JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 99, NO. B10, PAGES 19,975-20,010, OCTOBER 10, 1994

Deformation across the western United States: A local estimate of Pacific-North America transform deformation

EugeneD. Humphreysand Ray J. Weldon II Deparmaentof GeologicalSciences, University of Oregon,Eugene

Abstract. We obtaina locally basedestimate of Pacific-NorthAmerica relativemotion andan uncertaintyin this estimateby integratingdeformation rate alongthree different pathsleading west acrosssouthwestern North America from east of the Rio Grande Rift to nearthe continentalescarpment. Data areprimarily Quaternary geologic slip rate estimates,and resultingdeformation determinations therefore are "instantaneous"in a geologicsense but "longterm" with respectto earthquakecycles. We deducea rate of motionof the Pacificplate relative to North Americathat is 48 :E2mm/yr, a rate indistinguishablefrom that predictedby the globalkinematic models RM2 and NUVEL-1; however, we obtain an orientation that is 5-9 ø counterclockwise of these models. A morewesterly motion of the Pacificplate, with respectto North America, is calculatedfrom all threepaths. The relativelywesterly motion of the Pacificplate is accommodatedby deformationin the NorthAmerican continent that includes slip on relativelycounterclockwise-oriented strike-slip faults (including the SanAndreas fault),which is especiallyrelevant in andsouth of theTransverse Ranges, and a margin-normalcomponent of net extensionacross the continent,which is especially relevantnorth of the TransverseRanges. Deformationof the SW United Statesoccurs in regionallycoherent domains within which the styleof deformationis approximately uniform.In thevicinity of theTransverse Ranges, two importantshear systems splay fromthe SanAndreas fault: the easternCalifornia shear zone trending NNW fromthe eastemTransverse Ranges and the trans-Peninsularfaults trending SSE from the westernand central Transverse Ranges. Within the TransverseRanges the right-lateral SanAndreas fault steps left, seeminglyrequiring large amounts of convergencethere. However,most of thisconvergence is avoidedthrough a "funnelingflow" of the crust towardthe westernTransverse Ranges and into the relativelynarrow central California CoastRanges and the northernmotion of theMojave. The formerprocess involves an alternationof rotationdirection from counterclockwise(in and southof the central TransverseRanges) to clockwise(in the westernTransverse Ranges).

Introduction regional plate margin kinematics consistent with the globally derived Pacific-NorthAmerican plate velocity, The occurrenceof broadlydistributed young faulting in the westernUnited States(Figure la) affordsan oppor- these models included unrealisticallygreat deformation tunity to study the process of continental transform rates somewherein the southernCalifornia region. The deformationin a well-studiedregion. Essentialto sucha models of Hill [1982] and Bird and Rosenstock[1984] study is an understandingof the regionalkinematics of include excessiveconvergence rates acrossthe length of this broadlydistributed and diversedeformation field. An the TransverseRanges. Weldon and Humphreys[1986] importantaspect is knowledgeof the far-field plate introduceda model that has relatively low convergence motionsthat fundamentallydrive transformdeformation. rates across the central and eastern TransverseRanges Previous kinematic models of the southwestern United and high convergencerates acrossthe westernTransverse States have been constructed so that the total deformation Ranges by including relatively high rates of strike-slip across the region is consistentwith global kinematic faulting and a E-W orientedshortening in the continental models such as RM2 [Minster and Jordan, 1978] or borderland. They also included no deformation east of NUVEL-1 [DeMets et al., 1990]. In order to keep the San Andreas system south of the Garlock fault. Saucier and Humphreys [1993] and Humphreys and Weldon [1991] included deformation east of the San Copyright1994 by theAmerican Geophysical Union. Andreasfault and still found that NUVEL-1 boundary conditionsresulted in margin-normalcontraction. Papernumber 94JB00899. In this paper we considerthe deformationencountered 0148-0227/94/94JB-00899505.00 along three paths leading from the North American

19,975 19,976 HUMPHREYS AND WELDON: WESTERNU.S. DEFORMATION

a

/ GreatNorthernBasin -/- BasinandRange//

/ ColoradoPlateau

0 -,,--- • NorthAmerica

Iio

0 250 500 750 1000 I00

Figure1. Faukmap of axeascrossed by thethree paths of integrationacross the Pacific-North America boundary.(a) The deforming SW United States and adjacent regions. Heavy lines represent the most important faultsthat accommodate relative plate motion, and lighter lines represent other faults. Paired parallel lines are spreadingcenters. Bars are on downthrown sides of normal faults. Three paths are shown leading from stable NorthAmerica to thePacific plate. Arrows indicate velocity with respect to stableNorth America. Map symbolsare HF, Hurricane fault; LMF, Lake Mead fault; WFF, Wasatch Front fauk; and LVSZ, Las Vegas shearzone. (b) Faukmap of thesouthern California region. Projection is oblique Mercator projection about the RM2 Pacific-NorthAmerica pole [Minster and Jordan, 1978] (which is indistinguishablefrom the projection aboutthe NUVœL-1 Pacific-North America pole [DeMets et al., i990]). Map abbreviationsare BPF, Big Pine fault;CBF, Coronado Bank fault; ECSZ, Eastern Califomia shear zone; FCF, Furnace Creek fault; HMF, Hunter Mountainfault; IF, Imperialfault; LSF, Laguna Salada fault; N-IF, Newport-Inglewoodfault; NDVF, Northern DeathValley fault; OVF, Owens Valley fault; PVF, Panamint Valley fauk; RCF, Rose Canyon fault; SBM, San BernardinoMountains; SDTF, San Diego Trough fault; SDVF, Southern Death Valley fauk; SGF, San Gabriel fault;SGM, San Gabriel Mountains; SGP, San Gorgonio Pass; SIF, SanIsidro fault; SMB, Santa Maria Basin; SMF,San Miguel-Vallicitos fault; SSF, San Simeon fault; SYF, Santa Ynez fault; and VB, Ventura Basin. HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION 19,977

Sierra

Nevada Range

SMB--'

MojaveDesert ]

Outer ¸ Borderland

Ran

..

..

..

ß

ß . .

ß ß

ß .

...

.

ß . .

ß .. . ß ß . .. .

. .. .

. .

ß ß ß

...

ß ß : . \

ß

. .' \ . ..

ß .

ß ß \

ß ß . - \ ß .

Figure 1. (continued) interiorto thePacific plate (Figures la and2). In doing global kinematicmodels, and the precisionof our so, we obtainboth a descriptionof the transform-estimate is comparableto thatoffered by the presently accommodatingdeformation and a locallybased estimate availableglobal kinematic models RM2 and NUVEL-1. of Pacific-NorthAmerica plate motion. The estimate of We obtaina velocityfield that avoids high (--1 cm/yr) relativeplate motionthat we obtainis independentof convergencerates both in the central to eastern 19,978 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION

Figure9 •

Sierra Nevada Path

]%

Transverse RangesPath

Peninsular RangesPath

s

\

\ \

...

ß ..

Figure2. SouthernCalifornia index map. Shownare the three paths (and their subpaths) used in integrating deformationto obtainour locally based Pacific-North America relative velocity estimate. Also shown are the areasof Figures9-13. HUMPttREYS AND WELDON: WESTERN U.S. DEFORMATION 19,979

