TECTONICS, VOL. 19, NO. 1, PAGES 1-24 FEBRUARY 2000

Present-day motion of the Sierra Nevada block and some tectonic implications for the Basin and Range province, North American Cordillera

TimothyH. Dixon,• MeghanMiller, 2 FredericFarina, • Hongzhi Wang • and Daniel Johnson 2

Abstract. Global Positioning System (GPS) data from some continentalareas appear to behave like oceanic five sites on the stable interior of the Sierra Nevada block lithospherein one important respect:deformation is are inverted to describeits angular velocity relative to occasionallyconcentrated in narrow fault zones that stableNorth America. The velocity data for the five sites accommodaterelative motion betweenrigid blocks. The fit the rigid block model with rms misfits of 0.3 mm/yr Sierra Nevada block in the western United States is (north) and 0.8 mm/yr (east), smaller than independently probablya good exampleof a rigid continentalblock estimateddata uncertainty,indicating that the rigid block [Wright,1976]. Seismicityaround its marginsdelineates model is appropriate. The new Euler vector, 17.0øN, an aseismicregion east of the SanAndreas fault and west 137.3øW, rotation rate 0.28 degreesper million years, of the Basin and Range extensionalprovince (Figure 1). predictsthat the block is translatingto the northwest, However,a rigoroustest of the rigidity of this blockhas nearly parallel to the plate motion direction, at 13-14 neverbeen performed (nor, to ourknowledge, for anyother mm/yr, fasterthan previous estimates. Using the predicted continentalblock or microplate). Perhapsthe Sierra Sierra Nevada block velocity as a kinematic boundary Nevadablock is not rigid at all, and the absenceof condition and GPS, VLBI and other data from the interior seismicityindicates a weak block undergoingdiffuse and margins of the Basin and Range, we estimatethe aseismic deformation, or strain accumulation on a few velocitiesof some major boundary zone faults. For a locked faults cutting the block, previously assumed transect approximately perpendicularto plate motion inactiveor activeat low strainrates (e.g., the Kern Canyon throughnorthern Owens Valley, the easternCalifornia shear fault),to bereleased in futurelarge earthquakes. zone(westem boundary of the Basin and Range province) Spacegeodesy can rigorously test the conceptof rigid accommodates11+1 mm/yr of right-lateralshear primarily continentalblocks, just as it can measurethe angular on two faults, the Owens Valley-White Mountain (3+2 velocityand test the rigidity of largerplates [Argus and mm/yr) and Fish Lake Valley (8+2 mm/yr) fault zones, Gordon,1996; Dixon et al., 1996]. Sincethe motionsof basedon a viscoelasticcoupling model that accountsfor rigidblocks or plateson a spherecan be describedby Euler the effectsof the 1872 Owens Valley earthquakeand the (angularvelocity) vectors [e.g., Chase,1978; Minster and rheologyof the lower crust. Togetherthese two faults, Jordan, 1978; DeMets et al., 1990], we can test whether separatedby less than 50 km on this transect,define a velocitydata from siteson a continentalblock are well fit region of high surfacevelocity gradient on the eastern by a givenEuler vector, and whethervelocity predictions boundaryof the SierraNevada block. The WasatchFault onthe marginsof the blockbased on this Eulervector are zone accommodatesless than 3+1 mm/yr of east-west consistent with other observations. An accurate estimate of extensionon the easternboundary of the Basin and Range Sierra Nevada block motion also holds the key to province. Remaining deformationwithin the Basin and understandingcertain aspects of Pacific-NorthAmerican Rangeinterior is alsoprobably less than 3 mm/yr. plateinteraction [e.g., Atwater, 1970; 1989; Dixon et al., 1995; Wellset al., 1998;Hearn and Humphreys,1998]. 1. Introduction Minster and Jordan [1987] first appliedspace geodetic data to estimate motion of the Sierra Nevada block, Plate boundaries within continents are often definingnorthwest motion of the block at a velocityof characterizedby diffusezones of deformation,quite distinct -10+2 mm/yr. Argus and Gordon [1991] estimatedan from the narrowplate boundariesthat characterizeoceanic Euler vector for the block, defining counterclockwise lithosphereand define large rigid plates. However,at least rotationof the block about a pole locatedclose to and southwestof the block, and northwest to north-northwest motionof the block at 9-11 mm/yr, dependingon location. Both of thesepioneering studies used very long baseline • RosenstielSchool of Marineand Atmospheric Science, interferometry(VLBI) data from a few sites near the Universityof Miami, Miami, Florida. 2 deformingeastern margin of the block. The smallnumber Departmentof Geology, Central WashingtonUniversity, of sitesavailable at the time precludeda test of the rigid Ellensburg, Washington. block hypothesis,while the site locationslimited the accuracyof the relativemotion estimatesfor the Sierra Copyright2000 by the AmericanGeophysical Union. Nevadablock, sincethey are not actuallylocated on the stableblock (Figure 1). New Global PositioningSystem Paper number 1998TC001088. (GPS)data from siteson the stableinterior and margins of 0278-7407/00/1998TC001088512.00 the block are now available, enabling us to perform a 2 DIXON ET AL ß PRESENT-DAY SIERRA NEVADA MOTION

Longitude (W) 120 ø 116 ø 112 ø 40 ø

NA • o Z 38

_• 36 ø i.

PA ß

%. 34 ø 236 ø 238 ø 240 ø 242 ø 244 ø 246 ø 248 ø 250 ø Longitude (E)

Figure 1. Regionalneotectonic map for the westernUS, and (inset) major bounding plates: PA, Pacific plate, JF, Juande Fuca plate, NA, stableNorth America. Active faults on margins of SierraNevada block (SNB) are from Jenningsand Saucedo[1994]: bsf, Bartlett Springsfault; dvfcf, Death Valley-Furnace Creek fault zone; gf, Garlock fault; gvf, Green Valley fault; hcf, Hunting Creek fault; hlf, Honey Lake fault; hmpvf, Hunter Mountain-PanamintValley fault zone; mf, Maacamafault zone; mvf, Mohawk Valley fault zone; rcf, Rogers Creek fault zone; ovf, Owens Valley fault zone; saf, San Andreas fault. CNSB, Central Nevada seismic belt. Faults in eastern Basin and Range (BAR) are from Smith and Arabasz [1991] and include active faults and Late Cenozoic faults that are possibly active; wf, Wasatch fault zone. Seismicity from the University of California, Berkeley catalog at http://quake.geo.berkeley.edu/cnss/catalog-search.htmlshows all events after 1960 with magnitude>4.5 and depth less than 30 km. Note paucity of events within SNB. Spacegeodetic sites within the seismically active easternboundary of SNB and Basin and Range interior shown as open triangles with four letter identifiers. Sites on SNB are omitted for clarity here, but are shown in Figure 3 (location outlined by light solid line).

rigoroustest of the rigid block hypothesis,and refineour the interiorof the SierraNevada block, on its eastern understandingof Sierra Nevada motion and associated margin,and in theinterior of theBasin and Range province crustal deformation. These data, and the new (Figures1 and3; Table 1); interpretationsthey afford, are the focusof thispaper. 2. Observationswere made at semi-permanent, continuouslyrecording stations in California,on or near 2. Observationsand Data Analysis theSierra Nevada block, including Quincy (QUIN), part of theInternational GPS Service (IGS) network, operating for The dataused in this studywere obtained in several morethan 5 years;Columbia (CMBB) in California,part ways: of theBay AreaRegional Deformation (BARD) network, 1. Annual campaignswere conductedin August- operatingfor about5 years;ORVB, UCD1 and SUTB, Septemberbetween 1993 and 1998, althoughnot all sites alsopart of the BARD network,operating for about2 were occupiedeach year. Most sites were observedfor 24 years.Stations operating for lessthan 2 years(as of March hoursper day for 3-5 days,in campaignsin 1993, 1994, 1999)are not considered in thisreport; 1995 and 1998 (Figure2). Where available,the resulting5 3. Observationswere made at 16 semi-permanent, yeartime seriesprovide accurate site velocity data, as GPS continuouslyrecording stations widely distributed through velocityerrors depend strongly on the total time spanof easternand centralNorth America,operated by various observations[Mao et al., 1999]. These sites are locatedon agencies,used to definea stableregional reference frame for DIXON ET AL.' PRESENT-DAY SIERRA NEVADA MOTION 3

_ '''1'''1'''1'''1'''1''' 50-

o_ -50

50- _

o _

50-

_ o

-50 -

50-

-50

50

-50 OVRO North OVRO East OVRO Height

94 95 96 97 98 99 94 95 96 97 98 99 94 95 96 97 98 99 Year

Figure 2. GPS positions(relative to ITRF-96) as a functionof time, with an arbitraryconstant removed. Error bars(Table 3) omittedfor clarity. Slopeof bestfit line from weightedleast squaresgives velocitiesin Table I. 4 DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION

5O

-5O

5O

-5O

5O

ß

E -50 ß I::: 50 (" o * i o ß -5o 0 50

-5o

50- ** _

-50 -

i i I i i i i 1 i I i i r [ i i I i ' 5O

0 : I

-50

, WMTNNoah • WMTNEast • WMTNHeight '''1'''1'''1'''1'''1'''1'''1'''1'''1'''1'''1'''1'''1'''1'''1'''1'''1'''1 94 95 96 97 98 99 94 95 96 97 98 99 94 95 96 97 98 99 Year

Figure 2. (continued) DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION 5

Table 1. GPS Site Velocities Relative to ITRF-96

Latitude, Longitude, V•19•:ity, mm/yr degN degW North West Vertical

CEDA (CedarCreek) 35.75 118.59 -3.6 + 0.9 19.3 + 1.2 -1.4 + 2.9 CMBB(Columbia) a 38.03 120.39 -5.1 + 0.7 19.4 + 1.4 -9.6 + 3.1 GFLD (Goldfield) 37.82 117.36 -11.0 + 1.6 10.8 + 2.7 2.6 + 5.8 ELYA (Ely) 39.29 114.84 -11.6 + 0.8 14.8 + 1.1 -2.5 + 2.9 KMED (KennedyMeadows) 36.02 118.14 -3.0 + 0.6 18.5 + 0.9 -1.5 + 2.8 OASI (Oasis) 37.52 117.81 -8.5 + 1.6 11.5 + 2.7 -0.5 + 5.8 OVRO (OwensValley RadioObservatory) 37.23 118.29 -7.3 + 1.0 16.9 + 1.6 -5.8 + 3.4 ORVB (Oroville) 39.55 121.50 -6.6 + 1.1 21.9 + 1.7 -7.2 + 3.8 QUIN (Quincy) 39.97 120.94 -6.5 + 0.6 18.8 + 0.9 -4.0 + 2.2 SPRN (Springville) 36.18 118.73 -3.9 + 1.1 20.2 + 1.8 -3.1 + 3.7 SUTB (SutterButtes) 39.21 121.82 -5.9 + 1.2 22.2 + 2.0 -0.4 + 4.8 TIOG (TiogaPass) 37.93 119.25 -3.8 + 0.5 21.7 + 1.0 0.4 + 2.7 UCDI(UC Davis) b 38.54 121.75 -5.8 + 1.4 17.5 + 2.1 1.6 + 5.5 WGRD (WestguardPass) 37.27 118.15 -8.2 + 0.8 16.9 + 1.4 -2.8 + 3.5 WMTN (White Mountain) 37.57 118.24 -8,3 + 0,7 14.9 + 0.9 -0.9 + 2.9 a Verticaltime series has unexplained offset. b Verticalrate based on post-April, 1998 data because ofantenna change. the SierraNevada data. Thesestations provide time series in detailby Mao [1998]. We first definebest fitting ITRF- that are longerthan 3 years. 96 Euler vectorsfor eight Sierranstations (CEDA, CMBB, Datawere analyzed at theUniversity of Miami following KMED, ORVB, SPRN, SUTB, TIOG, UCD1) or some Dixon et al. [1997]. Briefly, we usedthe GIPSY software subset,and the 16 stationson stableNorth America [Dixon developedat the Jet PropulsionLaboratory [JPL] and et al., 1996; DeMets and Dixon, 1999]. Thesetwo Euler satellite ephemerisand clock files provided by JPL vectors are then used to define a relative Euler vector [Zumbergeet al., 1997]. Theseprocedures resulted in site describingrotation of the SierraNevada block with respect velocities defined in global reference frame ITRF-96 to North America. The 16 North American stations are [Sillard et al., 1998] (Table 1). Errors are estimated sufficientto ensurea robustregional reference frame (stable following Mao et al. [1999] and accountfor the influence North America), suchthat referenceframe "noise"is much of bothwhite and colorednoise (see section 2.1). A formal smallerthan the velocity error of any individual Sierran inversionprocedure [Ward, 1990] is used to derive site station,allowing us to def'methe velocitiesof the Sierran velocitiesrelative to stableNorth Americaand generatean sitesrelative to stableNorth America(Table 2) with high Euler vectordescribing motion of the SierraNevada block accuracy. Subsetsof data from the Sierran sites can be relative to stable North America. The method is described used to test block rigidity and to assessthe level of

