C. Jachens, R.J. Blakely, R.W. Saltus, A newisostatic residual gravity map of the conterminous United Stateswith a discussion on the significance of isostatic residual anomalies,J. Geophys. Res. 91 (1986) 83488372.

[37] W.E. Holt, C. Chase, T. Wallace, Crustal structure from three-dimensional gravity modeling of a metamorphic core complex: amodel of uplift, Santa CatalinaRincon mountains, Arizona,Geology 14 (1986) 927930.

[38] R. E. Sweeney, P. L. Hill, Arizona Aeromagnetic and GravityMaps and Data: A Web Site for Distribution of Data, U.S.G.S.Open File Report 01-0081 (2001).

[39] T. Parsons, J. McCarthy, G.A. Thompson, Very Different CrustalResponse to Extreme Extension in the Southern Basin and Rangeand Colorado Plateau Transition, in: M.C. Erskine, J.E. Faulds, J.M. Bartley, P.D. Rowley (Eds.), Am. Assoc. Pet. Geols., PacificSection, Guidebook, vol. 78, 2001, pp. 291304.

[40] G.A. Abers, A. Ferris, M. Craig, H. Davies, A.L. Lerner-Lam, J.C.Mutter, B. Taylor, Mantle compensation of active metamorphiccore complexes at Woodlark rift in Papua NewGuinea, Nature 418(2002) 862865.

[41] B.R. Hacker, G.A. Abers, Subduction Factory 3: an excelworksheet and macro for calculating the densities, seismic wavespeeds, and H2O contents of minerals and rocks at pressure andtemperature, Geochem. Geophys. Geosyst. 5 (2004),doi:10.1029/2003GC000614.

[42] S.B. Keith, S.J Reynolds, P.E. Damon, M. Shafiqullah, D.E.Livingston, P.D. Pushkar, Evidence for multiple intrusion anddeformation within the Santa CatalinaRinconTortolitacrystalline complex, southeastern Arizona, in: M.D. Critten-den, P.J. Coney, G.H. Davis (Eds.), Cordilleran MetamorphicCore Complexes, Geol. Soc. Am. Memoir., vol. 153, 1980,pp. 217267.

[43] W.R. Dickinson, Tectonic setting of faulted Tertiary strataassociated with the Catalina core complex in southern Arizona:special paper, Geol. Soc. Amer. Bull. 264 (1991) 106.

[44] L. Block, L.H. Royden, Core complex geometries andregional scale flow in the lower crust, Tectonics 9 (1990)557567.

http://dx.doi.org/10.1029/2005GC001008http://dx.doi.org/10.1029/2003JB002575http://dx.doi.org/10.1029/2003JB002730http://dx.doi.org/10.1029/2003JB002730http://dx.doi.org/10.1029/2003GC000614

Support of high elevation in the southern Basin and Range based on the composition and architec.....Tectonic setting and broadband deploymentTeleseismic receiver function analysesReceiver function observations and interpretationsBasin and rangeColorado plateau and Arizona transition zone

Support of high elevation in the Basin and RangeDensity modelsCompositional models

Concluding remarksAcknowledgementsReferences