TransverseRanges and in the offshoreregion by having of years). In the absenceof suchdata we use longer- thePacific plate move more westerly, relative to North termslip rate estimates, and in lieuof reliablegeologic America,than is derivedfrom the global models, and by data,we consider geodetic data. Where geodetic data are includingregional rotations and a greaterrole for faulting used,care is takento avoidthe elasticstrain field near eastof theSan Andreas fault. majorfaults. This priority in dataselection is motivated by a desirefor maximumconsistency. Descriptionof Method An importantaspect of thisanalysis is a formal inclusion of uncertainty. To describethe uncertainty Therelative motion between two points can be found associatedwith eachsn'ucmre encountered along the by integratingvelocity changes along the lengthof an lengthof a path,we ascribeprobability functions to the arbitrarypath connecting the two points[Minster and rateand orientation chosen for thatsu'ucture. If velocity Jordan,1984]. In practice,this calculationusually datafor activestructures encountered along a pathare involvesa summationof deformationrates for known independentof one another,the probabilityfunction activefaults encountered along the chosenpath [e.g., describingtotal motion encountered along the path can be Weldonand Humphreys, 1986]. Integrationalso should determinedby convolvingthe probabilityfunctions for includevelocity gradients across rotating rigid blocks and eachof the featuresconsidered. Because independent acrossregions of continuouslydistributed deformation. If velocityestimates can be determinedfor severalpaths a completeaccounting is madeof all velocitychanges joining two points,we cancombine the individual-path encounteredalong any chosenpath, then the velocity estimatesof relativevelocity by takingthe productof calculatedat theend of thepath gives the correct relative theirend-of-path probability functions to providea better motionbetween the two points.This is trueregardless of estimateof the relativevelocity between two points. If the natureof the deformationfield awayfrom the path, datafrom each path are not completely independent from eitheron the surface of theEarth or beneathits surface. one another,the estimatedprobability function for the Our path choices(Figure 2) are guided by the productwill tendto be too narrow.Therefore we have availabilityof data,by an attemptto avoidcontroversial chosen paths joining the North America and Pacific plates regionsand regionsof exceptionallycomplex or for whichthe data are nearlyindependent. However, distributeddeformation, and by a desireto makeuse of inferencesabout the ratesand stylesof individual strike-slipfaults where possible. Strike-slip faults are structuresoften are influenced bycomparison with similar emphasizedbecause their slip vector orientations usually structuresor kinematicrelationships between structures. are betterconstrained than those associated with other Indeed,few geologicdata are truly independent.For deformationalstyles. Furthermore, because the zoneof example,the late Cenozoicrotation of the Colorado deformation between the Pacific and North America Plateau is commonly inferred from the southward platesis principallyone of simpleshear, strike-slip faults increaseof activity acrossthe Rio GrandeRift, and thus best typify the regional deformational style. Where different paths that cross the Colorado Plateau and Rio different regions deformingby simple shear adjoin, an Grande Rift are not independent. Because of such activationof additionalslip systems,required by the von interdependence,systematic biases may enter into the Mises criterion[e.g., Hirth and Lothe, 1982], resultsin a analysis. We attempt to minimize this bias by chosing region of more complex deformation. We attempt to conservative(i.e., relatively large) uncertaintyestimates avoid complexly deforming regions; however, broadly and by using paths that cross su'ucturesof different distributeddeformation and the rotation of large blocks orientationsand, often, completelydifferent styles. The are explicitly included in our integration, where fact that each of our three paths yields a similar result, encountered. despite crossingthe San Andreas fault zone where its Rotationsare especiallydifficult to recognizebecause trend, rate, and associateddeformation differ in both they are not easily associated with discrete fault magnitudeand style attests to the robusmessof the boundaries;rather, their velocitycontributions occur over techniqueand the overallquality of the data. the portionsof paths that are locatedon the rotating To describe uncertaintiesassociated with tectonic block. To recognizeblock rotation,we must have either features encounteredalong a path, we create asym- high quality palmmagneticdata that indicate regionally metrical triangular-shapedprobability functionscentered consistentyoung rotation or an understandingof the on the best rate and orientationestimates and decreasing kinematicsin a regiongreater than that addressedalong a to zero probability at chosen limits. We choose this path. Becaum rotations are difficult to recognize (or distributionbecause it is a simple representationof the dismiss)and are difficult to characterizeuniquely, the data as reported by the field researchers,i.e., a most uncertaintiesintroduced by their possiblepresence must likely estimate rate and "absolute"uncertainty limits. be handledcarefully. Here we argue their presence(or Other center-weightedprobability distributions were tried absence) mainly on local kinematic grounds and use (such as Gaussianand trapezoidal)and producedvirtually palcomagneticdata to help quantifyvates where suchdata identical results. While rate information typically is are available and consistent. We deduce that derived directly from the work of others, orientation kinematicallyimportant rotations occur in severalregions, informationusually requires an interpretationof the style and that they are not importantin otherregions. of deformation. The preferredorientation of the velocity To avoid complicationsassociated with changesin vector associatedwith strike-slipfaults is taken to be the kinematicstyle on time males longer than hundredsof fault strike at the point of crossing. To allow for the thousandsof years and as short as the seismiccycle, possibility that the fault trend at this site may not which is hundredsto thousandsof years,data are chosen representthe orientationof the fault zone as a whole, or to emphasizelate Quaternaryslip rates(tens of thousands that the fault trend is oblique to the relative motion 19,980 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION vector, we allow a range of possibleorientation defined problem arises when the integrationpath traversesa by the rangeof fault trendalong its lengthnear the path. rotatingblock whose Euler pole is poorlyconstrained. In The maximum excursions in fault trend are considered particular,when path lengthsare longand the Euler pole very unlikely orientationsfor slip acrossthe entirezone, positionis near and poorlyconstrained, significant errors so the probability function tapers to zero at these may result(i.e., if the velocityis knownat onepoint on a orientations.Often we absorbinto thisrange of possible block, but the Euler pole position is mislocated,the orientationsminor amountsof fault-normaldeformation, velocityerror accumulated in integratingto anotherpoint either transtensiveor transpressive.For instance,an on the block is proportionalto both the distancebetween assignmentin orientaftonuncertainty of + 15øon a strike- the two points and the distanceerror in pole locations, slip fault allows for fault-normal deformationat 25% of and it is inverselyproportional to both the actual distance the strike-sliprate. Wherea pathcrosses an oblique-slip and the assumeddistance to the pole position). Because fault, an assignmentof orientationis more difficult and our pathstraverse several blocks whose pole positionsare relies on auxiliary informationsuch as fold orientations, poorly constrained,we use conservativevelocity uncer- earthquake mechanisms,geodetic data, or theoretical taintiesthat includepole positionuncertainties. resultsrelating deformationstyle to obliquekinematics. Generally,we assignan orientationuncertainty of +30ø for dip-slipfaults in the absenceof more detaileddata. Resultsof Path Integration We have selectedthree paths, shown in Figure2. The central of thesepaths, the TransverseRanges path, is A summaryof the rates,orientations, and uncertainties used as a reference. The other two paths are the usedto ascribedeformation velocities to the activefaults PeninsularRanges and SierraNevada paths. Table 1 lists anddeformation zones are presentedin the appendix. the preferred value and limits of the rate and orientation Thesevelocities are discussedin the followingthree functions that we have chosen to describe the paragraphsand are summarizedin Table 1 in the order deformation associated with each feature encountered theyare encountered by traveling west along each of the along each path. A more detailed account of the three paths shownin Figures la and 2. The vector structures encountered by the paths and pertinent contributionsof the activefeatures encountered along referencesare given in the appendix. eachpath are shownin Figures3-7. The velocity deviationspresented in this section represent the velocity We comment on a potential bias in the end-of-path rangeat 90% certainty.The stylesof deformationin the velocity estimate that may result from a use of regionsrepresented by these paths and the resulting path systematicallybiased uncertaintyestimates. A choice of productsare discussed in the following sections. uncertainty bounds may appear, at the time of their introduction, to be simply an expressionof caution. However, if one is systematically"cautious" in favoringa Peninsular Ranges Path given directionor rate, the end-of-pathvelocity estimate The PeninsularRanges path crossesfaults that accom- will be affected. For example, by assigningwide limits modate minor rates of deformation east of California: the to fault slip rates to allow a greaterrange of uncertainty, southern Rio Grande Rift and the southern Basin and the assignederror can be asymmetricbecause negative Range(Figures l a and 3). In California,the pathcrosses velocities are not permitted (i.e., the sense of slip is the San Andreas fault zone south of the Transverse known). To the casualreader this may simply appearas Rangesand crossesother NW trendingstrike-slip faults though conservativeerror estimatesare used. However, in southernCalifornia, northern Baja, and the continental if rates greater than the assignedbest estimate are no borderland(Figures lb and 2). West of the Elsinorefault more likely than lower rates, an end-of-path velocity we considertwo subroutes(Figures 2 and 4): one with estimatewill have too great a velocity. As an indication good orientation control (Figure 4b) that continues of the net effect of using generally asymmetric directlyinto the continentalborderland and the otherwith uncertaintydistributions, we show both the simple best better rate control (Figure 4a) that crossesthe faults of vector sum and the end-of-path uncertainty centroid northernBaja California and then continuesnorth in the determined through convolution. Comparisonof these borderland to the end location of the first subroute. The values allows an evaluation of the effects of using end-of-pathresult, at 52 +6 mm/yr N48øW +2ø, is the asymmetricuncertainty distributions. best constrainedorientation estimate of the threepaths Two additionalproblems may arise when integrating shownin Figure 3. As shownin Table 2, this orientation over long paths. First, becausethe Earth is spherical, is -7 ø more westerlythan that predictedby NUVEL-1 velocity varies as the cosine of distancefrom the Euler [DeMets et al., 1990] or NUVEL-1A (DeMets et al., pole with a wavelength equal to the Earth's circum- Effect of recentrevisions to the geomagneticreversal ference(compared to varyingdirectly with distancefrom time scale on estimatesof current plate motions, the Euler pole for motionon a plane). In the regionwe submitted to Geophysical Research Letters, 1994; consider,such effects are insignificant.For example, hereinafter referred to as DeMets et al., submitted correcting for this effect will diminish the rate estimate manuscript, 1994), which have formal uncertaintiesof of a path from the Rio GrandeRift to California(Figure about +3ø (at 90% certainty). Our more westerly la) by an insignificant1.4% (0.02 mm/yr), and for orientationcan be attributeddirectly to the relatively deformation across California the correction is 0.05% westerlyorientation of most of the strike-slipfaults (also -0.02 mm/yr becauseof the higher rates there). encounteredalong this path. One meansof reducingthe These examples illustrate that the Earth can be assumed orientation discrepancybetween our model and the fiat at the scaleswe are considering.Another related NUVEL-1 and NUVEL-1A modelsis to have greater HUMP•YS AND WELDON: WES• U.S. DEFORMATION 19,981