Table 2. Site Velocities Relative to Stable North America

North, West, Rate, Azimuth, mm/yr mm/yr mm/yr deg clockwisefrom N

CEDA 9.7 + 1.0 9.2 + 1.3 13.4 + 1.1 317 + 5 CMBB 8.7 + 0.8 9.0 + 1.4 12.6 + 1.2 314 + 5 ELYA 0.6 + 0.9 3.3 + 1.2 3.4 + 1.2 280 + 14 GFLD 2.0 + 1.6 0.0 + 2.7 2.0 + 1.6 359 + 77 KMED 10.2 + 0.7 8.2 + 1.0 13.1 + 0.9 321 + 4 OASI 4.6 + 1.6 0.9 + 2.7 4.7 + 1.7 349 + 33 OVRO 5.9 + 1.1 6.4 + 1.7 8.7 + 1.5 313 + 9 ORVB 7.6 + 1.2 11.4 + 1.8 13.6 + 1.6 304 + 6 QUIN 7.1 + 0.6 8.3 + 0.8 11.0 + 0.9 313 + 4 SPRN 9.5 + 0.6 10.0 + 1.3 13.8 + 1.6 313 + 6 SUTB 8.3 + 1.3 11.8 + 2.0 14.4 + 1.8 305 + 6 TIOG 9.9 + 0.5 11.2 + 1.0 14.8 + 1.3 311 + 5 UCDI 8.4 + 1.5 7.2 + 2.1 11.1 + 1.8 321 + 10 WGRD 5.0 + 0.9 6.4 + 1.5 8.1+ 1.3 308+8 WMTN 4,9 + 0,8 4,4 + 1.0 6.5+0.9 319+8 6 DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION possible"edge" effects from elasticstrain accumulation. (campaign or continuous,with varioustime spans)to be Similarprocedures are used to definethe velocitiesof sites correctlyweighted in the inversionto deriveEuler vectors. within the Basin and Rangeprovince or on its deforming Equation(1) is strictly valid only for evenly sampled margins(GFLD, ELYA, OASI, OVRO, QUIN, WGRD position time series. Our campaignmeasurements yield andWMTN) relativeto stableNorth America. very unevensampling, typically severalconsecutive days every year or two (Figure 2). Numerical experiments 2.1. Uncertainties suggestthat approximatingan unevenly sampledtime Unlessspecifically stated, all uncertaintiesin the tables seriessuch as this by an equivalentevenly sampled time andtext represent one standard error, while all errorellipses seriesyields rate errorestimates that differ by less than 1 in figures representtwo-dimensional 95% confidence mm/yr from actualvalues (usually much less). We make regions(1.7 timesthe two-dimensional one standard error). the even samplingapproximation in this study. As an Here we describethe proceduresused to estimatethese example,for stationCEDA, with a total of 19 observation uncertainties. daysover 5 years(Figure 2), we setg equalto 3.8 samples GPS velocity errorsmay be estimatedassuming that per year. measurementnoise is uncorrelatedin time ("white"). If Applicationof (1) also requiresestimates for white, time-correlated("colored")noise is present, the true flicker, and randomwalk noise. Thesecan vary from velocityuncertainty will be underestimatedif pure white stationto station,as they dependin part on local station noiseis assumed[Johnson and Agnew, 1995]. Possible characteristics.Mao et al. [1999] list noise values for a sources of colored noise in GPS include monument motion numberof North Americanstations including QUIN and unrelated to the larger tectonic motions of interest CMBB. For the remaining stations,we must estimate [Langbeinand Johnson,1997], uncertaintyin the satellite noisevalues using other criteria. Mao et al.'s data show a orbitparameters, and atmospheric and local environmentalgood correlationbetween the weightedroot-mean-square effects[Mao et al., 1999]. Zhanget al. [1997] and Mao et (WRMS) scatter of an individual time series and al. [1999] demonstratethat GPS velocity errorsmay be correspondingmagnitudes of white and flicker noise. These correlations can be used to estimate white and underestimatedby factorsof 2-11 if pure white noise is assumed.Mao et al. presenta simplemodel for estimating colorednoise magnitudes for individual GPS time series, the GPS rate error(Or) for individualvelocity components providedthat sufficientsamples (observation days) are available to ensure that WRMS is a reliable indicator of (north,east and vertical)for coordinatetime seriesin the presenceof combinedwhite and colored noise: dataquality. GFLD, OASI, and OVRO haveanomalously low (optimistic)WRMS values,probably reflecting the Gr2 12aw2 aGf2 Grw 2 small number of observations(Figure 2); their noise ---•+gT3 gbr 2 +• gT (1) estimatesare insteadbased on the averageWRMS of all stationsin the regionwith equivalentantenna type and whereg is the numberof measurementsper year, T is the sufficientobservations to be representative,listed in the total time spanof observations(2-5 yearsin this study), a footnoteto Table 3. Table 3 also lists the WRMS by andb are empiricalconstants (a = 1.78 and b=0.22), (Swand componentfor stations analyzed in this report, and (5iare the white and"flicker" noise magnitudes in mm, and correspondingwhite and flicker noise estimatesbased on Orwis "random walk" noise in mm/x/yr.Flicker noise and thesecorrelations (relevant equations are given in footnotes randomwalk noise are different types of colored noise, to the table). Our estimatesof velocityuncertainty are distinguishedby their time-dependence(Mao et al. give a basedon thesevalues, equation (1), andthe assumptions completediscussion). Briefly, flicker ("pink") noise has stated above. spectralpower that is inverselyproportional to frequency, while randomwalk ("red")noisehas spectralpower that is inverselyproportional to frequencysquared. Monument 2.2. Accuracy noise has been characterizedas a random walk process [Langbeinand Johnson, 1997] and is likely to be more One way to assessthe accuracyof our derivedvelocities significantfor sites in unconsolidatedalluvium than for is to compareresults to independentdata. This also sites in bedrock. From (1) it can be seen that if random providesa meansto assessthe errorestimates: if they are walk noiseis small, velocity errordepends strongly on the reasonable,our expectationis that most velocitiesreported total time spanof observations(T) andweakly on sampling for the same location should overlap within errors, frequency(g). Mao et al. suggestthat white noise and dependingon the confidencelevel quoted. OVRO and flicker noise dominate the GPS noise spectrum for QUIN eachhave very long baselineinterferometry (VLBI) coordinatetime seriesof the type used here, i.e., random and/orvery long baselinearray (VLBA) data availablefor walk noise is relatively small. Thus, our 5 year time series comparison. Table 4 lists publishedvelocities for these for Sierran stations set into bedrock and occupied in sitesrelative to stableNorth America,from the compilation periodic campaigns(CEDA, KMED, SPRN and TIOG) of Hearn and Humphreys[1998], updatedto reflect the shoulddef'me accurate site velocities,even thoughthey are latest VLBI/VLBA data, solution GLB 1102 [Ma and sampled much less frequently than semi-permanent,Ryan, 1998], and our own GPS data. The VLBI results continuouslyrecording stations such as CMBB, ORVB, listedare not independent,as they all rely on essentially SUTB and UCD1. The importanceof our noise model is the same data for the early part of their respectivetime that it allows the velocities from these diverse data sets series, but the manner in which stable North America is DIXON ET AL.' PRESENT-DAY SIERRA NEVADA MOTION 7

Table 3. Uncertainties

WRMS White N0is½ Flicker Noise • North East Vertical Northa Eastb Verticalc Northd Easte Verticalf

CEDA 3.0 4.9 9.8 2.1 3.6 6.5 3.5 4.8 11.9 CMBBg 3.9 8.6 19.7 2.7 6.4 14.8 4.6 8.6 18.6 ELYA 2.5 4.3 8.5 1.8 3.1 5.4 3.0 4.1 1 I. 1 GFLDi• 3.1 5.6 11.0 2.2 4.1 7.5 3.6 5.5 12.7 KMED 2.1 3.9 9.7 1.6 2.8 6.4 2.5 3.7 1 1.2 OASIh 3.1 5.6 11.0 2.2 4.1 7.5 3.6 5.5 12.7 ORVB 3.0 5.8 10.9 2.1 4.3 7.4 3.5 5.7 12.7 OVROh 3.1 5.6 11.0 2.2 4.1 7.5 3.6 5.5 12.7 QU1N 3.8 6.5 14.2 2.6 4.8 10.2 4.5 6.4 14.9 SPRN 2.8 5.5 9.9 2.0 4.0 6.6 3.3 5.4 12.0 SUTB 2.9 5.6 12.2 2.0 4.1 8.5 3.4 5.5 13.5 TIOG 3.1 6.1 9.3 2.2 4.5 6.1 3.7 6.0 11.6 UCD1i 3.6 6.1 11.0 2.5 4.5 7.5 4.2 6.0 12.7 WGRD 2.9 5.9 14.1 2.0 4.3 10.1 3.4 5.8 14.8 WMTN 2.6 3.9 10.8 1.9 2.8 7.3 3.1 3.7 12.6 NoAmJ 3.8 4.9 12.8 2.6 4.6 8.6 4.4 6.3 13.5 All uncertaintiesgiven in millimeters. WRMS is weightedroot-mean-square. a Whitenoise (N) = 0.613[WRMS(N)]+ 0.259. b Whitenoise (E) = 0.767[WRMS(E)]- 0.182. c Whitenoise (V) = 0.843[WRMS(V)]- 1.772. d Flickernoise (N) = 1.i 39[WRMS(N)] + 0.117. e Flickernoise (E) = 1.041[WRMS(E)] - 0.342. f Flickernoise (V) = 0.668[WRMS(V)]+ 5.394. g Verticalcomponent time series at thissite has unexplained offset associated with equipment change and may notgive a reliableestimate of verticalvelocity averaged over the entire time interval. h Noisevalues for these stations are based on means of CEDA, KMED, ORVB, QU1N, SUTB, TIOG, and UCD 1. i Verticalnoise values at UCD 1 arebased on time series after April 1997 to avoidantenna change effects. J Meanvalue for North America [Mao et al., 1999].

definedcan differ significantly,so the comparisonsare still locationsseveral hundred meters apart, themselves differ by useful. The various VLBI results for a given site often 1 mm/yr in rateand 9 ø in azimuth,which is much larger differ by much more than two standard deviations, thanquoted errors (Table 4). This discrepancymay reflect implying that errors for some of these results may be underestimationof erroror real variationreflecting local underestimated. Our GPS estimate for QUIN should be complexitiesin the velocityfield, sincethe sitesare located well resolved, as it is based on a large number of data within a right step in the OwensValley-White Mountain spanning5 years(Figure 2). QUIN's GPS-basedhorizontal fault zone[Dixon et al., 1995]. The weightedmeans of rate estimate(11 mm/yr) lies in the middle of the rangeof recent independentresults in Table 4 probably best publishedVLBI values (9-13 mm/yr). The GPS rate, representthe velocities of OVRO and QUIN. Unless azimuth,and vertical estimatesfor QUIN all lie within one specifically noted, these mean values are used in the standarderror of the most recentVLBI result, probablythe remainderof the paper. To avoid overweightingin most accurate estimate available at this site. Our GPS rate calculationof the mean, VLBI/VLBA errorswere increased estimatefor OVRO is based on far fewer data (Figure 2) by a factorof 2. At OVRO thisprocedure gives essentially and thus is less reliable, which is reflectedin its larger the VLBA resultowing to its very high precision,but at velocityerror (Tables 1, 2 and4). OVRO's GPS-basedrate QUIN the weightedmean is intermediatebetween the GPS estimate(9 mm/yr) lies at the low end of the range of result and the most recent VLBI result. availablevalues (9-12 mm/yr). The GPS horizontalrate estimate lies within two standard errors of the recent VLBA 3. Results result,probably the mostaccurate estimate available at this site, while the GPS azimuth and vertical estimates for The GPS velocity data are listed in Tables 1 and 2, and OVRO both lie within one standard error of the VLBA displayedin Figures2 and 3. Beforeusing thesedata to result. The VLBI and VLBA results at OVRO, from assessthe rigidityof the SierraNevada block or describeits 8 DIXON ET AL.' PRESENT-DAY SIERRA NEVADA MOTION

Table 4. Comparisonof OVRO andQUIN faults of interesthere, several times the 10-15 km thickness Velocities Relative to Stable North America of the elastic,brittle uppercrust. In this case,"far-field" sites (i.e., those fartherthan 50-100 km from the fault, Rate, Azimuth, Up, which includes most of our Sierran sites), should mm/yr degW of N mm/yr experiencesmall elastic strain effects, at or below the noise level of the observations(-1 mm/yr). If a simple elastic OVRO half-spacemodel appliesto the Sierras,our expectationis VLBI a 11 + 1 332 + 3 -- that most of the measured GPS velocities on the stable VLBIb 12.0+ 0.3 316+ 1 -2.7+ 0.9 block interior should be broadly representativeof longer- VLBI c 10.1 + 0.5 322 + 2 -- term motions, since our Sierran sites are located far from VLBId 11.7+ 0.2 323+ 1 -2.4+ 0.8 active faults and our measurementsoccurred long after VLBAd 10.6+ 0.1 314+ 1 -5.9+ 0.3 major earthquakeson boundingfaults. However, for a more realistic layeredrheology, with an GPSe 8.7 + 1.5 313 + 9 -5.8 + 3.4 elastic layer overlying viscoelasticmaterial, present-day Meanf 10.7+ 0.7 316+ 5 -5.5+ 1.4 elasticstrain effects due to lockedfaults may be influenced by past earthquakes,and theseeffects can extendfar from QUIN the fault and persist long after the last earthquake, VLBI a 11 + 1 310 + 5 -- dependingon earthquakehistory, fault depth,and crustand VLBIb 12.8+ 0.5 303+ 2 -4.5+ 3.6 uppermantle rheology [Savage and Lisowski, 1998]. This VLBI c 8.9 + 0.5 294 + 2 -- is especiallyimportant for kinematicstudies of continental VLBId 12.3+ 0.5 309+ 2 0.4+ 3.4 blockswith spacegeodesy, since the smallersize of these GPSe 11.0 + 0.9 313 + 4 -4.0 + 2.2 blocks comparedto larger plates means that it may be Meanf 11.6+ 0.9 311+ 3 -3.6+ 1.8 difficult to find sites whose velocities are uncontaminated Argusand Gordon [ 1991 ]. by edgeeffects. The generallysmaller size of continental blockscompared to platesalso makesit difficult to obtain Ma et al. [ 1994]. geographic"spread" in spacegeodetic site location,which Dixonet al. [1995]. is importantfor accurateEuler vectorestimation, especially Ma and Ryan [ 1998]. if the block boundariesneed to be avoided. Together,these Thisstudy. effects may limit accurate Euler vector estimates for Weighted mean and repeatability of recent continentalblocks based on spacegeodesy, unless accurate independent data: GPS (this study) and site velocities can be defined for the interior of the block, VLBI/VLBA data of Ma and Ryan [1998], appropriatemodels are available to "calibrate"the edge with VLBI/VLBA error scaledup by 2. effects,and statisticaltests are available to evaluateresults. The presentstudy takes a first step in addressingthese issues. present-daymotion, we first addressthe relationbetween One way to assessthe influenceof strain accumulation measuredGPS velocitiesand longer-termmotions. andearthquake history on the measuredsite velocitiesis to comparethe Euler vector based on velocities at all the Sierran sites to an Euler vector based on a subset of sites 3.1. Short-Term Versus Long-Term Velocities on the block interior likely to be less affectedby strain accumulationand earthquakes. If the fit of the velocity In the discussionbelow, we assumethat the short-term, datato a rigid block model is significantlyimproved by interseismicsite velocities measured by GPS canbe related elimination of suspectsites, this would suggestthat the to longer-termgeologic motions (e.g., fault slip ratesand effectsare important,and the resultingEuler vectorbased block motions) that averageover many seismic cycles, on interior sites would presumablybe a more accurate eitherdirectly or throughsimple mechanical models. There representationof long-termSierra Nevada block motion. are at least two effects to consider. First, the crust and We use the reducedZ 2 statisticto assessthe fit of the upper mantle have a postseismicresponse to stresses releasedin large earthquakes[Pollitz and Sacks, 1992; velocitydata to a givenrigid block model: Pollitz, 1997; Pollitz and Dixon, 1998]. Time-dependent postseismiceffects decrease with time aftera major event, Z2=• (Oi-Ci)2 /2N 3 (2) and may not be significantafter many decades. Large i=l O'• -- earthquakeson the major faultsbounding the SierraNevada whereO• is a velocityobservation (north or east),Ci is the occurredin 1857and 1906on the SanAndreas fault, and in • calculatedvelocity at the samesite basedon a givenEuler 1872on the OwensValley fault. The othermajor process vector,c•i is the velocity error,N is the numberof sites is elastic strain accumulationon and near locked faults, ' usedin the inversion(5,7, or 8), and 2N-3 is the numberof which may reducethe relativevelocity of sites acrossthe degreesof freedom(number of sites,with two datapoints fault comparedto the long-termgeologic slip rate. For each, minus number of adjustable parameters). The simpleelastic half-space models of strainaccumulation minimum Z 2 indicatesthe bestfit model. Valuesof [Savageand Burford, 1973; Okada, 1992], the critical reducedZ2-1.0 indicate a goodfit of datato modeland distanceis of order50-100 km for the verticalstrike-slip suggestthat error estimatesare reasonable. Reduced DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION 9

122 ø W 120 ø W 118 ø W

40 ø

39 ø

'• 38ø

OVRO 37 ø

36 ø M EDA•

237 ø 238 ø 239 ø 240 ø 241 ø 242 ø Longitude (E) Figure 3. GPS site velocities (arrows)with respectto stable North America (Table 2) and their two- dimensional95% confidenceellipses. Faultsand fault namesare sameas Figure 1, except dsf, Deep Springs Fault; flf, Fish Lake Valley fault zone. Star showsthe location of shallow pressuresource at Long Valley Caldera [Langbeinet al., 1995b]. Arrows without ellipsesshow predicted velocities for sites on stable Sierra Nevadablock usingGPS-based Euler vectorin Table 6 (5 station solution). Note that velocities of all sites on the stable block interior fit the predictedvelocity within uncertainty ellipse, while sites on the block margin(QUIN, OVRO, WMTN andOASI) do not. WGRD omittedfor clarity,ELYA and GFLD are off this map (seeFigure 1 for locationof thesesites).

Z2l.0suggest either that errors are underestimated orthat overestimated. a given model poorly fits the data. Theserules of thumb For comparison,we also list an Euler vector basedon assumethat errorsare normally distributedand that sample eliminationof the threesites most likely to be affectedby size is sufficientlylarge to be statisticallyrepresentative; elasticstrain accumulation: TIOG, becauseit is affectedby neithercriteria is particularlywell satisfiedhere, so caution deformationat Long Valley Calderaand is alsoclose to the is warranted. Table 6 lists an Euler vector based on the SierraNevada range front; KMED, becauseit lies only 18 velocities of all 8 sites located on the Sierra Nevada block, km west of the Owens Valley fault zone; and UCD1, withreduced Z2 = 0.44,indicating that the data fit therigid becauseit is the closestof the eight sites to the San 10 DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION

Andreasfault, andonly -45 km eastof the GreenValley recurrenceinterval, and expressionsfor ak are given in the fault,part of the activeCalaveras - Concord- Hunting appendixto Savageand Lisowski. The time constantXo is Creek- BartlettSprings fault system(Figures 1 and 3). relatedto the earthquakerecurrence interval and relaxation An Euler vector describingSierra Nevada block motion time for the viscoelastichalf-space, Xo=[tT/2•l, where it is basedon the remainingfive sites on the stable Sierra therigidity of thehalf-space (set here to 3x10•øPa),and rl Nevadablock interior has reduced Z2=0.23, a factorof 2 is the viscosity(set here to 2x1019Pa-sec). These values improvementcompared to the Eulervector based on all 8 for rigidityand viscosity are equivalentto a relaxationtime sitesincluding ones near the blockmargins. This suggests (rl/g) of-20 years,intermediate among the rangeof values that elastic strain effects may be important, may be foundby otherauthors [e.g., Thatcher,1983; Li and Rice, detectablein our data, and that the Euler vector basedon 1987]. Each fault segmentis assumedto be long and the subsetof GPS sites closerto the block interior gives a vertical,and local strain effects at theend of fault segments more reliableestimate of long-termblock motion. It also are ignored. Each of the Sierran sites is assumedto be suggeststhat errorsmay be overestimated. Reducing influencedby both the San Andreasfault systemto the averagevelocity errorby a factor of 2 for the five site west and the eastern California shear zone to the east solutiongives reduced •Z2-1.0. This may reflect the fact (Figure1). The directionof velocity incrementfor these thatthe regressionsused to estimatethe errors(Table 3) are two fault systemsis assumedto be parallelto plate basedon data from acrossNorth America, encompassinga motion. Strain effectsfrom the San Andreassystem are widerange of climatezones, soil or bedrocklocations, and suchthat measuredSierran site velocitiesare mainly faster monumentstyles. In contrast,most of the Sierransites are comparedto the long-term average,while strain effects locatedat high altitudein a dry environment(minimizing from the eastern California shear zone are such that troposphericnoise) and set into glaciatedbedrock measuredSierran site velocitiesare mainly slowerthan the (minimizingground noise). Giventhe smallnumber of long-termaverage. In andnorth of the San FranciscoBay data(10 versus16), we cannotpreclude the possibilitythat areawe explicitly include the effectsof the San Andreas someof the improvementin fit betweenthe 5 site and 8 fault (sensustricto), the Calaverasfault, the Hayward- site solution is due to chance. Note that the 5 station and RogersCreek- Maacamafault system,and the Concord- 8 station Euler vectors are indistinguishablewithin Green Valley- Hunting Creek- Bartlett Springs fault uncertainties,suggesting that the magnitudeof strain system.Effects from the 1989 Loma Prietaearthquake are effectsis near the error level of our data. This partly ignored,as are the effectsof creepingsections of faults. reflectsthe relativelylong distancebetween known active For the three southernmostsites only (CEDA, KMED, and faultsand most of our sites(minimizing strain effects), and SPRN) we also includethe effect of the Garlock fault. may alsoreflect the factthat the straincorrections for the We approximatethe effectsof the easternCalifornia SanAndreas fault systemand easternCalifornia shear zone shearzone to the eastby consideringonly the effectof the areopposite in signand thus partly cancel. westernmostfault (i.e., the fault closestto our sites) in the strain accumulationmodel, and assumingthat this fault 3.2. Correcting for Elastic Strain Accumulation With carriesapproximately half of the total -11 mm/yr of slip a Coupling Model that needsto be accommodated.For example,south of 37ø N latitude,we consideronly the OwensValley fault zone, Given the apparentimprovement in fit when sitesmost whoseslip rate is taken to be 6 mm/yr. North of 38ø N likely to be affectedby strainaccumulation are eliminated, latitude,the main bounding fault or faults for the Sierra we also investigatedstrain effectsin a more quantitative Nevada block have not been determined. We assume that way, as follows. TIOG is omitted from the calculations,as one of the main boundary faults follows the band of it may be influencedby deformationat Long Valley seismicitythat, from south to north, begins near Mono Calderawhich is time-varyingthrough our measurement Lake, California, continues northwest near the California- period. For the remaining7 sites, we use the viscoelastic Nevada border to Lake Tahoe, and then continues with a couplingmodel of Savage and Lisowski [1998], with an small right step (i.e., Lake Tahoe may be a pull-apart elasticlayer overlyinga Maxwell viscoelastichalf-space. basin) northwestalong the Mohawk Valley fault zone The elasticlayer correspondsto the upper crust,and the [Goter et al., 1994] (Figures 1 and 3). A single viscoelastichalf-space corresponds to the more ductile throughgoingfault with clear surfaceexpression is not lower crust. The boundarybetween them, at depth H, developed in this region, but Thatcher et al. [1999] correspondsto the lower limit of brittle faultingand the observea strong gradientin GPS site velocitiesnear here. maximumdepth of crustalearthquakes, typically 10-15 A changein northwestvelocity of-6-8 mm/yr occurs km. For thismodel the surfacevelocity v is betweenthe centerof the SierraNevada block, moving at -14 mm/yr northwestrelative to North America (this v= Z(bn / zr)[arctan{x/((2n - 1)H)}-arctan{x/((2n +1)H)}] (3a) study) and the region immediatelywest of the central n=l Nevadaseismic belt, moving at-6-8 mm/yr northwest relativeto North America [Thatcheret al., 1999]. In bn+1= b•(*:O n/ n!) Y•n!/[k!(n - k)!]ak+ •(t / T)n-k (3b) additionto the MohawkValley fault zone,the HoneyLake k=0 fault zone (Figure 1) may accommodatepart of this where x is the horizontal distancefrom the fault, t is the motion. Wills and Borchardt[1993] estimatea Holocene time since the last earthquake,T is the earthquake slip rateof 1.1-2.6 mm/yr for the Honey Lake fault zone DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION 11

Table 5. Fault ParametersUsed in ViscoelasticCoupling Model

CEDA KMED SPRN CMBB UCD 1 SUTB ORVB

Fault SAF SAF SAF SAF SAF SAF SAF Distance,km 119 171 142 167 109 146 191 SlipRate, mm/yr 34a'b 34a'b 34a'b 17a 17c 17c 17c LastRupture, Year 1857 1857 1857 1906 1906 1906 1906 Recurrence,years 206b 206b 206b 400 210 210 210 Depth,km 12 12 12 14 12 12 12

Fault - - - CLV RGC MAC MAC Distance,km - - - 143 7 5 108 15 0 SlipRate, mm/yr - - - 15a 14c 14c 14c LastRupture, Year - - - 1979 ? ? ? Recurrence,years - - - 3 3 222 22 0 2 20 Depth,km - - - 5 10 12 12

Fault GAR GAR GAR - GVL BSP BSP Distance,km 7 2 7 6 108 - 4 5 6 2 106 SlipRate, mm/yr 6d 6d 6d - 8c 8c 8c LastRupture, Year ? ? ? - ? ? ? Recurrence,years 1000 1000 1000 - 176 194 218 Depth,km 12 12 12 - 12 12 15

Fault OWV OWV OWV Unnamed MHK MHK MHK Distance, km 63 17 6 7 100 166 145 7 0 SlipRate, mm/yr 6e 6e 6e 5f 5f 5f 5f LastRupture, Year g (-) (-) (-) ? ? ? ? Recurrence,years 1000h 1000h 1000h ? ? ? ? Depth km 13 13 13 12 15 15 15 Fault abbreviations:SAF, San Andreas;CLV, Calaveras;RGC, RogersCreek; MAC, Maacama;GAR, Garlock; GVL, GreenValley; BSP, Bartlett Springs-Hunting Creek; OWV, OwensValley; MHK, Mohawk. a CaliforniaDivision of Minesand Geology [1996]. b Working Group on California Earthquake Probabilities [1995]. c Freyrnuelleretal. [1999]. d McGill and Sieh [1993]. e 50%of total 11 mm/yr of integrateddeformation across eastern California shear zone, augmented by 0.5 mm/yr to accountfor far-field effects of Death Valley-FurnaceCreek fault zone. f Basedon assumed relation to Owens Valley fault zone slip rate, kinematic consistency, anddata from this study and Thatcher et al. [1999]. g TheOwens Valley fault last ruptured in 1872. However,CEDA, KMED and SPRN are near or beyondthe southernlimit of surfacerupture. We have assumedthat the last rupture occurredsuch that we are midway through the earthquakecycle. h Basedon time to accumulate1872-1ike slip (-6 m) at a rate of 6 mm/yr.

(preferredvalue 2.0 mm/yr). We thereforeassume a total of model are taken from the compilationby the California 7 mm/yr,with 5 mm/yr accommodatedon the seismically' Divisionof Mines and Geology[1996], and the time and activeMohawk Valley fault zone. New GPS data from locationof surfaceslip from the last major earthquakeare QUIN areconsistent with this modelof slip distribution taken(rom Jennings and Saucedo [1994]. In caseswhere (Section4.3). fault depth is unknownwe arbitrarilyset it to 12 km. If Unless noted, all slip rates, earthquakerecurrence earthquakerecurrence and time of last earthquakeare intervals,and fault depthsfor the viscoelasticcoupling unknown,we arbitrarilyset thesevalues to 500 yearsand 12 DIXON ET AL ß PRESENT-DAY SIERRA NEVADA MOTION

Table 6. Best Fitting Euler Vectors

Latitude, Longitude,to, ErrorEllipse a (5•o, Z2b deg N degE deg/my (SmaxOmin •max deg/my

NorthAmerica (ITRF-96) c - 1.2 279.8 0.193 2.7 0.9 -1 0.005 0.80 Sierra Nevada (ITRF-96) All sites,uncorrected d 13.3 245.6 0.44 6.0 0.4 -lO 0.07 0.44 7 sites,velocity correction e 15.8 244.8 0.48 7.3 0.5 -9 O.11 0.95 Northernblock f 17.3 244.8 0.52 12.3 0.5 -12 0.20 0.65

Southernblock g 13.4 245.5 0.45 8.4 0.4 -9 O.11 0.34

Interiorsites, uncorrected h 10.9 246.1 0.41 6.8 0.4 -11 0.07 0.23

ß SierraNevada-North America • Argusand Gordon [1991 ]J 32 232 0.61 6 1 51 0.15 Hearnand Humphreys [1998] k - 13.4 205.6 0.13 All sites,uncorrected d 19.8 224.0 0.31 8.1 1.0 -37 0.07 7 sites,velocity correction e 22.5 225.1 0.35 9.2 1.2 -39 0.11 Northernblock f 24.0 226.7 0.39 13.8 1.1 -34 0.20

Southernblock g 19.9 224.2 0.32 11.4 1.0 -38 0.11

Interiorsites. uncorrected h 17.0 222.7 0.28 9.9 1.1 -34 0.07 a •maxis orientationof long axis, degrees clockwise from north. Axes are two-dimensional one standard error; for 95% confidence,multiply by 1.7. b Z2 is squaredsum of residuals(observed data minus calculated model) divided by standard deviation squared(observation error from equation 1 and noise values in Table 3), normalized by 2N-3, where N is numberof sitesused in inversion(5,7 or 8) (equation2, text). c Basedon velocities of 16sites described in DeMets and Dixon [ 1999],updated with additional data. d Based on horizontal velocities of 8 sites, CEDA, CMBB, KMED, ORVB, SPRN, SUTB, TIOG and UCD1, uncorrected for strain accumulation. e Basedon horizontalvelocities of 7 sites,CEDA, CMBB, KMED, ORVB, SPRN, SUTB and UCD1 (i.e., TIOG omitted) with velocities correctedfor strain accumulationas describedin text and Table 5. f Basedon horizontal velocities of CMBB, ORVB, SUTB, TIOG and UCD 1. g Basedon horizontal velocities of CEDA,CMBB, KMED, SPRNand TIOG. h Preferred(minimum Z2) model, based on horizontal velocities of5 siteson stable interior of Sierra Nevada block, CEDA, CMBB, ORVB, SPRN and SUTB, uncorrectedfor strain accumulation. i SierraNevada rotates relative to North America. J Basedon VLBI velocities of OVRO and QU1N relative to stableNorth America, and deweighted VLBI velocity for Mammoth Lakes. k Basedon a regional kinematicmodel.