TABLE 1. Relative Velocities

Azimuth, Featureor Region Rate, mm/yr West of North , NUVEL-1 velocity at TR path end =>> 48.8 :1:3 38.7 ø :t:3ø Net result (correctedto TR path end) => 48.2 +2 45.6 ø +2 ø

PeninsularRanges path => 52 +6 48o +2o East of the San Andreas fault => 1.1 +0.9-0.8 44 ø +45 ø Rio Grande 0.25 +0.25-0.15 90 ø :t:.30ø SouthernBasin and Range 0.5 +0.50-0.25 100ø :t:30ø California east of the San Andreas fault 1.0 +2-1 20 ø +30o-20 ø San Andreas fault 30 +7 48 ø +7ø-3 ø San Jacinto fault 12 :t:4 53 ø +5ø-10 ø Elsinorefault and nearbyfaults to the east 5 +3-5 49ø +20ø-12ø West of Elsinore => 5.3 +1.7 29 ø +7o-9 ø NorthernBaja option => 5.5 +1.9 32ø +20o-30ø Rotation of southern California blocks 4 :t:4 -83 ø +20 ø San Miguel-Vallecitosfault zone 1 :!:l 55ø +8 ø Agua Blancafault 5 +3-2 67ø +10ø-5ø San Isidro fault zone 2 +2 32 ø +15 ø SouthernCalifornia option => 4.4 :L3.0 32ø +9ø-18ø Rotation of southern California blocks 3.5 :L3.5 -132 ø +10 ø RoseCanyon fault zone 1 +2.0-0.5 37ø +7 ø Coronado Bank fault zone 1 +2.0-0.5 33 ø +7 ø San Diego Troughfault 1 +3.0-0.5 32ø +6ø San Clemente fault 2 +5.0-1.5 46 ø :kSø

TransverseRange path => 46 +5 50ø +7ø East of California => 0.6 :[-0.5 81 ø :t:.35ø Rio Grande 0.14 +0.06-0.04 90 ø :t:.30ø Colorado Plateau rotation 0.1 :L-0.1 0 ø :130 ø Basin and Range 0.5 +0.5 90ø :t:.30ø Mojave faults 8 +3-4 15ø +25 ø San Andreas fault 36 +8 64 o +2o Rotation of San Gabriel block ...... +11 ø +4 ø Contractionin westernTransverse Ranges 11 +10-3 -10ø +20ø Left lateral in TransverseRanges 2 :t:2 -85ø +10ø-20 Rotationof westernTransverse Ranges 2 :t:2 95ø :t:.30ø

SierraNevada path => 48 +4 41ø +4o Southern Sierra Nevada => 8.6 +2.0 51 ø :[-9ø East of California => 2.1 +2.1-1.1 86 ø +22 ø Rio Grande 0.05 +0.05 90 ø :130 ø Colorado Plateau rotation 0.1 :L-O.1 0 ø :k30 ø Great Basin east of California 2.0 +3.0-1.5 90 ø :130 ø California border to Sierra Nevada => 8.5 +1.5 57 ø +8 ø Great Basin option => 8.2 +1.8 57ø +7ø Death Valley system 1 +3-1 47ø +8 ø Hunter Mtn system 3 +1 57ø +6 ø OwensValley system 1.5 +1.0-0.5 17ø +20ø-13ø Mojave/Garlockoption => 9.7 +2.5 62ø +19ø SouthernNevada fault system 1 +1 -65ø +25 ø Mojave faults (from TR path) 8 +3-4 15ø +25 ø Gatlock fault 5 +6-2 125 ø +30-8 ø Sierra Nevada block rotation 1.0 +1.0-0.5 180 ø :135 ø Contraction near and west of San Andreas 2 +5-1.5 -45 ø +5ø-15 ø San Andreas fault 34 +3 41 ø +4 ø Right lateral westof San Andreas 6 +6-4 40ø +15ø

TR is TransverseRanges. Arrows indicate velocities calculatedfrom subsequentvalues (double-headedarrow for NUVEL). Uncertaintiesfor valueswith arrowsare at 90% certainty. rates of slip on borderlandfaults; however, as can be rate is an increase the counterclockwise rotation rate of seen in Figure 4, increasedborderland activity increases the southern California blocks. The kinematic the rate discrepancy. A kinematic modification that consequencesof such a modificationon deformationin reducessomewhat the discrepancyin both orientationand the TransverseRanges are discussedbelow. 19,982 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION

N N (a) (b) RM2

NUVEL westofElsinore F: Elsinøre( E San Jacinto

southern San Andreas

east of San Andreas

Peninsular Ranges Path Figure 3. Vectorplots for the PeninsularRanges path. (a) The velocitydata for the slxucturesconsidered along the PeninsularRanges path, shown in Figures la and 2. Enclosedareas about the ends of velocity vectors representthe 30, 60, and90% confidencelevels in the velocity,and vector tips lie at the mostprobable velocity. The north arrow represents50 mm/yr, for scale. See Figure4 for "westof Elsinore"estimate. (b) The end-of- path vectorresulting from the convolutionof uncertaintyregions shown in Figure3a, representingour estimate of Pacific-NorthAmerica relative velocity derivedfrom geologicinformation found along this path. Shownfor comparisonare the Pacific-NorthAmerica velocity estimates predicted by RM2 [Minsterand Jordan, 1978] and NUVEL-1 [DeMets eta/., 1990]. Although our rate estimateis indistinguishablefrom that of NUVEL-1, the velocityis 7ø more counterclockwisein orientation.The plus symbolsindicate the resultof simplevector sums.

Transverse Ranges Path associated with deformation kinematics in the western Deformationeast of California, includingspreading in TransverseRanges result in end-of-pathuncertainties that the central Rio Grande Rift, rotation of the Colorado are greaterthan thoseobtained for the other two paths. Plateau, and Great Basin faulting in southernNevada, Sierra Nevada Path contributesonly minor rates (Figure 5; see appendixfor details). In California, three major componentscon- The Sierra Nevada path (Figures2 and 6) crossesthe tribute to the rate estimate: the "eastern California shear northern Rio Grande Rift, Colorado Plateau, the active zone" in the central Mojave Desert [Dokka and Travis, Great Basin, the San Andreas fault in central California, 1990], the San Andreas fault zone in the Transverse regionsof contractioneast and west of the San Andreas, Ranges, and convergence in the western Transverse and right-lateral and contractile deformation near the Rangesfold and thrustbelt. As shownin Figures 1 and California coast. We also consider an alternate eastern 5, deformation in the Mojave and in the western portionto this path (Figures2 and 7), which passessouth Transverse Ranges contributes velocities to the path of the actively deforming Great Basin, crosses the integral that are more northerly than any Pacific-North Mojave Desert, and joins the main path on the Sierra America velocity estimate, whereas the San Andreas Nevadablock by crossingthe Garlockfault. Consistency fault, where crossed,is more westerlythan any Pacific- of southern Sierra Nevada motion derived from these two North America velocity estimate. Because the San pathsrequires either a relativelyinactive western Gatlock Andreas fault is the dominant structure encountered, the fault (i.e., southof the Sierra Nevada) or the occurrence end-of-path velocity estimate is oriented with a more of deformation that we have not included in our cal- westerly trend relative to the globally based estimates. culations. Our end-of-pathvelocity estimate of 48 +4 Accumulateddeformation along this path is calculatedto mm/yr N41øW +4 ø (Figure 6) is more westerly than be 46 +5 mm/yr N50øW +7 ø comparedto a NUVEL-1 NUVEL-1 and NUVEL-1A by-5 ø and is 1 mm/yr predictedvelocity of 48.8 mm/yr N39øW (Figure 5). Our smaller in magnitude. The very long baseline inter- locally basedrate is indistinguishablefrom the NUVEL-1 ferometry (VLBI) stationVAND, which is thoughtto be and NUVEL-1A rates, whereas the orientation is different essentiallyon the Pacific plate, lies near the end of our at a confidencelevel of greater than 95%. Uncertainties path. Its motion relative to North America is estimated HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION 19,983