250 yearsago, respectively. If recurrenceinterval is known The velocity incrementsestimated from the coupling but the time of last earthquakeis unknown,we arbitrarily model, to be added to or subtractedfrom the observedsite setthe latter to a time beforepresent equal to one-halfthe velocities,are generallyless than 2 mm/yr, nearthe level of recurrenceinterval, i.e., the fault is assumedto be in the themeasurement error. The largestcorrection is at CMBB, middle of its earthquakecycle. The most recentsurface with a ratecorrection of 2 mm/yr. The Euler vectorbased rupturefor the centralGarlock fault occurred sometime after on "corrected"site velocitiesis equivalentwithin errorsto 1490 A.D. [McGill, 1992]. We arbitrarilyassume it to be the Euler vectors based on uncorrected velocities, but midwaybetween that date and the present time (i.e., 1745). misfitsthe data (Z 2= 0.95)significantly more than either 2 Modelparameter values and data sources are summarizedin of the uncorrectedEuler vectors(Z = 0.23 and 0.44; Table Table 5. 6). This may reflect our simplistic applicationof the DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION 13

couplingmodel (e.g., ignoring the effect of finite fault Both of these sites are located in alluvial valleys. length)or uncertaintyabout key parametersin the model, Subsidence is also observed at both the VLBI and VLBA such as earthquakerecurrence interval or time of last sites at OVRO, but not at the VLBI site at QUIN (Table rupture, both of which can have a big effect on the 4), so only OVRO's subsidencemay be real. Subsidence magnitudeand even the sign of the correction. In the at OVRO may be relatedto rangefront normal faulting near remainingdiscussion, we usethe Euler vectorbased on the the stations,or it may be due to alluvial compaction.All five interior sites, with velocities uncorrectedfor strain three OVRO stations are located on unconsolidated accumulation("Solution h, Preferred"in Table 6), as it is alluviumin OwensValley, where groundwater levels have thebest fit (minimum•2) solution.This solution predicts declined in the past due to diversionof easternSierra ran- northwest motion of the Sierra Nevada block at 13-14 off into the Los Angelesaqueduct. The resultingloss of mm/yr, closeto an estimateof presentday motion based pore pressurecan lead to compactionof unconsolidated on regional kinematic considerations(N50øW+5 ø at alluviumin the region abovethe decliningwater table and 12.7+1.5 mm/yr)[Hearn and Humphreys,1998]) and close consequentsurface subsidence. Local compactionmight to the geologicalaverage motion since 8-10 Ma (NW-NNW compromisethe utility of this site for regional tectonic at 15 mm/yr)[Wernicke and Snow, 1998]. Thus Sierra studies, as it may also affect the horizontal velocity Nevadamotion appearsto have been remarkablysteady components. However, we see no evidenceof this in our over the last 8-10 Ma. data: OVRO's velocity fits regionalfault models within uncertainty,as shownin a subsequentsection. 3.3.2. Monument noise. If the Sierra Nevada block were 3.3. Vertical Motion, Data Quality, Monument Noise, and perfectlyrigid, the residuals(difference between observed Local Effects velocity and velocity calculatedfrom our Euler vector; Table 7) wouldreflect the root-sum-square(rss) of all error 3.3.1. Uplift, subsidence,and data quality. With a time sources,including monument noise. North residualsare spanof observationstotaling 5 yearsat many of our sites, smallerthan east residuals,suggesting that the residuals the verticalcomponent of velocityis sufficientlyprecise for likely reflecterror sources unique to GPS (east errorsare some applications. To minimize vertical errors at the generallylarger than north errors for GPS results where campaignsites, we took four steps. First, we employed carrierphase biases are not fixed, as in this analysis),and the same antennadesign for most observations,a Dom- do not reflect monumentnoise, which shouldbe random in Margolin antennawith choke ring back plane, a design direction. The effectof monumentnoise on the velocity -.:withgood multipath rejection. (CMBB had a different estimatesis thus likely of the orderof or smallerthan the antennadesign through mid-1997 and underwentan largestnorth residual(0.6 mm/yr; Table 7). While the antennachange in August 1997. This may be relatedto the permanentsites (CMBB, ORVB, SUTB and UCD1) have offsetin the verticaltime seriesabout this time (Figure 2), relatively large monuments,our campaignsites on the and the vertical velocity from this site is omitted in the SierraNevada block or near its easternmargin consistof following discussion). Second, for most campaign small,inexpensive stainless steel pins or brassplaques set observationswe usedfixed height "spike" antennamounts, in bedrock,from the surfaceto a depth of-•15 cm. The designedto eliminate the uncertainty associatedwith high quality of the data from thesemarks (e.g., Tables 1, measurementof variabletripod height that often plagues 6, and 7; Figures2 and 3) suggeststhat when unweathered thistype of measurement(a pictureof our spikemount set- bedrockis available,such inexpensive marks are adequate up is visible on the World Wide Web at for most GPS experiments. Deeply drilled monuments www.geodesy.miami.edu).Third, we were able to locate [e.g., Langbein et al., 1995a] which reducethe effect of many of the sites on barehill tops far from obstructions, monument noise in unconsolidatedalluvium, are thus not reducing multipath and giving good sky visibility. requiredin the glaciatedSierran bedrock environment. Finally, all of our campaignobservations occurred in the 3.3.3. Effectof Long Valley. TIOG's velocity is slightly fall (September-October),minimizing the influence of faster(by 1 mm/yr) than the predictedvelocity at this annual or seasonaleffects on the time series. In addition, location. This may reflect the influence of local many of the sites, and all four Sierran campaign sites, are deformationfrom the resurgentdome at Long Valley locatedat relativelyhigh altitude,reducing the influenceof Caldera, located-40 km to the southeast(Figure 3). a variablelower troposphereon the heightestimates. These During the period of our measurements,the dome was sites thereforehave high-qualitydata, especiallyin the inflatingdue to a pressuresource at 5-7 km depth,possibly verticalcomponent. For example,all 4 Sierrancampaign a magmachamber, with the centerof thedome uplifting by siteshave vertical WRMS values less than 10 mm, which a total of-15 cm duringthis 5 year period [Langbeinet is lower than the correspondingvalues at most other sites al., 1995b; Dixon et al., 1997; unpublishedUniversity of (Table 3). Miami GeodesyLaboratory data, 1999.]. Simple elastic All 4 Sierrancampaign sites, includingthe two nearthe half-spacemodels suggest that strainpropagation from this rangefront (KMED and TIOG) havevertical velocities that centerof inflationwill add a northwestvelocity component are zero within one standard error. Thus there is no at TIOG relative to stable Sierra Nevada. Marshall et al. evidencefor rapiduplift alongthe easternSierra range front [1997] show detectablevelocity effectsnear this location fault, or anywhereelse in the block, within uncertainties. for a periodwhen Long Valley Calderawas less active. The GPS data at OVRO and QUIN suggestmarginally However,the magnitudeof this local volcaniceffect may significantsubsidence, 6+3 and 4+2 mm/yr, respectively. be partiallyoffset by elasticstrain accumulation associated 14 DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION

Table 7. Misfits for Best Fit Sierra Nevada Euler Vector Relative to ITRF-96

North Velocity, mm/yr EastVelocity, mm/yr Observed Predicteda Residual Observed Predicteda Residual

CEDA -3.60 -3.74 0.14 -19.3 -19.11 -0.19

CMBB -5.10 -5.15 0.05 -19.40 -20.66 1.26

ORVB -6.60 -6.02 -0.58 -21.90 -21.66 -0.24

SPRN -3.90 -3.85 -0.05 -20.20 -19.42 -0.78

SUTB -5.90 -6.27 0.37 -22.20 -21.40 -0.80 b Mean .... 0.24 .... 0.65

RMS .... 0.32 .... 0.77

a Predictionis from best fit Eulervector in Table6 (solutionh, 5 stations,uncorrected). b Mean of absolute value.

with right-lateral strike-slip faulting along the eastern are reachedif we use the larger, 8 stationdata set. Thus, California shear zone, which would tend to slow TIOG's we conclude that the GPS velocities for all sites on the velocityrelative to the SierraNevada block interior. TIOG's SierraNevada block are consistentwith the rigid block velocity data are includedin the Euler vector determination model within observationalerror. In other words, the basedon all of the "uncorrected"data (solution"d", Table kinematicsof the SierraNevada block canbe adequately 6), but omitted in the other two Euler vector estimatesfor describedby rotation of a rigid block on a sphere, the wholeblock (solutions "e" and "h", Table6) including analogousto the motionof largerrigid plates. the best fit vector. The abovearguments also suggestthat any velocity anomaliesdue to non-rigid block behaviorat our southern sites(SPRN, KMED and CEDA) associatedwith motion 4. Discussion onthe Kern Canyonfault, or long-termpostseismic effects associatedwith the 1952 earthquakeon the White Wolf 4.1. Rigidity of the Sierra Nevada Block fault,are small and below our data uncertainty. We cannot precludethe possibilitythat sucheffects are presentand Velocity data from the five GPS sites on the stable coincidentallyconspire to producea rigid blockmodel that interiorof the SierraNevada block are very well fit by a is biased,but this seemsunlikely. If suchwere the case, singleEuler vector, to much better than one standarderror we might expect that the northern Sierra Nevada block (Figure3, Tables1, 6 and7). The root-mean-square(rms) would behavedifferently from the southernSierra Nevada valuesof the residualsare quite small, 0.3 mm/yr (north) block;such is notthe case (see next paragraph). Bowden et and 0.8 mm/yr (east), considerablysmaller than the at. [1997] note that postseismiceffects associated with the independentlyestimated GPS velocityerrors, with rms 1952 Kern County earthquakewere high in the decade values1.0 mm/yr (north) and 1.6 mm/yr(east) ( Table1). followingthe earthquakeand then dropped significantly in The maximumrate residualis 1.3 mm/yr, andthe mean the following decades,consistent with our observations. rateresidual is 0.8 mm/yr,compared to themean rate error, We split the stationsinto two groups,defining separate 1.5 mm/yr. Theseresiduals are smallerthan the residuals Euler vectorsfor the northernblock (CMBB, ORVB, for similardata and models for major plates [e.g., Dixon et SUTB, TIOG and UCD1) and southernblock (CEDA, at., 1996, Dixon and Mao, 1997; DeMetsand Dixon, CMBB, KMED, SPRN and TIOG)to investigatepossible 1999;Norabuena et at., 1999]. The residualsreflect either differential motion. The data are insufficient to der'me themagnitude of non-rigidblock behavior or, morelikely, theseblocks with completelyindependent data, so CMBB the effect of observationalerrors. Thus we cannot conclude and TIOG near the centerof the block (Figure 3) are thatthe Sierra Nevada block is asrigid or morerigid than commonto eachgroup. The 95% confidenceregions for majorplates; it couldbe lessrigid, but the level of non- the two Euler poles have significantoverlap, and the rigidity in each case is well below the current GPS rotationrates agree within one standarderror (Table 6), detectionlimit. The relativemagnitudes of the residuals suggestingthat the entireregion behaves as a singlerigid and independentlyestimated velocity errors suggest that block within observational error. our errorestimates may be conservative,consistent with the The GPS residuals for the best fit Sierra Nevada block inferencefrom the Z2 tests. If the residualswere Euler vector (Table 7) are similar to or smaller than the significantlylarger than the error estimates,this would velocityresiduals for the old, cold continentallithosphere imply either that we had underestimatedthe errors or that comprisingstable North America [DeMets and Dixon, the SierraNevada block is not rigid. Similarconclusions 1999]. Seismicdata [Wernickeet al., 1996] as well as DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION 15 geothermometryfrom xenoliths in Cenozoic volcanics Nevada motion providing the kinematic boundary [Duceaand Saleeby,1996] have suggestedthe presenceof condition,is one way to quantifysuch missing motion, hot asthenospherenear the baseof thinned easternSierran and more generallyprovides a test of kinematicmodels. crust. Apparentlythe rigidity of the Sierra Nevada block We illustratethis here with severalexamples. hasnot yet beenaffected by the additionalheat by amounts A simple transectapproximately perpendicular to a greater than our observationalerror. Perhaps crustal small circle describingPacific-North America motion thinningand basal heating on the easternmargin are recent passesfrom the centerof the stableSierra Nevada block, phenomena,such that therehas not yet beensufficient time acrossnorthern Owens Valley near OVRO, south of the for the strong, brittle upper crust to warm and weaken centralNevada seismic belt, through Ely Nevada, and significantly. acrossthe Wasatchfault zonenear Salt Lake City to the ColoradoPlateau (Figure 1). A more northerntransect 4.2. Block Rotation and "Missing" Motion throughcentral Nevada encountersa differentstyle of deformationpartitioning, in particularthe centralNevada Rotation of continental blocks about vertical axes within seismicbelt [Pezzopaneand Weldon,1993; Bennettet al., or nearthe blocks is an importantcomponent of continental 1998; Thatcher et al., 1999] (Figure 1) but the same deformation,as documentedby numerouspaleomagnetic kinematic boundary condition (Sierra Nevada motion) should apply. The Wasatchfault zone and the faults studies. Argus and Gordon [1991] suggestedthat the SierraNevada block undergoescounterclockwise rotation comprisingthe easternCalifornia shear zone (e.g., Owens about a proximal vertical axis, located-•10 ø from the Valley andFish Lake Valley fault zones;Figures 1 and 3) southwestmargin of the block. The new GPS data do not are the only two major activefault systemscrossed in the supportrotation of the block abouta proximalaxis, rather, more southern transect. Deformation accommodating northwestmotion of the SierraNevada is partitionedon a the block undergoesslow counterclockwiserotation, with regionalscale between these two zones, respectivelythe the pole of rotationrelatively far from the block boundary easternand westernboundaries of the Basin and Range (Table 6). Our new rotation rate (0.28ø/Myr) is province, with east-west extension on the north-south intermediatebetween the rotationrates proposed by Argus striking Wasatchfault zone, and dominantly right-lateral and Gordon (0.61ø/Myr) and Hearn and Humphreys strike-slipon the northweststriking eastern California shear [1998] (0.13ø/Myr). Most of the Sierran site velocity zone [Dokka and Travis, 1990; Savageet al., 1990; Dixon azimuthslie close to the azimuthof present-dayPacific- et al., 1995; Hearn and Humphreys,1998]. Dixon et al. North America motion [DeMets and Dixon, 1999] [1995] comparedregional VLBI and SLR (satellite laser calculatedat the respectivesite locations. The northern stationson the block interior (ORVB and SUTB) have ranging)data to local deformationdata and suggestedthat velocity azimuths---10 ø more westerlyrelative to southern Basin and Range deformationwas largely restrictedat interiorstations (CEDA and SPRN), perhapsreflecting the presentto thesetwo marginalzones. Subsequentgeodetic kinematicconsequences of fastereast-west extension in the studieshave both refutedthis simple picture [Bennettet al., 1998] and confirmedit [Thatcheret al., 1999]. Part of northernBasin and Rangecompared to the southernBasin and Range [e.g., Shen-Tu et al., 1998; Bennett et al., the problem is that the nature of partitioning is almost certainly not constantthroughout the province. Also, 1999]. This resultsin a small componentof convergence partitioningof extensional(-east-west) motion and strike- normal to the San Andreasfault, dependingon latitude. slip motion associatedwith the easternCalifornia shear We can calculatethis using the new Sierra Nevada Euler vector, the new estimate of Pacific-North America motion zone needto be consideredseparately (extension could be [DeMetsand Dixon, 1999] and known motion on the San diffuse,while strike-slipmotion could be concentratedin Andreasfault (e.g., Table 5). At 36øN on the San Andreas the shearzone; e.g., Bennettet al., [1998]). Finally, it has beendifficult to accuratelyaccount for the effectsof time- fault,we calculate4+2 mm/yr of fault-normalconvergence, correlatederror inherentin any spacegeodetic technique, as well as an additional3+2 mm/yr of right lateral slip, some of which is likely accommodatedwest of the San leadingto conflictinginterpretations of the samedata [e.g., Savage, 1998]. Since even small amounts of deformation Andreasfault basedon the distributionof seismicityand canhave important tectonic and naturalhazard implications mappedfaults (Figure 1). [Wernickeet al., 1998; Connor et al., 1998], we revisit this issueusing vector summationand the new data and 4.3. Deformation in the Basin and RangeProvince error model. We will show that Sierra Nevada block motionsupplies a usefulkinematic boundary condition for Motion of the Sierra Nevada block relative to stable modelsof interiorBasin and Range deformation,that the NorthAmerica is a measureof integratedBasin and Range currentdata suggest that deformationis largelyrestricted to deformation,and should be equivalentto the vectorsum of the easternand westernboundary zones, but that up to 3 slip rates on individual faults and other deformation mm/yr of interior deformationis allowed by the vector sourcesacross the Basin and Range[Minster and dordan, summationapproach within uncertainties,a value that is 1987; Humphreysand Weldon, 1994; Holt and Haines, consistentwith backgroundseismicity in the interior(e.g., 1995]. Of course,some motion may be missedbecause it Figure 1). occurson unrecognizedfaults, or is accommodatedvia Implicit in the vectorsummation approach as used here diffuseaseismic deformation or as magmaticstrain [e.g., is the assumptionthat the site velocity relativeto stable Parsons et al., 1998]. Vector summation, with Sierra North America is a monotonicallyincreasing function of 16 DIXON ET AL.: PRESENT-DAYSIERRA NEVADA MOTION distancefrom the craton,with no changesin the sign of the equivalentwithin errorsto the VLBI-basedvalue given by velocitygradient. In otherwords, given only the kinematic Dixon et al. [ 1995] (98øñ12øwest of north). Sincethe two boundarycondition (Sierra Nevada motion), we cannot azimuthsare derivedby independenttechniques and with precludethe possibility that the velocity "signal" associated independently defined referenceframes, we take the with a zoneof convergenceis coincidentallycancelled by indicatedwesterly motion of the site to be a robust an adjacentregion of extensionin the sparselysampled estimate, and note that it differs significantly (95% interior of the Basin and Range, with correspondinglyconfidence) from the directionof overallplate motion and highertotal deformationthan that estimatedfrom our the direction of Sierra Nevada motion, both northwest simplevector summation approach. However,available within small uncertainties. This is consistent with models kinematic data for the interior [Bennett et al., 1998; thatdrive Basin and Range extension by stressesother than Thatcheret al., 1999] currentlyshow no evidenceof such or in additionto plate boundarytraction [Sonder et al., velocityreversals beyond 95% confidenceintervals. Also 1986; Joneset al., 1996; Thatcher et al., 1999; Sonder we areaware of no significantthrust fault earthquakesin and Jones,1999]. However,our datado not requiresuch the historicrecord for the region. models,as the displacement(or strain)direction does not 4.3.1. Deformation near the eastern boundary of the Basin uniquelydefine the stressdirection. ELYA's westward andRange province. Dixon et al. [1995] estimated4.94-1.3 velocityazimuth probably does define the meanextension mm/yrof motionat Ely Nevadain the centralBasin and direction across the Wasatch fault zone, essentially Rangerelative to stableNorth America, on the basisof a perpendicularto the local,north-south trend of the range limited numberof VLBI experimentsconducted between front(Figure 1). Thus, thereis no evidencefor oblique 1984 and 1990. This site is sufficientlyfar from the extension. Wasatchfault zoneand nearby faults that localizedelastic The similaritybetween the local and regionalvelocity straineffects can be ignoredfor purposesof estimating estimates across the Wasatch fault also places some long-termslip rate from the geodeticdata, but slip on constraintson the magnitudeof elasticstrain effects on the possibleadditional faults near the Wasatch fault would also fault and fault geometry. Savage et al. [1992] fit bereflected in this site'svelocity. The new GPS datafor trilateration data from the local Wasatch network to a Ely (ELYA in Tables1-3) indicate3.44-1.2 mm/yr of standard(fully locked)elastic strain model to estimate westward motion relative to stable North America. Our long-term(far-field) fault slip rates. They obtained 5.64-2.0 newvelocity is similarto the valueof 2.5 mm/yrof west mm/yrof fault slip assuminga 60ø dippingplanar fault directedextension for the easternprovince estimated by geometry,equivalent to 2.7 mm/yr of horizontaleast-west Hearn and Humphreys[1998] on the basisof minimum extension,and 7.64-1.6 mm/yr of fault slip for a listricfault strainenergy considerations. geometry,equivalent to 7.54-1.6mm/yr of horizontaleast- The velocityof ELYA probablyrepresents an upperlimit west extension. The new velocity data for ELYA are not for the rate of horizontal extension across the Wasatch fault permissiveof the high far-fieldrates implied by the latter zone. Published extension rates across local networks model, suggestingeither that a listtic model is spanningthe Wasatchfault zone near 41ø N include inappropriate,or if the fault is listtic, that a simplefully 2.64-0.6mm/yr from trilateration[Savage et a/.,[1992] and locked model is inappropriate,i.e., some motion may 2.74-1.3mm/yr from GPS [Martinez et al., 1998]. The occur aseismically. small differencebetween the ELYA velocity and local rate 4.3.2. Integrated deformationacross the westernboundary estimates(0.84-1.3 mm/yr usingthe local rate estimateof of the Basin and Rangeprovince. The easternCalifornia Savageet al. [1992]) may reflectthe influenceof elastic shearzone [Dokkaand Travis, 1990; Savageet al., 1990] strain accumulation on the fault zone and/or the accommodates124-2 mm/yr of right-lateralslip in the contribution of additional faults outside the local networks; MojaveDesert, oriented N39øW4-5 ø [Sauberet al., 1994]. in eithercase the sum of these effectsis apparentlyquite North of the Garlock fault this motion is accommodated on small. Based on the similarity between our regional two or threeactive strike-slip faults, whose total slip rate extensionestimate from ELYA and the two independent sums to 114-2 mm/yr, based on VLBI and SLR local extension rate estimates for the Wasatch fault zone observationsat OVRO and a simple elasticstrain model [Savageet al., 1992; Martinez et al., 1998], we suggest [Dixonet al., 1995]. Our new datarefine this estimate. thatthe Wasatchfault zoneis the only significantpresently The vector difference between the motion of the Sierra activeextensional structure between Ely Nevada and eastern Nevadablock, defined by our Euler vector,and the motion North America. Dixon et al. [1995] suggestedthe same of ELYA is an estimateof integratedeastern California thing using less precisedata. Thatcher et al. [1999] shearzone deformation plus any additionaldeformation in reached a somewhat different conclusion based on a GPS the interiorof the Basin and Rangewest of ELYA (Figure transectnear 39 ø N, distributingslip betweenthe Wasatch 4). Thisvector difference (114-1 mm/yr oriented N37øW) is fault zone and the Drum Mountain fault 100 km to the equivalentwithin errorsto the local estimateof Sauberet west. This may indicatea changein the partitioningof al. [1994] (12 mm/yr, N39øW), implying that additional deformationalong strike, but distinguishingamong the motion in the interior of the Basin and Range is small. possibilitiesis probablynear the resolvingpower of current The sameconclusion was reached by Dixon et al. [1995] by data. comparingVLBI and SLR regionaldeformation data to Our estimateof the azimuth of ELYA's velocity relative localdeformation data, and by Thatcheret al. [1999] based to stableNorth America, 80ø4-14ø west of north (Table 2) is on a GPS transectacross the Basin and Rangenear latitude DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION 17

[ I ] I [ I ] I [ I

- -10

. ? E xx 6

-- 2 Z

- \ %%,•,,, •.//

' I ' I ! I ' I ' I ' - 12 - 10 -8 -6 -4 -2 0 West Velocity (mm/yr) Figure 4. Velocityvector diagram, illustrating predicted motion of the SierraNevada block relativeto stable NorthAmerica (solid line, SIERRA)for a transectperpendicular to platemotion near OVRO. Measured motionat ELYA(dashed line, EBR; Table 2) representsmotion of theeastern Basin and Range relative to stableNorth America, mainly extension across the Wasatchfault zone. Vectordifference between SIERRA and EBRrepresents motion on the easternCalifornia shear zone (ECSZ)plus any additional deformation in the Basinand Range interior. ECSZ deformation includes right lateral shear on theFish Lake Valley fault zone (dashedline, FLVF) at 8.4 mm/yr, and right-lateral shear on the Owens Valley fault zone (OVF) at 3.0mm/yr (Figure5). Measuredmotion on thesetwo faultsclosely matches total motionbetween the centralSierra Nevadaand ELYA, implyingthat no significantmotion is "missed"on this transect.

39ø N. A velocityprofile through northern Owens Valley with the velocitypredicted by the new Euler vector, i.e., (next section)directly measureseastern California shear the velocity expectedif thesestations were locatedon the zonedeformation, and gives an equivalentresult. stableblock interior. In all cases,sites within the block 4.3.3. Partitioning of deformation within the eastern have velocitiesthat match the Euler vector prediction California shear zone. The slip rates of many of the withinthe 95% confidenceregion, including those not used individualfaults comprisingthe easternCalifornia shear to generatethe bestfitting Euler vector(TIOG, KMED and zone are not well known. In this section we show how the UCD 1), while siteson the margin(QUIN, OVRO, WGRD, newEuler vector can place constraints on theseslip ratesat WMTN and OASI) have velocitiesthat are significantly certainlocations, even where local data are limited (of slowerthan the predictedvelocity. This is not surprising; course uncertainties are reduced if additional local data are all of the latter group of sites are located within the available). We also show that accuratefault slip rate seismicallyactive boundary zone (Figure 1). An Euler estimatesin this regionbased on geodesyrequire some vectorfor the SierraNevada block basedmainly on data knowledgeof past earthquakehistory due to time- fromOVRO andQUINCY wouldtherefore predict a Sierra dependentviscoelastic effects in the lower crust. Nevadavelocity that is biasedto slow values,as thesesites Figure 3 shows the GPS-basedvelocities relative to do not lie on the blockinterior and are affected by elastic stableNorth America for OVRO, QUIN, and otherstations strainaccumulation on its easternmargin. We cantake this nearthe eastern margin of theSierra Nevada block, together one stepfurther: the differencebetween observed motion at 18 DIXON ET AL ß PRESENT-DAY SIERRA NEVADA MOTION

thesesites and motion predictedby our new Euler vector show a total of 5 M>_5.5 events on this fault since 1836. can be used to estimate slip rates on the major nearby Although Holoceneslip rates are not publishedfor this faults, as describedbelow. fault, one segment(the Indian Valley fault) is known to 4.3.3.1.Deformation near QUIN: At Quincy, California, have experiencedHolocene displacement [Jennings and the difference vector between predicted Sierra Nevada Saucedo,1994]. Assumingthat the velocity differenceat velocity and the observedvelocity at QUIN (e.g., 3.1+1.1 QUIN is due to slip only on thesetwo faults, we can solve mm/yr northwestusing the GPS resultin Table 2), requires for the slip rate of the Mohawk Valley fault assumingthe a minimum of severalmm/yr of right-lateralstrike-slip geologicalslip rate for the HoneyLake fault,using the new motion on northwest striking faults near QUIN. One Euler vectorto predictthe motion of a "pseudosite"on the candidateis the Honey Lake fault zone, -50 km northeast block interior,far from the fault (Figure 5a). Earthquake of QUIN, with a Holoceneslip rate of 2 mm/yr [Wills and history and recurrenceinterval are not known here, so we Borchardt, 1993]. Another candidate is the Mohawk revertto the simple elastichalf-space model [Savageand Valley fault zone, the linear valley with intense Burford, 1973] in placeof the viscoelasticcoupling model, microseismicitywhere QUIN is located [Goter et al., linearly superposingresults to representtwo parallelfaults 1994]. A magnitude 6 earthquakeis believed to have [e.g., Dixon et at., 1995]. The limited data are fit by a occurredon the Mohawk Valley fault zone in 1888 model that hasthe Mohawk Valley fault zone slipping at [Toppozadaet al., 1981]. Hill et al. [1991] show two 6+3 mm/yr(Figure 5a). Modelsthat distribute 6-8 mm/yr focal mechanisms northwest of Lake Tahoe near Truckee, of slip equallyon the MohawkValley and HoneyLake California that are consistentwith right lateral strike-slip fault zonesfit the data nearlyas well. Thatcheret at. events on a northwest-strikingfault in this area, with [1999] showa velocity changeof 6-7 mm/yr near this magnitudesof 5.4 (1980) and 6.0 (1966) (seealso Tsai and region,consistent with eithertype of model. Aki [ 1970] and Hawkins et al. [ 1986]). Goter et al. [ 1994] Deformation across the central Nevada seismic belt (Figure1) canalso be estimatedby comparingtotal Sierra Nevadamotion near QUIN to the vector sum of motion in