N N N (a) t.•, (c) (b)

N /

San Isidro San CI San Diego Trough

Agua Bianca Vallecitos-San Miguel oronado Bank Rose Canyon-N. I.

block rotations block rotations Northern Baja Option Southern California Option West of Elsinore Fault, Peninsular Ranges Path

Figure4. Vectorplots for thewestern portion of thePeninsular Ranges path. Plotis asdescribed in Figure3, exceptscale is expanded(north arrow represents 10 nun/yr). (a) The velocitydata and net convolutionresults for the northernBaja option. (b) The velocitydata and net convolutionresults for the southernCalifornia option. N.I. refersto theNewport-Inglewood fault. (c) The productof the estimatesfrom Figures 4a and4b. This is the "westof Elsinore"estimate in Figure3. The southernCalifornia option provides a well-constrained estimateof velocityorientation, whereas the northern Baja option provides better rate control.

at 48 +1 mm/yr N39øW +1 ø [Ward, 1990] and 48 +1 Kinematics of Pacific-North America N42øW +1 ø [Gordon et al., 1993]. Interaction

Combined-Path Pacific-North America Velocity In additionto providingan estimateof Pacific-North America relative motion, the deformation fields discussed Estimate for eachof the pathscan be usedto infer the deformation To obtain our best Pacific-North America velocity we of the southwestUnited States. As noted by Atwater take the product of our three end-of-pathprobability [1970], this deformation is broadly distributed and functionsfor these three independentvelocity estimates. heterogeneousin style. Figures3, 5, and 7 illustratethat Before doing this, however,we rotate the SierraNevada deformation in the southwest United States east of and the PeninsularRanges velocity estimatesby an California is accommodatedlargely by west directed amountneeded to continuethese paths to the Transverse extensionon normal faults and is slow comparedto the Rangespath end. We usethe NUVEL-1 poleposition in deformation in California. Active systems east of makingthis correction(see Table 2). This correctionis California include the Rio Grande Rift, frontal faults of small, and any error introducedby potentialmislocation the Wasatch Mountains, and the western front of the within the NUVEL-1 pole position,including moving the Colorado Plateau. Differential motion between the areas poleto accommodateour velocityestimate, is very small. north and south of southern Nevada are accommodated The productof the three end-of-pathuncertainty dis- by minor rates of left-lateral slxike-slip activity in tributionsis a vector 48 +2 mm/yr N46øW +2 ø (Figure southernmostNevada. As California is approachedfrom 8). This is -7 ø more westerly than NUVEL-1 and the east, increasingamounts of transform-relatedright- NUVEL-1A. lateral shear are encountered. In the area of southern and 19,984 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION

N N (b) (a) RM2

NUVEL

wTR contraction wTR rotation

• eralinwTR SanAndreas slip and • SanGabrielrotation Mojav•

east of California Transverse Ranges Path Figure5. Vectorplots for the TransverseRanges path. Plot is as describedin Figure3. The estimated Pacific-NorthAmerica velocity is about11 ø counterclockwiseof the globallybased estimates RM2 and NUVEL-1. This resultsfrom the nearlywesterly velocity contribution of the SanAndreas fault for thispath, whichis not completelycompensated for by thenortherly orientations of contraction in thewestern Transverse Ranges(wTR in thefigure) and shear across the Mojave faults of theeastern California shear zone.

(a) N (b) N RM2

NUVEL Hosgri•• • SanAndr •/••wCA shortening •

southernSierraNevada• Sierra Nevada rotation

Sierra Nevada Path

Figure6. Vectorplots for the Sierra Nevada path. Plot is asdescribed in Figure 3. EstimatedPacific-North Americavelocity is about5ø counterclockwiseof the globally based estimates RM2 and NUVEL-1, owing to a slightlycounterclockwise orientation of the Sa•i Andreas fault in centralCalifornia and to SanAndreas-normal extensionrates east of the SierraNevada that exceed contraction rates in westernCalifornia (wCA in the figure). SeeFigure 7 for "southern.Sierra Nevada" velocity estimate. HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION 19,985

N N (a) /\ (c) /\ (b) OVRO •- •Dixon et al ,..

E •

N Product• /\

E E

o OwensValley• southern Nevada. / .k...%•...DeathValley

Huntereast Mountain'•••of California%.._ east of southern Nevada Mojave Option Great Basin Option Southern Sierra Nevada, Sierra Nevada Path Figure7. Vectorplots for theeastern portion of theSierra Nevada path. Plotis asdescribed in Figure3, exceptscale is expanded(north arrow represents 10 mm/yr).OVRO represents the velocity of a geodeticstation in theOwens Valley, near the Sierra Nevada. (a) Thevelocity data and net convolution results for theMojave option.(b) Thevelocity data and net convolution results for thesouthern Great Basin option. (c) Theproduct of the estimatesfxom Figures 7a and7b. This is the "southernSierra Nevada" estimate in Figure6. The geodeticallybased velocity estimate for theSierra Nevada of Dixonet al. (submittedmanuscript, 1994) is shown for comparisonand discussedin the text. central California, deformationis organized into three If no significantrotation is occurring,activity west of distinctive areas; each area expressesa characteristic the San Jacinto fault moves the western Peninsular patternof deformationstyle, and each is representedby Ranges-1 cm/yr to the SE with respectto the Pacific one of our paths. The primary goal of this sectionis to plate; slip on the San Jacintofault addsa similar amount extendthe kinematicdescriptions beyond the vicinity of of velocityto the easternPeninsular Ranges and northern the three paths so as to include the southernCalifornia SaltonTrough. As shownin Figurelb, the generaltrend region, thereby permitting constructionof kinematic of these faults, like the San Andreas fault, is -10 ø modelsfor local regionsof interest. counterclockwiseof the NUVEL-1 trend. This produces a Pacific-NorthAmerica relative velocityestimate that is Southof the TransverseRanges itself counterclockwise of the NUVEL-1 orientation. Contributingto this more westerly motion is the -1 Southof the TransverseRanges, NW trending,right- mm/yr of extensionin the southernBasin and Range. lateral, strike-slipfaults lying to the west of the San The two VLBI sites lying unambiguouslywithin these Andreasfault accountfor nearlyall of the platevelocity blocksthat have a well-estimatedvelocity are locatedat that is not carried on the San Andreas,giving the MonumentPeak, about 50 km ENE of San Diego, and appearanceof a broad simple-sheartransform accom- Pinyon Flats, east of the San Jacinto fault south of the modation zone. Faults active in this zone include the San GorgonioPass (Figure lb). Their velocitiesrelative subparallelSan Jacinto,Elsinore, and southernCalifornia to North America [Ryan et al., 1993] are: Monument continental borderland faults. However, if the crust Peak, 41.5 +1 mm/yr N52øW +2ø, and PinyonFlats, 25 through which these faults trend is rotating +1 mm/yr N47øW +2 ø. These velocities are counterclockwise, then the combination of rotation and approximately 10ø and 5ø counterclockwise of the shearcreates a componentof pure-sheardeformation (as NUVEL-1 trend (respectively),and consistentwith our discussedin the lastparagraph of this subsection). estimateof motionat thesepoints. 19,986 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION

TABLE 2. Predicted Pacific-North America Relative Motion at Selected Points

End-of-Path Location RM2 NUVEL-1 Path Latitude Longitude Rate, mm/yr Orientation Rate, mm/yr Orientation Sierra Nevada 35.7 ø -121.6 ø 56.9 N35.8øW 48.8 N36.1øW TransverseRanges 33.7ø - 119.7ø 56.8 N38.1øW 48.8 N38.7øW PeninsularRanges 32.4ø - 118.2ø 56.6 N39.7øW 48.6 N40.6øW