I I I

Figure5. Simplestrain accumulation models for two parallel 12 strike-slipfaults in an elastichalf-space [Dixon et at., 1995] 14 or anelastic layer over a viscoelastichalf space[Savage and :• • c!:_ -r• • mm/yr Lisowski, 1999] comparedto fault-parallel velocity •1o rJ3rJ3 0 •• - components(triangles, one standarderror bar) on transects _1 I • 6 mm/yr perpendicularto majorfaults on the easternboundary of the SierraNevada block near (a) Quincy,California, and (b,c) Stable Sierra , x OwensValley, California. In Figure5a, elastichalf space 9 mm/yr modelonly is used,and major faults are Mohawk Valley fault zone(MVF)and Honey Lake fault zone(HLF). HLF slip rate fixedat 2 mm/yr[Wills and Borchardt, 1993] and MVF slip ratevaried (3, 6 and 9 mm/yr). Velocity for centralSierra Nevada(SNEV) near Briggs, California (39.4øN, 121.7øW) I I VF •HLF calculatedfrom Euler vector or site velocity data at SUTB, ORVBand QUIN (Tables 2, 6 and8). In Figures5b and 5c, -100 -50 0 5O similarmodels are shownfor the OwensValley (OVF)and FishLake Valley (FLF) fault zones. 5b: FLF rate is fixed at 6 I I I I mm/yr [Dixonet at., 1995; Reheisand Dixon, 1996; Bennett et aL, 1997] (seetext), andOVF slip rateis varied(3 and 7 SNEV OVRO I FLF=6mm/yrJ - mm/yr,dashed lines, and 5 mm/yr, solid line). SNEVrate calculatedfor a pointeast of Fresno(36.8øN, 119.5øW). In El0 this model, no active faults are assumed between FLF and ' WGRD ELYA.Figure 5c showsa closeup of datain centralpart of Figure 5b, with additional strain accumulationmodels. First •WMTN numbershows slip rateon OVF,second number shows slip rateon FLF. Heatflow data (small circles)from Blackwellet ,OASI al. [1991]. Notecoincidence of highlateral gradients in heat flow and surfacevelocity. Thin solidlines showelastic half- mm/yr- GFLD space(EHS) models with equivalentsummed slip rate (11 ELYA mm/yr)but differentpartitioning of slip betweenOVF and FLF. No EHS model fits all data within one standard error. mm/yr••x,-- Viscoelasticcoupling (VEC)model for 3.0 mm/yrslip on mm/yr OVF, 8.4 mm/yrslip on FLF (heavysolid line is summedrate, dashedlines are individual fault rates), with viscosity10 20 -100 0 100 200 300 400 Pa-sec,and lockingdepth, recurrence interval and date of last earthquakefor OVF and FLF of 8 km, 1200 yearsand 1872 Distance (km) AD, and12 km, 900 years and 1400 AD, respectively. DIXON ET AL ß PRESENT-DAY SIERRA NEVADA MOTION 19

I I I I 90 VEC (3/8) 12 o 7/4 0 80 E10 E ..... FLF- . _ • 8 60.• •

o o .... OVF. 50 •-

40

3O 60 -40 -20 0 20 4O 6O Distance (km) Figure 5. (continued) the easternBasin and Range (from ELYA) and motion (Table 8). Figures5b and 5c show a velocity transect acrossthe Mohawk Valley-Honey Lake fault zone, if we perpendicularto the meanshear zone trend, comparing the assume no intervening deformation sources (i.e., our fault-parallelrate component of siteswithin 50 km of the estimateis a maximum). Taking the motion acrossthe transect(all sitesexcept ELYA actuallylie within 15km of MohawkValley-Honey Lake fault zoneat 6-8 mm/yr (from thetransect) with modelsfor two parallelstrike-slip faults, the model in Figure 5a) at an azimuth 320ø (from a map) theOwens Valley-White Mountain fault zoneand the Fish we estimate44-2 mm/yr at 300ø4.10ø for motion acrossthe LakeValley fault zone, the two fault zonescomprising the central Nevada seismic belt (e.g., extension roughly shearzone in thisregion. The two faultsdiffer in strikeby perpendicularto the trend of major normal faults), -•20ø, but givendata and model uncertainties, the parallel equivalentwithin uncertaintiesto the directmeasurement of fault approximationis probablyadequate. The transect Thatcher et al. [ 1999]. beginsat a point in the centralSierra Nevada (36.8øN, 4.3.3.2. Deformation near OVRO: OVRO's velocity is 119.5øW),passes through our sitenear Goldfield, Nevada alsodeficient compared to the centralSierra Nevada block, (GOLD),and ends near ELYA (Figure1), passingbetween in largepart because it is locatedwithin the OwensValley- sites WMTN and WGRD on the White Mountain block, White Mountain fault zone and thus is affectedby strain andpassing 10 km northof OVRO. Thedata constrain the accumulation on both this fault zone and the Fish Lake slip rateof the OwensValley-White Mountain fault zone Valley faultzone immediately to the east(Figures 1 and 3). north of 37øN and the Fish Lake Valley fault zone Here we distinguishbetween the Owens Valley-White immediately to the northeast. We have no site on the Mountainfault zoneslip rate southand north of 37øN, the stable Sierra Nevada block near this transect, but the new latitudeof the valley'sintersection with the Deep Springs Euler vectorallows an accurateestimate of the velocity at fault, and close to the latitude (37.2øN) where the Owens arbitrarylocations on the block. Our data are sensitiveto Valley fault transitionsto the White Mountainfault via a the integratedslip rate across the Owens Valley-White right step. The DeepSprings fault actsto transferright- Mountain fault zone and the Fish Lake Valley fault zone, lateralslip from the OwensValley fault zoneto the south, but are lesssensitive to the partitioningof slip betweenthe acrossthe Deep Springsnormal fault, to the Fish Lake two fault zones, separatedby only about 40 km on this Valley fault zoneto the north [Reheisand Dixon, 1996]. transect. The integratedslip rate from these data (114.1 Recentwork on the Deep Springsfault suggeststhat its mm/yr, Figures 5b and 5c) is equivalentto the estimate long-termhorizontal slip rate is -1 mm/yr [Lee et al., based on vector difference (Figure 4) that included a 2000]. Thus we expectthe slip rate of the OwensValley componentof deformationin the interior of the Basin and fault zone south of 37øN to be -1 mm/yr fasterthan the Range. Thusthe fault-parallelcomponent of any additional rate to the north. deformationbetween the Fish Lake Valley fault zone and It is usefulto resolvethe velocity componentsof sites ELYA is small and possiblyzero. At 95% confidence,the nearthe easternCalifornia shearzone into componentsthat maximum integrated fault-parallel slip rate across the areparallel and perpendicular to the meanshear zone trend interiorBasin and Rangeallowed by our data is 3 mm/yr. 20 DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION

The stronggradient in surfacevelocity across this part of summedslip rate acrossthe Death Valley-FurnaceCreek thewestern Basin and Range (Figure 5b) is striking. This and Hunter Mountain-PanamintValley fault zonesis 5+1 velocity gradientis incompatiblewith models requiringa mm/yr [Bennettet al., 1997]. By differencingour new more or lessconstant strain rate acrossthe Basinand Range estimateof total slip rate acrossthe easternCalifornia shear between stable North America and stable Sierra Nevada. zone(11+ 1 mm/yr)andthe publishedslip rate estimatefor Instead, it suggests that most of the present-day the Death Valley and Hunter Mountain systems (5+1 deformation in the Basin and Range province that mm/yr), the slip rate of the OwensValley fault zone south accommodates northwest motion of the Sierra Nevada is of 37øN would be 6+ 1 mm/yr. Publishedestimates for the concentratedin the westernboundary of the province, at slip rate on the Owens Valley fault zone include 7+1 leastfor thistransect (see next paragraph for the limitations mm/yr from trilaterationdata fit to an elastic strain model of this argument). In the vicinity of the transect, [Savageand Lisowski, 1995], 4+1 mm/yr from VLBI and essentially all of this deformation appears to be SLR data at OVRO and an elastic strain model [Dixon et concentratedin a zone of mainly strike-slip faulting less al., 1995], and 2+1 mm/yr based on geological data than 50 km wide on the Sierra Nevada block's eastern [Beanlandand Clark, 1994]. Thus the publishedestimates boundary. This region of high surfacevelocity gradient differ significantly. coincideswith a region of rapid change in heat flow North of 37.3øN,the Hunter Mountain-PanamintValley betweenthe easternSierra margin and the westernBasin fault zone disappearsin a seriesof north striking normal andRange (Figure 5c), as well as high lateralgradients in faults,near the intersectionof the Death Valley- Furnace electricalresistivity in the middle crust [Park et al., 1996] Creekfault zoneand the Fish Lake Valley fault zone(e.g., and P wave velocity in the uppermantle [Humphreysand Figure 1 of Dixon et al., 1995]. Kinematic arguments Dueker, 1994; Dueker, 1998]. These correlationssuggest suggestthat essentiallyall of the motion on the combined that the surfacevelocity distributionclosely reflectsdeep- Hunter Mountain-PanamintValley plus Death Valley- seatedprocesses or structure,and not simply local variation Furnace Creek fault zones is transferred into the Fish Lake in uppercrustal structure. Valley fault zone [Dixon et al., 1995; Reheis and Dixon, This narrow zone of deformationand correspondingly 1996; Reheisand Sawyer, 1997]. In addition, north of its high velocity gradienthas implicationsfor contemporary intersectionwith the Deep Springsfault (the locationof our strain rate estimatesfor the Basin and Range province. transect),the Fish Lake Valley fault zone acquiresan Theserates are sometimescalculated by taking estimated additional 1 mm/yr of slip, suggestinga total of 6+1 SierraNevada motion, and dividing by the width of the mm/yr if we assumethe elastichalf-space model of Bennett entireBasin and Range province (-103 km) or anassumed et al., [1997]. This would suggestthat the OwensValley- width for the actively deforming boundary zone, often White Mountainfault zonenorth of 37øN slips at a rate of takenas severalhundred km. The datain Figure 5b and 5c about5+1 mm/yr (11+1 mm/yr - 6+lmm/yr; Figure5b). If indicate that, at least for this location, the zone of active the Fish Lake Valley fault zone insteadslips at 4 mm/yr, deformation is restricted to a few 10's of km in width. which is allowedby availabledata within 95% confidence, However, the width of the actively deforming zone, as then the slip on this segmentof the Owens Valley fault indicatedby the location of active faults and seismicity, zone would be correspondinglyhigher, 7 mm/yr, an widensconsiderably both northand south of this particular estimate consistent with the available GPS data if we use location(e.g., Figure 1). This suggeststhat strainrates in an elastic half-spacemodel (Figure 5c) but inconsistent the regionmay exhibit a high degreeof spatialvariability, with availablegeologic data and inconsistentwith the GPS and mean strain rate estimates for the entire Basin and data if we insteaduse a viscoelasticcoupling model (see Rangeprovince, as well as generalizationsconcerning the nextparagraph and Figu•'e 5c). Forthe data in Figure5b, width of the deformingboundary zones, may not be the bestfit (minimumZ ) estimatefor the elastichalf-space particularlymeaningful. Mapping the spatialvariability in solution is 6+2 mm/yr for the Owens Valley-White deformation,and understanding its origins,is an important Mountain fault zone and 5+2 mm/yr for the Fish Lake futuregoal for Basinand Range studies. Valleyfault zone, with reducedZ2= 0.5. Thisis Why is the deformationzone associated with the eastern essentiallyidentical to the estimateof Savageand Lisowski California shear zone so narrow at the north end of Owens [1995] for the Owens Valley fault zone south of 37øN Valley comparedto regions north and south? One basedon terrestrialgeodetic data and an elastichalf space possibilityis thatLong Valley Caldera,part of a long-lived model (7+ 1 mm/yr), recallingthe expectedrate differential volcaniccenter active for the last few million years, has of 1 mm/yr north and southof 37øN. locally heated and weakened the crust, focussing Our fault slip rate estimatefor the OwensValley-White deformationin thisparticular area (Figures 1 and3). Mountainfault zoneusing an elastichalf-space model (6+2 Slip rateestimates for the individualfaults comprising mm/yr, or 7+2 mm/yr south of 37øN) is fasterthan the the shear zone can be estimated as follows. South of 37øN, geologic estimate of 2+ 1 mm/yr [Beanland and Clark, the eastern California shear zone consists of three sub- 1994]. Part of this difference could reflect the fact that the parallelfaults. From east to west theseare the Death geodeticestimate integrates motion acrossthe entire fault Valley-FurnaceCreek, HunterMountain-Panamint Valley, zone and is indicative of the overall differential motion at and Owens Valley-White Mountain fault zones, all depth,while the geologicestimate reflects motion on well- northwestto north-northweststriking faults (Figure 1). developedstrands of the fault with surfaceexpression. GPS datafit to an elastichalf-space model suggestthat the Dixon et al. [1995] suggestedthat the OwensValley fault DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION 21 is a relativelyyoung feature. Such faults may not haveyet Table 8. Velocity Components Perpendicular and developeda clear, throughgoingsurface expression, and Parallel to Eastern California Shear Zone hencethe difference between geodetic and geologic slip rate estimatescould be significant. Veloci_ty, mm/yr However,the long-termviscoelastic effects of the 1872 . ab Owens Valley earthquakemay also explain part of the Parallela'b Perpendicular' discrepancybetween the geologicrate and geodeticrates basedon an elastic half-spacemodel. We can use the SNEVc 12.5+ 0.7 4.5 + 0.9 viscoelasticcoupling model (equation3) to investigate OVROd 10.2+ 0.8 3.1+ 0.8 this. Ratherthan use the defaultparameter values listed in Table 5 to constrainthe model, we can exploit the WGRD 7.4 + 1.1 3.4 + 1.4 relativelyhigh datadensity in our transectto estimatesome WMTN 6.4 + 0.8 1.7 + 1.0 of the poorlydetermined parameters. Specifically, we fix OASI 4.5 + 1.9 -1.3 + 2.5 the dateof the last earthquakeon the OwensValley fault zoneto 1872, fix the lockingdepths on the OwensValley GFLD 1.8 + 1.9 -0.9 + 2.5 andFish Lake Valley faultzones to 8 and 12 km [Dixonet ELYA 2.0 + 1.0 2.7 + 1.1 al., 1995], and adjustthe remainingsix parametersto a Velocityrelative to stableNorth America, Table 2. obtainbest fit values (two slip rates, two earthquake b recurrenceintervals, date of last event on the Fish Lake Assumed fault azimuth is 333 ø, mean of Owens Valleyfault zone, and lower crustal viscosity•. The best fit Valley and Fish Lake Valley fault zones modelis shownin Figure5c, with reducedZ = 0.65. We c Predictedvalue is from SierraNevada Euler vector at a point 36.8øN, 119.5øW. obtainslip rates of 3.0+1.9 mm/yr and 8.4+2.0 mm/yr for d the Owens Valley and Fish Lake Valley fault zones, Mean velocity from Table 4. respectively.Best estimatesfor the otherparameters are listed in the figurecaption. Total right lateralslip rate acrossboth faults in thismodel, 11.4+1.1 mm/yr, is better Sawyernote that the slip ratemay havevaried with time, constrainedthan the individualfault slip rates,as with the and may be slowerat later times. elastichalf spacemodel. The velocityprofiles for the two The three-dimensionalgeometry of the faultshas been faults(Figure 5c) differgreatly, because the OwensValley ignoredin oursimple coupling model. Thatgeometry, as fault zoneis in the early stagesof its earthquakecycle, well as betterdata on earthquakehistory to constrainthe while the Fish LakeValley fault zoneapparently is in the viscoelasticresponse of the lower crust and reducethe late stagesof its cycle. Applying an elastichalf-space numberof adjustableparameters, needs to be incorporated modelin this casewould causethe slip rate of the Owens in futuremodels to improvethe accuracy of faultslip rate Valley fault to be over-estimated(due to the "additional" estimatesderived from geodetic data. Sincethe modelhas velocitynear the fault), and the slip rate on the Fish Lake somesensitivity to earthquakerecurrence interval and date Valley fault zoneto be slightly underestimated.Note also of lastearthquake, it is alsopossible that a densearray of thatthe summed velocity profile for the two faultsis quite highquality geodetic data could provide useful constraints asymmetric,and that the Fish Lake Valley fault zone on these parametersindependent of estimates from causessignificant strain accumulation well eastof the fault paleoseismicstudies. dueto its late stagein the earthquakecycle. For example, Figure4 is a vectordiagram, showing motion of the 50 km fromthe fault,a geodeticnetwork spanning a 25 km SierraNevada block from our Eulervector (equivalent to aperturewould experienceabout 0.5 mm/yr of differential integratedBasin and Range deformation), compared to the velocitydue solelyto strainaccumulation on the Fish Lake sum of deformationacross major known faults: eastern Valleyfault zone (see also Savage et al., [1999]). Accurate Basin and Range deformation(velocity of ELYA, estimationof slip ratesfor minor faults in the Basin and essentiallyrepresenting extension across the Wasatchfault) Rangeinterior near the westernboundary will therefore andstrike-slip motion on the Fish LakeValley and Owens requirean accurateaccounting of such viscoelasticeffects Valley-WhiteMountain fault zones. The equivalence of the andthe earthquakecycle. Sierra Nevada vector and the summed deformationvectors The fault slip rate estimatesfrom our viscoelastic (thefit is almostas good if we insteaduse the faultslip couplingmodel are in reasonableagreement with geological ratesestimated from the elastichalf spacemodel) suggests estimates.The geologicalestimate for slip on the Owens that no significantdeformation has been missed in this Valley fault zonesouth of 37øN (2+1 mm/yr; Beanland path,equivalent to thestatement that there is no significant and Clark, 1994) is equivalentto our estimate(3.0+1.9 deformationwithin the interiorof theBasin and Range on mm/yr+ 1 mm/yr = 4+2 mm/yr) within one standarderror. thistransect, within observational error. However,given Reheisand Sawyer[ 1997]give a detailedslip ratehistory errorpropagation in this vectorsummation approach, this for varioussegments of the Fish Lake Valley fault zone. is not a strongconstraint on interiorBasin and Range For the main, straightsection of the fault zone (Oasis deformation,especially the extensional(east-west) section),the averagerate after 620,000 years BP is 9.5+2.2 component. Also, as noted above, it assumesa monotonic mm/yr, equivalentto our estimateof presentday rate velocityprofile; only denseprofiles [Bennett et al., 1998; (8.4+2.0 mm/yr) within one standarderror. Reheis and Thatcheret al., 1999]can test this assumption. 22 DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION

A smallamount of extensionperpendicular to the trend example,near Quincy, California, summed right-lateral slip of the easternCalifornia shearzone is suggestedby the acrossthe Mohawk Valley and Honey Lake fault zonesis differencein fault-perpendicularvelocities between ELYA 8+3 mm/yr,based on thevelocity of Quincyand a geologic and the central Sierra Nevada (1.8+1.4 mm/yr; Table 8). sliprate estimate for theHoney Lake fault zone. Integrated Most of this extension is accommodated between the Sierra right lateralslip acrossthe easternCalifornia shear zone in Nevadaand OASI, just eastof the Fish Lake Valley fault the vicinity of OVRO is 11.4+1.1 mm/yr, basedon a zone, i.e., extensionis likely accommodatedon normal velocitytransect and a viscoelasticcoupling model. Of faults within OwensValley and Fish Lake Valley (Figure this total, 3+2 mm/yr is accommodatedacross the Owens 3; Table 8). Savage and Lisowski [1995] estimated Valley-White Mountain fault zone north of 37øN, and 8+2 1.0+0.3 mm/yr of horizontal extension across Owens mm/yr is accommodatedacross the Fish Lake Valley fault Valley usingleveling data and an elasticstrain model for a zone,consistent with geologicalestimates. Together, these dip-slipfault. Active extensionhere is consistentwith the two faultsdefine a regionof steepvelocity gradienton the valley'srelief, the observeddip-slip componentof motion easternboundary of the Sierra Nevada block, spatially on the Owens Valley fault during the 1872 earthquake coincident with steep gradients in other geophysical [Beanlandand Clark, 1994], and observeddip-slip motion parameterssuch as heat flow. Earthquakehistory and on the nearbyIndependence fault, an activenormal fault on viscoelastic effects need to be considered in order to obtain the west side of Owens Valley. A small amount of accuratefault slip rate estimatesfrom geodeticdata in this extension(few tenths of mm/yr) also likely occursacross region. Vector summationand the kinematic constraint Fish Lake Valley basedon geologicalstudies [Reheis and from our Sierra Nevada Euler vector suggest little Sawyer, 1997]. deformationin the interiorof the Basin and Rangeon one We can estimate the amount of net extension between particular transect approximatelyperpendicular to plate Fish Lake Valley and ELYA by subtractingextension motion through northern Owens Valley, assuming a across Owens Valley (1.0+0.3 mm/yr) and Fish Lake monotonicvelocity profile across the region. Valley (taken as 0.2-½0.2mm/yr) from total extension betweenthe Sierra Nevada and ELYA (1.8-½1.4 mm/yr). This gives0.6+1.4 mm/yr, permissiveof the -1 mm/yr of Acknowledgments. We are grateful to the many east-westextension acrossthe proposed nuclear waste studentsand colleagueswho contributed to the field repository at Yucca Mountain, Nevada estimated by campaignsdescribed in this report,including Steve Fisher, Wernickeet al. [1998]. Bjom Johnsand Karl Feauxof Unavco. High-qualitydata from permanentstations within the BARD, CORS and 5. Conclusions IGS networks, and accessto archives at JPL, SIO and GSFC were criticalto this study. We thank the operators We have derived the first Euler vector describing of thosenetworks and archives for generouslysharing these present-daymotion of the Sierra Nevada block using data.Measurements between 1993 and 1996were supported geodeticdata from siteslocated on the stableinterior of the by NASA's Dynamicsof the Solid Earth program. The block. The data indicatethat the block is rigid within 1998observations were performed with help from PeterLa velocityuncertainties. Our Euler vectorpredicts northwest Femina, Ron Martin and Lena Krutikov, Center for Nuclear motion of the block relative to stable North America at WasteRegulatory Analyses, and supportof JohnTrapp and fasterrates than previousestimates. We see no vertical Phil Justusat the U.S. NuclearRegulatory Commission motionwithin uncertainties,except at a site within Owens (NRC), under contractNRC-02-97-009. This paper is an Valley east of the Sierra Nevada block, which may be independentproduct and does not necessarilyreflect the subsiding. We see no evidenceof rapid rotation of the views or regulatoryposition of NRC. Todd Williams and Sierra Nevada block about a proximal pole; rather, it EdmundoNorabuena helped construct several figures. We translatesnearly parallel to overall plate motion, indicating thank Jim Savagefor discussionson viscoelasticity,and a distalEuler pole. The new Euler vectorcan be usedas a Wayne Thatcher,Tom Parsons,Don Argus and Editor kinematic constraintfor regional tectonicproblems. For Dave Schollfor commentson the manuscript.

References

Argus,D. F., and R. G. Gordon, Current Sierra America, Geol. Soc. Am. Bull., 81, 3513- earthquake,U.S. Geol. Surv. Bull., 1982, Nevada-NorthAmerica motion from very 3536, 1970. 29p., 1994. long baselineinterferometry: Implications Atwater, T., Plate tectonic history of the Bennett, R. A., B. Wernicke, J. Davis, P. for the kinematics of the western United northeast Pacific and western North Elosgui,J. Snow,M. Abolins,M. House,G. States,Geology, 19, 1085-1088,1991. America, Geologyof North America, vol. N: Stirewalt,and D. Ferrill, Global Positioning Argus, D. F., and R. G. Gordon, Tests of the EasternPacific Ocean and Hawaii, edited by Systemconstraints on fault slip rates in the rigid plate hypothesis and bounds on E. L. Winterer, D. M. Hussong,and R. W. Death Valley region, California and intraplatedeformation using geodetic data Decker, Geol. Soc. Am., Boulder, Colorado, Nevada,Geophys. Res. Lett., 24, 3073-3076, from very longbaseline interferometry, J. 21-72, 1989. 1997. Geophys.Res., 101, 13,555-13,572,1996. Beanland, S., and M. M. Clark, The Owens Bennett,R. A., B. Wernicke, and J. Davis, Atwater,T., Implicationsof platetectonics for Valley fault zone, eastern California, and Continuous GPS measurements of the Cenozoic evolution of western North surface faulting associated with the 1872 contemporary deformation across the DIXON ET AL.: PRESENT-DAY SIERRA NEVADA MOTION 23