Tending to offset the orientationdiscrepancy between into the Mojave block, where the San Andreasfault is NUVEL-1 and our model for California south of the misalignedwith respectto the transformdirection. In our TransverseRanges is the velocity contributionarising model, this pod moves around the curved San Andreas from a counterclockwise rotation of this crust about the fault and is accommodatedat its southernend with slip arcuate southernCalifornia San Andreasfault. Although on the Agua Blanca fault, which separatesthe rotating the style of this rotation is fairly well defined by the pod from the nonrotatingBaja peninsulato the south. geometry of the bounding faults, the rate is poorly The pod of crustis also shearedby severalnorthwesterly determined. If all of southern California SW of the San trending right-lateral strike-slip faults within it. The Andreasfault rotatesat the full rate implied by slip on combinationof right-lateralshear in the pod and rotation the San Andreas fault, as suggestedby Weldon and producesa componentof pure shearthat allows regional Humphreys [1986], rotation occurs at -4ø/m.y. As shorteningin the north-southdirection and lengtheningin discussedin the appendix,we have chosena rotationrate the east-westdirection. Effectively, this contributesto for the blocks southof the TransverseRanges that is half moving most of southernCalifornia in a more westerly of the full rate (see Figure 4). direction so as to avoid the large left step of the San The arcuate Agua Blanca fault and continental Andreasfault throughthe TransverseRanges. borderlandfaults appear to play a role in the rotationof SW California that is the mirror image of the arcuateSan Andreas fault in southern California. This is illustrated Transverse Ranges in Figure 9, which is in a Mojave frame of reference. Deformation within the TransverseRanges region is The almond-shapedpod of crust lying between the San typified by a combinationof strike-slipfaulting (both Andreas fault and the Agua Bianca-inner borderland right-lateral and left-lateral) and convergence. Both faults rotatescounterclockwise so as to avoid converging stylesare accompaniedby block rotations. This activity is largely a consequenceof kinematicsimposed by the large left step in the San Andreasfault throughthis N region (Figure lb). The high easternSan BernardinoMountains lie north RM2 of a relatively small left step in the San Andreasfault through San Gorgonio Pass and are associated with thrustingthat has uplifted the San Gorgoniomassif and NUVEL northernplatform of the San BernardinoMountains. This restrainingSan Andreasfault geometryis local to the San Gorgonio Pass region; the western San Bernardino Mountainshave been left standinghigh from an earlier western U.S. uplift event (the spectacularsouthwest escarpment is estimate largely the result of young right-lateral San Andreas (this paper) faultingjuxtaposing high and low regions[Meisling and Weldon, 1989]). The San Gabriel block moves so as to avoid most of the convergenceapparently required in this regionwhere the San Andreasfault trends in a stronglyrestraining direction (i.e., very counterclockwisefrom its trend elsewhere). However, the broadly arcuateSan Andreas fault slips in a strike-slipfashion through this region, which is accomplishedby a counterclockwiserotation andrelatively great westerly component of velocityof the San Gabriel block (Figure 10a). The crust southof the San Gabriel block is also thoughtto be involved in the rotation (Figure 9), but at a lower rotation rate than the Pacific-North America San Gabriel block itself. Evidence supportingthe Figure 8. Our final Pacific-NorthAmerica velocity estimate, rotationof California southof the TransverseRanges comparedto Pacific-NorthAmerica velocity estimatespredicted include the curved south San Andreas fault [Weldon and by RM2 [Minster and Jordan, 1978] and NUVEL-1 [DeMets et Humphreys,1986], the curvedAgua Biancafault (Figure al., 1990]. Plot is as describedin Figure 3. EstimatedPacific- 9 and preceding section),paleomagnefic results in the North Americavelocity is about7 ø counterclockwiseof the glo- San Gabriel Mountains (Appendix), and geodetic bally basedestimates of RM2 and NUVEL-1. information(p•g section). Evidencesupporting a HUMPttREYS AND WELDON: WESTERN U.S. DEFORMATION 19,987

the anomalouslytrending San Andreasfault to the north result in the San Gabriel block moving more rapidly to the west than any of the surroundingcountry, like a watermelonsee• being displacedto the west betweenthe Mojave and the blocks convergingfrom the south. A consequenceof this mechanismis that the San Andreas fault north of the San Gabriel block slips at a rate somewhat greater than one would infer by simply accountingfor slip rates on the San Andreas fault and associatedfaults away from this sectionof fault. This is supportedby the available data [Weldonet al., 1993]. The combined set of observations cited in the preceding paragraph lead us to conclude that the San Gabrielblock rotatescounterclockwise as it slidesalong the arcuate San Andreas fault, while the crest south of the San Gabriel block rotates more slowly. This dif- ferential rotation between the San Gabriel block and crest to the south, which was not included by Weldon and Humphreys [1986], is thought to occur about a pole located near the eastern end of the San Gabriel block because little deformation occurs between the two blocks there. Deformation in the westernTransverse Ranges includes both thrusting and strike-slip faulting, as well as kinematicallyimportant rotations. Luyendyket al. [1980] and Jackson and Molnar [1990] describe a deformation

...... model for the western Transverse Ranges in which thrustingand left-lateral slip occur on east-westtrending faults that separateelongate, clockwise rotating blocks (similar in style to that shown in Figure 11). Alternatively, left-lateral slip on these faults can result from the relatively westerly motion of the western Transverse Ranges with respect to the continental After Movement borderland,independent of block rotation. In this latter role, these faults are extensions of the left-lateral faults Figure 9. Simplifiedmodel illustratingthe simultaneousright- active at the southern front of the San Gabriel block that lateral shearing and rotation of southern Califomia. With allow this block to move west comparedto crust to the respectto a Mojave reference,this pod rotatescounterclockwise south. Figure 12 showsboth processesbeing active in about the curved San Andreas and Agua Blanca faults and thus western TransverseRanges, with rotational mechanisms avoidsconverging into the Mojave block (comparewith Figure being more active in the westernmostTransverse Ranges lb). The combinationof rotation and right-lateral shear within and nonrotational shear being dominant near the San the pod producesa pure-shearstyle of deformationwith N-S Gabriel block. shorteningand E-W lengtheningthat accompaniesthe overall simple shearoccurring across the plate margin. In the Ventura Basin region in particular, several kinematic processesacting togetherto producehigh rates of shortening(Figure 12). The crest north and east of rotation of southern California crust at a rate slower that Ventura Basin rotates counterclockwise due to the motion of the San Gabriel block is found in the left-lateral of the San Gabriel block, while the block south of the faulting at the southernmargin of the San Gabriel block. Ventura Basin rotates clockwise [Terres, 1984; Terres If the only importantdeformation of the California crest and Luyendyk, 1985; Hornafius, 1985] about a hinge south of the San Andreasfault arose from slip on the located near where this block adjoins the San Gabriel trans-Peninsulafaults (such as the San Jacinto, Elsinore, block. This activity can be likened to the closing of a and borderland faults), then thrusting would occur at folding fan about a pivot located in the vicinity of Los increasingrates to the west along the southernmargin of Angeles, with the Ventura Basin and adjoining regions the San Gabriel block, as would a componentof right- lying between the closing edges of the fan. Also lateral faulting there (see Figure 10b). Suchthrusting is contributingto westernTransverse Ranges convergence is observed,but the associatedstrike-slip faulting is clearly the NNW motion of the continental borderland. This left-lateral. If the San Gabriel block was moving as componentof convergenceresults from slip on several shown in Figure 10a and the crust to the south was trans-Peninsularfaults (Figure lb), includingthe Elsinore converging towards the Mojave without rotating, as fault and faults kinematicallyrelated to the Agua Blanca shownin Figure 10b, then strike-slipfaulting along the fault, the Rose Canyon and San Clemente faults, which southernmargin of the San Gabriel block would be left- allows the borderland to translate toward the NNW and lateral. The combination of left-lateral slip on faults converge onto the western Transverse Ranges [Hum- southof the San Gabriel block and right-lateralslip on phreysand Weldon, 1991]. 19,988 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION

• MojaveBlock

San ...... •.•..B.l..•.k...... i•.•,• ......

Ctabrt•'lFault BeforeMovement

BeforeMovement

AfterMovement After Movement Figure10. Schematicrepresentations of thrusting mechanisms in the central Transverse Ranges. (a) Illustration showingthe San Gabriel block rotating as is movesaround the curved San Andreas fault, which is thesouthern boundaryof the fixedMojave block. Left-lateralfaulting occurs near the southern margin of theSan Gabriel block. Thrustingaccompanies left-lateral faulting to the west. (b) Illustrationshowing the contributionto TransverseRanges thrusting provided by right-lateralslip on NNW trendingfaults lying south of theTransverse Ranges.The rate of convergenceincreases to thewest. The oblique direction of convergencealso results in a componentright-lateral shear along the rangefront. Thissense of strike-slipmotion is differentfrom that predictedin Figure10a, and is alsodifferent from that observed in geologicoffset and earthquake slip vectors alongthe southernmargin of the SanGabriel block.

N N

? ?

BeforeMovement AfterMovement Figure11. Modelwhere the rotation of elongatedblocks create thrusting and left-lateral shear in thewestern TransverseRanges. The model is similarto thatsuggested by Luyendyket al. [1980]and Jackson and Moltmr [1990],except we showthese rotating blocks distributed in a morefan-like pattern. Also unlike Luyendyk et al. andJackson and Molnar, we do not thinkthat this mechaxfism applies to regionseast of the VenturaBasin, and clearlyclaes not hold for the SanGabriel block. HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION' 19,989

TransverseRanges rotate into the relatively narrow zone of the CoastRanges. These two mechanisms trade off with one other, and x, although both mechanisms appear to be active, the relative importance of crustal thickening and NNW acceleration is not well constrained. Both of these mechanismshave important consequences.In the first mechanism, crustal thickening of the southern Coast Ranges occurs at a relatively great rate as this crest passesthe Big Bend of the San Andreasfault. Therefore

VAND a significantfraction of the Coast Range shorteningthat is generallyinferred to have occurrednormal. to the San Andreasfault north of the Big Bend can be attributedto this interactioninstead of ongoingrapid convergencein the central California Coast Ranges. In the second mechanism, acceleration of the crest between the San Andreasand Hosgri faults addsto the slip rate of the San Andreas fault and subtractsfrom the slip rate of the Hosgri fault. The faults of the western Transverse Ranges form a left step in a right-lateral zone of faults found near the southern and central California coast and hencecan be consideredas a transferfault [e.g., Weldon and Humphreys, 1986]. However, the participating right-lateral faults in central California (principally the Hosgri system) are less active than either the par- Figure 12. Kinematicmodel representingcurrent deformation ticipating right-lateral faults in southern California of the westernand cenl:ralTransverse Ranges, using components 0xxwe•n the Elsinore and borderlandfaults, inclusive)or from Figures 10 and 12. The combinedeffect of counterclock- the transfer faults in the western TransverseRanges. wise rotation of the San Gabriel block and clockwise rotation of Crustal accelerationthat adds slip to the San Andreas the block south of the Ventura Basin results in rapid conver- fault and subtractsslip from the Hosgri fault can help genceacross the VenturaBasin, with thrustingon both sidesof decreasethis apparentdiscrepancy. the basin. West of the Ventura Basin in our model, clockwise rotation of blocks results in thrustingand left-lateral faulting. The lines labeled a-a' and b-b' are used to illustrate the net North of the Transverse Ranges shorteningof crust in an E-W direction that occurs as crust movesnorthward past the TransverseRanges. Thrustingin the North of the TransverseRanges, shear systemsnearly westernmostTransverse Range results in crustal thickening, parallel to the San Andreasfault are active both east and whereas clockwise rotation contributes toward an acceleration in west of the San Andreas,giving rise to three transform- velocity and crustallengthening (both in a San Andreas-parallel accommodatingshear zones. The easternsystem (ECSZ direction). Very long baseline interferometrystations VAND of Figure lb) splaysfrom the San Andreasfault in the and IPL are shown for reference. region of the easternTransverse Ranges, trends through the Mojave [Sauber, 1989], and continueson the Death Valley, PanamintValley, and OwensValley fault systems The crest of the central and westernTransverse Ranges [Dokka and Travis, 1990; Powell and Weldon, 1992], acquiresa relatively westerlyvelocity as a result of its where it is associated with crustal extension in the interactionwith the relatively west trendingSan Andreas westernGreat Basin. (Additional Great Basin extension, fault in the TransverseRanges. As this crustmoves into orientedapproximately E-W, occurseast of the Walker the westernmostTransverse Ranges from the southeast,it Lane belt.) In contrast,the San Andreas fault and the rotates clockwise,thereby loosing some of its westerly westernfault system(including the Hosgri fault system) velocity (Figu• 12). In doing so, this crust acquiresa are associated with crustal shortening [Namson and velocity nearly parallel to the San Andreasfault north of Davis, 1988;Mount and Suppe,1987; Feigl et al., 1990]. the big bend. This resultsis a general"flow" of the crest Both fight- and left-lateralfaults accommodatedefor- immediately south of the San Andreas fault that is mation within the eastern California shear zone (Figure parallel to the trace of this (sinuous)fault. The funneling 13), which allows the western Mojave to move NNW of the crest west of the southern California San Andreas with respectto North America at -1 cm/yr [Gordon et fault into the zone west of the central California San al., 1993]. The right-lateralfaults trend northwesterly, Andreasfault (line b-b' to line a-a' in Figure 12) can be and if this orientation alone is used to infer the motion of accomplished by either (1) a loss of surface area the westernMojave, the accommodationzone would pull associatedwith thrustingand crustal thickeningor (2) apart. This deformationzone is complex, and local San Andreas-normalshortening and San Andreas-parallel regionsof extensionand contractionoccur where block lengtheningof the crest as a resultof strike slip faulting motions cannot be accommodatedsimply by strike-slip and block rotation. This second mechanism accelerates faults. However, becausethis zone in the Mojave is the crest of the CoastRanges toward the NNW (parallel neitherdilating nor contractingsignificantly [Dokka and to the San Andreasfault) as crustalblocks of the western Travis, 1990], western Mojave block motion must 19,990 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION

Before Movement

After Movement

Figure 13. Kinematicmodel for the motionof the Mojaveblock relative to easternCalifornia. The Mojave block translatesNNW, roughlyparallel to the zoneof left- andright-lateral faults that accommodateits motion (thesenumerous faults constitutethe easternCalifornia shear zone of Figure lb). In this model,rotation of the westtrending blocks occurs in conjunctionwith left-lateralslip on thefaults that define the blocks (including the easternOarlock fau10. The motionof the Mojave block is more northerlythan the right-lateralfaults that contributeto its motion,which must be accommodatedeither by deformationnormal to the right-lateralfaults or by smallamounts of localclockwise block rotation. Slip on thewestern Oarlock fault is not associatedwith the easternCalifornia shear zone; rather, the left-lateralwestern Oarlock fault separatesthe northerlymoving Mojave block from the northwesterlymoving Sierra Nevada block. approximatelyparallel the overall NNW trend of this rotationof elongatedblocks separatedby west trending shearzone, and not the trend of individualfaults. In our left-lateral faults [Luyendyket al., 1985] permits the model, the NNW motion of the Mojave block is translationof the Mojave block (Figureslb and 13). The accommodatedby a slight clockwiserotation of the eastem Garlock fault is included with these faults [Dokka blocks in the central Mojave portion of the shearzone and Travis, 1990]. that are boundedby the right-lateralfaults. In regions The nearly northerlymotion of the Mojave block has south of the eastern Garlock fault and north and southeast two importantconsequences. First, convergencerates in of the easternTransverse Ranges, a significantclockwise the TransverseRanges south of the Mojave are reduced HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION 19,991 as the Mojaveblock moves away from the convergentis-7 ø morewesterly than the globallybased estimates zone (at about 6-8 mm/yr). Second,the northwesterly (Table 1). motion of the Sierra Nevada block and the nearly Severalpossible explanations for this discrepancyare northerlymotion of the Mojave block resultin the listed below, and somecombination of thesefactors observedleft-lateral slip on the Gatlockfault whereit probablyexplains the differencein orientation.(1)There separatesthese two blocks;if the Mojaveblock were is a smallchance (less than 1%)that the models are con- movingnorthwesterly also, then no strike-slipmotion sistentwithin their probability distributions as described. wouldbe resolvedon the northeasterlytrending Garlock In orderfor the modelsto be consistent,velocities for fault there. individual features in the western United States must be North of the Garlockfault (Figure lb) the style of chosentoward the northeastlimits of their respective deformationchanges; significant north trendingnormal probabilityfields, whereas velocities for individualfea- faultsjoined by NW trendingstrike-slip faults define a turesin the NUVEL-1 modelmust be chosenso as to set of pull-apartbasins that accommodateboth fight- resultin a Pacificplate motion that is biasedtoward a lateral shearand extension[Burchfiel et al., 1987; southwestorientation. If the two modelsare consistent, Stewart,1988]. Becausethe trend of thisbelt aligns with then the kinematicsolution for the southwestUnited theMojave zone (ECSZ of Figurelb) yetis dilating,the Statesis very well constrainedby the productof these SierraNevada must be movingaway from the average two velocityprobability distributions. (2) The NUVEL-1 trend of this belt, i.e., moving in a directionmore velocityestimate for Pacific-NorthAmerica motion does westerlythan -N15øW. We inferthat the orientationof not representthe southwesternUnited Statesproperly fight-lateralfaults (--N50øW)represents the orientationof becausedeformation of the Pacific plate occurssome- relative motion acrossthe zone as a whole, resultingin a where between the southwest United States and the creationof surfacearea. T. H. Dixonet al. (Constraintsregions where kinematicconstraints on Pacificplate on present-dayBasin and Rangedeformation from space motionexist. (3) Thereis significantdeformation in the geodesy,submitted to Tectonics,1994; hereinafter southwestUnited States that we havenot includedin our referredto as Dixonet al., submittedmanuscript, 1994) model. The possibilityof unaccountedfor deformation discussthe occurrenceof additional Great Basin resultsin end-of-patherror estimates that are too small. extension,directed roughly eastward,occurring at However,we havebeen careful to includeall significant significantrates east of the WalkerLane belt. It is this deformationof whichwe are aware,and we considerit extension,combined with a near absenceof southern unlikely that significantdeformation has gone unrec- Basin and Rangeextension, that resultsin a Sierra ognized.The factthat our rate estimates are very similar Nevadamotion that is more westerlythan the Mojave to the NUVEL-1 rate supportsthe contentionthat block. The left-lateralwestern Garlock fault accom- importantdeformation has not beenomitted, especially modatesthis difference in relative motion in a manner becausemost deformationcontributes a componentthat similarto that proposedby Davis and Burchfiel[1973]. adds to the Pacific-NorthAmerica rate. The velocity In our model,eastern Garlock fault activityis dueboth to fields that we would most likely fail to recognizeare its role in separatingclockwise rotating blocks in the thoseassociated with rotations.(4) Significantdefor- Mojaveand to its rolein accommodatinga component of marionoccurs in the regionwest of ourmodel (such as Gatlock-parallelextension to its north; the western shorteningnear the continental rise), and hence we have Garlockfault is activeonly in the latterrole, and is not integrateddeformation all the way to the Pacific presumablyless active. The northwardcontinuation of plate. If this were the sole causeof the discrepancy, the easternCalifornia shear zone includesthe Walker about6 mm/yrof obliquecontraction would be required Lane belt [Stewart,1988] and severaldeformation zones offshore. Such contractionseems unlikely considering trendingthrough the PacificNorthwest [Pezzopane and the structuralsimplicity of young sedimentsand the Weldon,1993]. relativeseismic quiescence of this region. Discussion NonidealTransform Deformation The Pacific-North America transform deformation field Pacific-NorthAmerica Relative Motion differsfrom its oceanic counterparts in two important Our estimateof Pacific-NorthAmerica relative motion ways. First, transform-relateddeformation is broadly (Figure8) is calculatedusing data associated with the distributed,extending across California and portions of deformationzone in the southwesternUnited Statesthat westernmostNevada. Second, significantamounts of accommodatesthe relativemotion, and it is independentdeformation involve either the creation or thedestruction of otherestimates of platemotion. In this senseit is of surfacearea. This sometimesoccurs where fault complementaryto the modelsNUVEL-1 [DeMets et al., patternsare complexand local spaceproblems exist. 1990],NUVEL-1A (DeMetset al., submittedmanuscript, However,we also find large regionsthat experiencea 1994),and RM2 [Minsterand Jordan,1978], which are gainor lossof surfacearea. Figure14 showsan interior derivedfrom plate kinematic information that is exclusive belt of dilationand a westernbelt of contraction,each of data from our studyarea. Our locally basedrate occurringwithin the zone of transformaccommodation. estimateis not significantlydifferent from that of The dilatingbelt extendsacross the entireregion studied NUVEL-1 or NUVEL-1A (andis abouta centimeterper except throughthe Mojave. The belt of contraction year slowerthan the olderRM2 rate). Thusour rate includesthe CoastRanges and the Transverse Ranges. A estimateis consistentwith the rate givenby the global southwardcontinuation of the contractionalbelt within dataset. However,our locally based orientation estimate the southern California continentalborderland is 19,992 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION

.::.:.:.:.:.. Z;:i:i:i:i:i:i:i:i.

::5::::::. i:!:i:!:i:i

...... '"!:!:i:i:i'i 'i:i:!:!:i:i'i'!:i:i:!:!:ii

s••emajør -slip fault (San Andreas fault) %. •secondary strike-slipfault

regionof contraction

..... regionof dilation '.. .::::-:.i•i • ...... ' ß :::::::::::::::::::::o,-,o . - .' :i:;:i:i:?i:i:i:i: :".:•::::::::::::' ß:-:-:.3:¾::::: ',

...... ::i:i:i:i:!:!:i:i:•

ß:-:-:-:.:-:-5:::: :•':':':':':':-:-• ::5::::::::::::x :US:U:::::::::• ?';'X':-:.:-:.X -:.;.;.:.:.:.;.2.:.• •:::::::::::::x