northern Basin and Range province, Goter, S. K., D. Oppenheimer, J. Mori, M. Ma, C., J. W. Ryan, and D. S. Caprette,NASA Geophys.Res. Lett., 25, 563-566, 1998. Savage, and R. Masse, Earthquakes in Space Geodesy Program, GSFC data Bennett, R. A., J. Davis, and B. Wernicke, California and Nevada, U.S. Geol. Surv., analysis,1993, VLBI geodeticresults, 1979- Present-day pattern of Cordilleran OpenFile Rep. 94-967, 1994. 1992, NASA Tech.Memo. TM-104605, 1994. deformation in the western United States, Hawkins, F. F., R. LaForge and R. A. Hansen, Mao, A., Geophysical applicationsof the Geology,27, 371-374, 1999. Seismotectonicstudy of the Truckee/Lake Global Positioning System, Ph.D. thesis, Blackwell, D.D., J.C. Steele and L.S. Carter, Tahoe are, northeastern Sierra Nevada, Univ. Miami, Miami, Florida, 1998. Heat flow patterns of the North American California, U.S. Bureau Reclamation Mao, A., C. G. A. Harrison, and T. H. Dixon, continent;a discussionof the geothermal SeismotectonicRept., 85-4, 210p., 1986. Noise in GPS coordinate time series, d. map of North America, in: Neotectonicsof Hearn, E. H., and E. D. Humphreys, Geophys.Res., 104, 2797-2816, 1999. North America, edited by D.B. Stemmons, Kinematics of the southern Walker Lane Marshall, G., J. Langbein, R. Stein, M. M.D. Zoback, and D.D. Blackwell, Geol. Belt and motion of the Sierra Nevada block, Lisowski,and J. Svarc, Inflation of Long Soc. Am., Boulder, Colorado, 1991. California, d. Geophys. Res., 103, 27,033- Valley caldera,California, Basin and Range Bowden, G. W., A. Donnellan, L. H. Kellogg, 27,049, 1998. strain, and possible Mono Craters dike D. Dong, and J. B. Rundle, Geodetic Hill, D. P., J.P. Eaton, W. L. Ellsworth, R. S. opening from 1990-1994 GPS surveys, measurements of horizontal strain near the Cockerham, F. W. Lester, and E. J. Corbett, Geophys.Res. Lett., 24, 1003-1006, 1997. The seismotectonic fabric of central White Wolf fault, Kern County, California, Martinez, L. J., C. Meertens, and R. B. Smith, 1926-1993, d. Geophys. Res., 102, 4957- California,in The Geologyof North America, Rapid deformationrates along the Wasatch 4967, 1997. eds. D. B. Slemmons,E. R. Engdahl,M.D. fault zone, Utah, from GPS measurements California Division of Mines and Geology, Zoback and D. D. Blackwell, Geol. Soc. with implications for seismic hazard, Californiafault parameters,DMG Open File Am., DecadeMap Volume1, 107-132, 1991. Geophys.Res. Lett., 25, 567-570, 1998. Report,96-08, Sacramento,California, 1996. Holt, W. E., and A. J. Haines, The kinematics Chase, C. G., Plate kinematics: The Americas, of northern South Island, New Zealand, McGill, S., Paleoseismologyand neotectonics East Africa, and the rest of the world, Earth determined from geologic strain rates, d. of the central and eastern Garlock fault, Planet Sci. Lett., 37, 355-368, 1978. Geophys.Res., 100, 17,991-18,010, 1995. Ph.D. thesis, Calif. Inst. Tech., Pasadena, California, 1992. Connor, C., Stamatakos,J. A., D. A. Ferril, B. Humphreys,E. D., and K. G. Ducker, Western E. Hill, Detecting strain in the Yucca McGill, S., and K. Sieh, Holocene slip rate of U.S. upper mantle structure,d. Geophys. the central Garlock fault in southeastern Mountain area, Nevada: Comment, Science, Res., 99, 9615-9634, 1994. 282, 1007b,1998. Humphreys, E. D., and R. J. Weldon, SearlesValley, California, d. Geophys.Res., DeMets C., and T. H. Dixon, New kinematic Deformation across the western United 98, 14,217-14,231, 1993. models for Pacific North America motion States: A local estimate of Pacific-North Minster, J. B., and T. H. Jordan, Present-day from 3 Ma to present:Evidence for steady America transform motion, d. Geophys. plate motions,d. Geophys.Res., 83, 5331- motion and biasesin the NUVEL-1A model, Res., 99, 19,975-20,010, 1994. 5354, 1978. Geophys.Res. Lett., 26, 1921-1924, 1999. Jennings,C. W., and G. J. Saucedo, Fault Minster, J. B., and T. H. Jordan, Vector constraints on western U.S. deformation DeMets,C., R. G. Gordon,D. F. Argus, and S. Activity Map of California and Adjacent Stein, Current plate motions,Geophys. d. Regions,With Locationsand Ages of Recent from spacegeodesy, neotectonics and plate Int., 101,425-478, 1990. VolcanicEruptions, scale 1:750,000,Calif. motions,d. Geophys.Res., 92, 4798-4804, 1987. Dixon, T. H., and A. Mao, A GPS estimate of Geol.Data Map Ser., map 6, Calif. Div. Of relative motion between North and South Mines and Geol., Sacramento,California, Moore, J. G., and C. A. Hopson, The America, Geophys.Res. Lett., 24, 535-538, 1994. Independence dike swarm in eastern 1997. Johnson,H. O., and D.C. Agnew, Monument California, Am. d. Sci., 259, 24-259, 1961. Dixon, T. H., S. Robaudo, J. Lee, and M. motion and measurement of crustal Norabuena, E. O., T. H. Dixon, S. Stein, and C. Reheis,Constraints on present-dayBasin and velocities, Geopbys. Res. Lett., 22, 2905- G. A. Harrison, Decelerating Nazca-South Range deformation from space geodesy, 2908, 1995. America and Nazca-Pacific plate motions, Tectonics,14, 755-772, 1995. Jones,C. H., J. R. Unruh, and L. J. Sonder,The Geophys.Res. Lett., 26, 3405-3408, 1999. Dixon, T. H., A. Mao, and S. Stein,How rigid is role of gravitationalpotential energy in Okada, Y., Internal deformation due to shear the stable continental interior of the North active deformationin the southwesternUS, and tensile faults in a half-space, Bull. American plate, Geophys. Res. Lett., 23, Nature, 381, 37-41, 1996. Seismol.Soc. Am., 82, 1018-1040, 1992. 3035-3038, 1996. Langbein,J. and H. Johnson,Correlated errors Park, S. K., B. Hirasuna, G. R. Jiracek, and C. Dixon, T. H., A. Mao, M. Bursik, M. Heftin, J. in geodetictime series:implications for time Kinn, Magnetotelluric evidence of Langbein, R. Stein, and F. Webb, 1997, dependentdeformation, d. Geophys.Res., lithospheric mantle thinning beneath the Continuous monitoring of surface 102, 591-603, 1997. southernSierra Nevada, d. Geophys. Res., deformation at Long Valley caldera, Langbein,J., F. Wyatt, H. Johnson,D. Hamann, 101, 16,241-16,255, 1996. California,with GPS, d. Geophys.Res., 102, and P. Zimmer, Improved stability of a Parsons,T., G. A. Thompson,and R. B. Smith, 12,017-12,034, 1997. deeply anchored geodetic monumentfor More than one way to stretch: A tectonic Dokka, R. K. and C. J. Travis, Role of the deformationmonitoring, Geopbys. Res. Lett., modelfor extensionalong the plumetrack of Eastern California shear zone in 22, 3533-3536, 1995a. the Yellowstonehotspot and adjacent Basin accommodating Pacific-North American Langbein,J. O., D. Dzurisin, G. Marshall, R. and Rangeprovince, Tectonics, 17, 221-234, plate motion,Geophys. Res. Lett., 17, 1323- Stein, J. Rundle, Shallow and peripheral 1998. 1326, 1990. volcanicsources of inflation revealed by Pezzopane, S. K., and R. Weldon, Tectonic Ducea, M. N., and J. B. Saleeby, Buoyancy modelingtwo-color geodimeter and leveling role of active faulting in central Oregon, sources for a large, unrooted mountain data from Long Valley caldera, California, Tectonics,12, 1140-1169, 1993. range, the Sierra Nevada, California: 1988-1992,J. Geopbys.Res., 100, 12,487- Pollitz, F. F., Gravitational visco-elastic evidencefrom xenoliththermobarometry, d. 12495, 1995b. postseismicrelaxation on a layeredspherical Geophys.Res., 101, 8229-8244, 1996. Lee, J., C. M. Rubin,A. Blythe,P. Gans,and A. Earth,d. Geophys.Res., 102, 17,921-17,941, Ducker,K., New velocity imageof the western Calvert,Quaternary faulting history along 1997. US upper mantle: is heat input from below the Deep Springsfault, California, Geol. Pollitz, F. F., and T. H. Dixon, GPS the lithosphererequired? Eos Trans. AGU, Soc.Am. Bull. (in press,2000). measurements across the northern 79 (45), Fall Meet. Suppl.,F-559, 1998. Li, V. C., and J. R. Rice, Crustal deformation in Caribbean plate boundary: Impact of Freymueller,J. T., M. Murray, P. Segall, D. great California earthquakes,d. Geophys. postseismicrelaxation following historic Castillo, Kinematics of the Pacific-North Res., 92, 11,533-11,551, 1987. earthquakes,Geophys. Res. Lett., 25, 2233- America plate boundary zone, northern Ma, C., andJ. W. Ryan,NASA SpaceGeodesy 2236, 1998. California, d. Geophys. Res., 104, 7419- Program, GSFC data analysis, VLBI Pollitz, F.F., and I.S. Sacks, Modeling of 7441, 1999. geodeticresults, 1979-1998, 1998. postseismicrelaxation following the great 24 DIXON ET AL.: PRESENT-DAY SIERRANEVADA MOTION

1857 earthquake,southern California, Bull. United States determined t¾om earthquake Motion of the Sierran-GreatValley block, SeismoLSoc. Am., 82, 454-480, 1992. moment tensors, very long baseline Int. Geol. Rev., 40, 403-410, 1998. Reheis,M. C., and T. H. Dixon, Kinematics of interferometry,and GPS observations,d. Wernicke,B., et al., Origin of high mountains the eastern California shear zone: Evidence Geophys.Res., 103, 18,087-18,117,1998. in the continents: The southern Sierra for slip transfer from Owens and Saline Sillard,P., Z. Altamimi, and C. Boucher,The Nevada,Science, 271, 190-193, 1996. Valley fault zonesto FishLake Valley fault ITRF96 realization and its associated Wemicke,B., J. Davis,R. Bennett,P. Elosegui, zone,Geology, 24, 339-342, 1996. velocityfield, Geophys. Res. Lett., 25, 3222- M. Abolins,R. Brady, M. House,N. Niemi, Reheis, M. C., and T. L. Sawyer, Late 3226, 1998. andJ. Snow,Anomalous strain accumulation Cenozoichistory and slip rates of the Fish Smith,R. B., and W. J. Arabasz, Seismicityof in the Yucca Mountain area, Nevada, Lake Valley, Emigrant Peak and Deep the Intermontane seismic belt, in Decade Science,2 79, 2096-2100, 1998. Springsfault zones,Nevada and California, Map.,vol. 1, Neotectonicsof NorthAmerica, Wessel,P., andW. H. F. Smith,New versionof Geol. Soc. Am. Bull., 109, 280-299, 1997. editedby D. B. Slemmonset al., pp. 185-228, genericmapping tools released, Eos Trans. Ross, D.C., Correlation of granitic plutons Geol. Soc.Am., Boulder,Colorado, 1991. AGU, 76, 329, 1995. across faulted Owens Valley, California, Sonder,L. J., and C. H. Jones,Western United Wills, C. J., and G. Borchardt,Holocene slip U.S. Geol. Survey, Pro•ssional Paper 450- States extension: How the west was rate and earthquake recurrence on the D, D-86-D-88, 1962. widened, Ann. Rev. Earth Planet Sci., 27, Honey Lake fault zone, northeastern Sauber, J., W. Thatcher, S. Solomon, and M. 417-462, 1999. California,Geology, 21, 853-856, 1993. Lisowski,Geodetic slip rate for the eastern Sonder, L. J., P. C. England, and G. A. Working Group on California Earthquake California shear zone and the recurrence Houseman, Continuum calculations of Probabilities,Seismic hazards in southern time of Mojave Desert earthquakes,Nature, continental deformation in transcurrent California earthquakes, 1994-2024, Bull. 367, 264-266, 1994. environments,d. Geophys.Res., 91, 4797- Seismol.Soc. Am., 85, 379-439, 1995. Savage, J. C., Detecting strain in the Yucca 4810, 1986. Wright, L. Late Cenozoicfault patternsand Mountain area, Nevada: Comment, Science Thatcher,W., Nonlinear strainbuildup and the stress fields in the Great Basin and westward 282, 1007, 1998. earthquakecycle on the SanAndreas fault, displacementof the Sierra Nevada block, Savage, J. C., and R.O. Burfurd, Geodetic d. Geophys.Res., 88, 5893-5902,1983. Geology,4, 489-494, 1976. determinationof relative plate motion in Thatcher,W., G. R. Foulger, B. R. Julian,J. Zhang, J. Y. Bock, H. Johnson,P. Fang, S. central California, d. Geophys. Res., 78, Svarc, E. Quilty, and G. W. Bawden, Williams, J. Genrich, S. Wdowinski, and J. 832-845, 1973. Present-daydeformation across the Basin Behr, SouthernCalifornia permanentGPS Savage, J. C., and M. Lisowski, Strain and Rangeprovince, western United States, geodetic array: Error analysis of daily accumulation in Owens Valley, Bull. Science,283, 1714-1718, 1999. positionestimates and site velocities,d. Seismol. Soc. Am., 85, 151-158, 1995. Toppozada,T. R., C. R. Real, andD. L. Parke, Geophys.Res., 102, 18,035-18,055,1997. Savage, J. C., and M. Lisowski, Viscoelastic Preparation of isoseismalmaps and Zumberge, J., M. Heftin, D. Jefferson, M. couplingmodel of the San Andreas Fault summariesof reportedeffects for pre-1900 Watkins,F. Webb, Precisepoint positioning alongthe big bend, southernCalifornia, d. California earthquakes, Open File Report for efficientand robustanalysis of GPS data Geophys.Res., 103, 7281-7292, 1998. 81-11, 182 p., Calif. Div. Of Mines and from largenetworks, d. Geophys.Res., 102, Savage,J. C., M. Lisowski,and W. H. Prescott, Geol., Sacramento,California, 1981. 5005-5017, 1997. An apparent shear zone trending north- Tsai, Y., andK. Aki, Sourcemechanism of the northwest across the Mojave Desert into Truckee, California, earthquake of Owens Valley, Geophys. Res. Lett., 17, September12, 1966,Bull. Seismol. Soc. Am., T. H. Dixon, F. Farina, and H. Wang, 2113-2116, 1990. 60, 1199-1280, 1970. RosenstielSchool of Marine and Atmospheric Savage,J. C., M. Lisowski,and W. H. Prescott, Ward, S. N., Pacific-North America plate Science, University of Miami, 4600 Strain accumulation across the Wasatch motions: New results from very long RickenbackerCauseway, Miami, FL 33149- fault near Ogden, Utah, d. Geophys.Res., baseline interferometry, d. Geophys. Res., 1098.(tdixon•rsmas.miami.edu). 97, 2071-2083, 1992. 95, 21,965-21,981, 1990. D. Johnsonand M. Miller, Departmentof Savage,J. C., J. L. Svarc, and W. H. Prescott, Wells, R. E., C. S. Weaver, and R. J. Blakely, Geology, Central WashingtonUniversity, Strain accumulation at Yucca Mountain, Fore-arc migration in Cascadia and its Ellensburg,WA 98926. Nevada, d. Geophys. Res., 104, 17,627- neotectonicsignificance, Geology, 26, 759- 17,631, 1999. 762, 1998. (ReceivedNovember 23, 1998; Shen-Tu, B., W. E. Holt, and A. J. Haines, Wemicke, B. and J. K. Snow, Cenozoic revisedSeptember 13, 1999; Contemporary kinematics of the western tectonismin the central Basin and Range: acceptedOctober 27, 1999.)