:¾:::::2;::2:::2::.X

Figure14. Schematicregional map emphasizing regions of positiveand negative dilation that are associated withimportant shear zones. Surface area is beingcreated in theSalton Trough region and the region of theSW GreatBasin, both shown with stipples.Surface area is beinglost in the centralCalifornia Coast Ranges, the TransverseRanges. and perhaps in thenorthern borderland, shown with ruled pattern. See Figure lb for names and locations of structural features. suggestedby the occurrenceof borderlandearthquakes NUVEL-predictedPacific-North America relative position showingthrust components[Legg, 1988; Pachecoand vector and essentiallyparallel to our estimateof this Nabelek, 1988; Haukssonand Jones, 1988] and by young vector. The transition between the dilafional and con- folds and thrusts [Mills and Fischer, 1991]. The San tractional San Andreas domains occurs near the eastern Andreas fault zone is associated with the Gulf of TransverseRanges. California and the Coast Rangesregions of positiveand Deformation in the western Great Basin and the negativedilation, respectively. In both settings,the San California CoastRanges involving positive and negative Andreas fault is oriented counterclockwise of the surface dilation (respectively) defines paired tectonic HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION 19,993 belts. The western Great Basin has active normal faults oriented nearly parallel to the Pacific-North America relative motion vector (e.g., the frontal faults of the Sierra Nevada), and the dominant strike-slip faulting I there is oriented so as to move the Sierra Nevada/Great Valley block toward the Pacific (see Figure 7b). At the same latitude, contraction occurs in the central California Coast Ranges [e.g., Mount and Suppe, 1987] that is oriented at a high angle to the Pacific-North America relative motion vector (see Figure 6). Together, these zones of deformation give rise to a transform-normal componentof motion to the Sierra Nevada/GreatValley block. This suggeststhe action of forces createdlocally that are not directly attributableto transformdeformation [e.g., Eaton, 1932]. Another region where deformationoccurs that is not related simply to transform faulting is the Transverse Ranges. Geodynamicmodeling of this region indicates that locally generatedforces, attributed to anomalously dense mantle sinking beneath the TransverseRanges, drive crust toward the Transverse Ranges [Bird and Baumgardner, 1984; Sheffelsand McNutt, 1986, 1987; Humphreysand Hager, 1990]. Becauseactual rates of shorteningacross most of the TransverseRanges are relatively slow (i.e., <5 mm/yr), approximate force SaltonTrough Basin & Range balanceappears to exist betweenconvergence-driving Figure 15. Examplesof the styles of deformationwhere exten- forces created in the mantle and divergence-drivingsional features are incorporated into the strike-slipfault strut- forces createdby the high topography[Humphreys and nares.For reference,lines parallel to thePacific-North America Hager, 1990]. In contrast,shortening rates in the western velocityestimates fxom from NUVEL-1 (sidesof figure) TransverseRanges probably exceed 10 mm/yr.This loss [DeMetset al., 1990]and our work (dashed lines) are shown. of surfacearea occurs primarily west of theBig Bend, If shearwere applied across the region of transtension,the i.e., in the westernTransverse Ranges, presumably NUVEL-1 orientation would cause compression normalto the most of the strike-slipfaults, whereasthese faults are well becausethelithosphere ofthe western Transverse Ranges aligned forsimple shear with respect tothe dashed line. Zone isrelatively weak compared tothat of the Great Valley of foldingaccommodating shortening isshown with double- (supportedby uppermantle images [Biasi and sidedarrows. See Figure lb fornames and locations ofstrue- Humphreys,1992; Humphreysand Dueker, 1994]), and turalfeatures. becausethe crust in the westernTransverse Ranges is being drawn toward a zone of mantle downwelling beneathat least part of this region [Humphreysand Conclusions Hager,1990]. Weestimate Pacific-North America relative motion by Stylesof deformation aredistinctive within the areas of integratingdeformation along three paths across the SW nonidealtransform accommodation. Surface area dilation United Statesfrom east of the Rio GrandeRift to near tendsto occurwithin pull-apart steps of majorfight- thecontinental escarpment, thereby crossing the defor- lateralstrike-slip systems. The associated normal faults marionfield accommodating relative transform motion. areincorporated directly into the fault system and trend The kinematicinformation from each path is nearly obliquelyto theorientation of the strike-slip elements independent of that from the otherpaths, providing (seeFigure 15). TheSalton Trough has been attributed independent estimates of the Pacific-NorthAmerica to thisstyle of kinematics[e.g., Elders et al., 1972],as relativevelocity. These estimates are consistentwith havethe California valleys east of thesouthern Sierra eachother and combine to producea Pacific velocity Nevada(as discussedabove). In bothregions, large estimateof 48 :k3mm/yr N45øW +_2 ø with respect to amountsof crustalextension have occurred on fault NorthAmerica (for a locationsouth of the western systemsthat contribute nearly directly toward transform Transverse Ranges). Most of thekinematic information shearaccommodation (although, at leastin theregion usedin ourstudy is basedupon deformation representing eastof the southernSierra Nevada, some extension average rates younger than 20,000 years. The global occurson faultsthat contribute motion that is at a high modelsNUVEL-1 [DeMetset al., 1990]and NUVEL-1A angleto the strike-slipsystem). In contrastto the (DeMetset al., submittexlmanuscript, 1994)also strive kinematicsof transtension,transpression typically is for an estimateof ongoingmotion (by usingmagnetic accommodatedwith folds and thrust faults that lie off of lineationdata younger than about 3 m.y.). Comparedto thedominant local shear system and allow contraction in thevelocity predicted by NUVEL-1and NUVEL-1A, our a direction that is nearly normal to this system. This estimateis similar in rate but our directionfor the Pacific styleis describedabove for the TransverseRanges and is plate relative velocity is more counterclockwise(i.e., well expressedin the Coast Ranges[e.g., Mount and westerly)by 7ø +2ø . The strongestorientation constraint Suppe,1987] and has been proposed for thesouthern Los is obtained from the southernpath, where the Angelesbasin [Hauksson and Saldivar, 1989]. deformationoccurs primarily on strike-slipfaults that 19,994 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION trend -10 ø more counterclockwisein orientationthat of TransverseRanges path is discussedfirst becauseit is our the NUVEL-1 trend. Deformationon the other minor referencepath and becausethe Rio GrandeRift is most featuresencountered along this pathdoes not changethis easily introducedfor this path. A disproportionate orientationsignificanfiy. amountof the following discussionfocuses on the low The kinematic information used in our velocity deformationrates occurringacross the westernUnited calculationsfor the three paths representthree distinct States east of California. Although these rates are regions,each expressinga different style of transform relativelyinconsequential, it is importantto demonstrate deformation.The northernregion (central California) and that they do not accommodatethe differencebetween the southernregion (southernCalifornia), which accom- globally estimatedPacific plate velocity and the local modate transformmotion with relatively simple alefor- Pacific plate velocity estimatedbelow. In fact, defor- mationfields, are kinematicallyjoined across the complex marion east of California tends to increasethis dis- Transverse Ranges region. Loosely defined, two crepancy. importantshear zones are active: one near the continental marginand another near eastern California. The zone TransverseRanges Path near the margin includesthe San Andreasand Hosgri systemsin thenorthern region and borderland and trans- Rio GrandeRift. The firstsignificant deformation Peninsularfaults in the southernregion. The easternencountered along the Transverse Ranges path is theRio zoneincludes the San Andreas in the southern region and GrandeRift (Figurela), a seriesof basinsbounded the shearzones in the Mojaveand the westernGreat principallyby normalfaults, which extends -1000 km Basin.Throughout this entire area the San Andreas fault fromthe Mexican border to northernColorado [Chapin, is the dominatestructure. However, its rate variesfrom 1979;Kelley, 1979]. The rift hasbeen associated with regionto regionas it interactswith the two lesser extensionof the greaterBasin and Range region [e.g., transform-accommodatingshear zones. Both of these Stewart,1978; Tweto, 1979]. Also,because deformation shearsystems splay from the San Andreas in thevicinity acrossthe rift is the only significantdeformation of the TransverseRanges (Figure lb), with the trans- occurringbetween North Americaand the Colorado Peninsularsystem trending SSE from the western and Plateau,modem motion of the Plateauhas been centralTransverse Ranges and the easternCalifornia kinematically related to deformation within the rift. Most shearzone trending NNW fromthe eastern Transverse workers agree that extension across the rift increasesto Ranges. the south[e.g., Chapin, 1979] in a mannerconsistent Secondaryfaulting near the SanAndreas fault also with rotationof the ColoradoPlateau about a pole variesfrom region to region,accommodating dilation in locatednear the northern termination of the rift [Kelley, the southernregion and contractionin the northern 1979;Golombek et al., 1983]. Awayfrom the central region.The San Andreas fault defines a large left step rift,which trends 15 ø moreeasterly than the overall rift, whereit passesthrough the Transverse Ranges, which majorfaults within the rift as well as the rift itself trend woulda• torequire high rates of convergence in this north.Within the central rift, bothstructural [Kelley, region;however, southern California crust moves so as to 1979]and palcomagnetic data [Golombek et al., 1983] experienceonly relativelyminor rates of convergenceindicate extensionalfaulting with a componentof left- there.This is accomplishedby a combinationof the lateralslip. Left-lateraldeformation could result from movementof the Mojave block to the northand away rotationof the plateauor simple west directedextension fromthe zone of convergence(Figure 13) andby a acrossthe relatively easterly trending central rift. In fight counterclockwiserotation of the crust southof the of thestyle of deformationnorth and south of thecentral TransverseRanges as it movesaround the arcuate San rift,the second option appears more likely. Therefore we Andreasfault (Figures 9 and 10a), causing areduction of assume that extension occurs N90øW *, perpendicular to its northernvelocity (and an increasein its westernthis trend. Because rifts commonly open in directions velocity).Crust in the westernTransverse Ranges oblique to theirtrend we assignan uncertaintyin shortensand rotatesclockwise to flowin a NNW orientationof +30 ø*, which is greaterthan the variations direction,nearly parallel to the SanAndreas fault in in trendof themajor faults within the rift. We choose centralCalifornia. The counterclockwise andthen clock- 30ø becausethis is therange in extensiondirection that wiserotations of southern California crust, in conjunction may occuron obliquenormal faults without the withincreasing amounts of crustalcontraction, result in a accompanimentof significant amounts of strike-slip "funneling"of crust south of theTransverse Ranges into faulting[Withjack and Jamison, 1986]. Our uncertainty therelatively narrow central California Coast Ranges to rangeadmits the possibilitythat strike-slip faulting thenorth (Figure 12). relatedto ColoradoPlateau rotation occurs at abouthalf the extension rate. The long-termaverage extension rate acrossthe central rift has been -0.2 mm/yr, and the late Quaternaryrate is Appendix:Detailed Description ofDeformation inferredtobe -0.14 mm/yr [Golombek, 1981; Golombek EncounteredAlong the Three Paths 0.14etal., * mm/yr1983]. as We the use mostthe likely Quaternary rate for extension the central raterift. of In this appendixwe presentand discussthe kinematic For the upper limit we use the long-term averageof informationused to inferthe pattern of deformationin 0.20' mm/_•,and for the lowerlimit we arbitrarily the southwestUnited Statesand the Pacific-Northchoose 0.10' mm/yr.There are no published data for the America relative motion that it accommodates.Values relativemagnitude of left-lateralshear, but it is inferred used in Table 1 are indicatedbelow with daggers. The to occurat a fractionof the extensionrate. HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION 19,995

Colorado Plateau rotation. The kinematics of NE Arizona and southernmost Nevada. The most Colorado Plateau rotation can be described with an Euler significantdeformation encountered along this portionof pole. We estimatethis pole positionwith informationon the path occurs on a family of predominantlynormal Rio Granderifting, assumingthat both North Americato faults near the westernmargin of the Colorado Plateau. the east of the rift and the ColoradoPlateau are rigid Faults strike essentiallynorth where our path crossesthis entities. Above, we used a N90øW direction and an system,suggesting a westward* extension direction. As extension rate diminishing to zero at the Colorado- above,we assign+_30 ø* from this directionto permit Wyoming border, which define our preferred pole oblique extension. The northern continuationsof these position. The kinematic implicationsof rotation about faults have been studied in SW Utah [Hamblin et al., this pole are consistentwith mild north directed short- 1981; Mayer, 1986], the most active of thesebeing the ening along the northern Colorado Plateau margin Hurricane fault. From this work we infer a rate of about [Hamilton, 1988a]. The previously assignedrifting 1.0 mm/yr for the entire zone bounding the western orientationuncertainty of _+30ø correspondsto a pole Colorado Plateau. As discussed below in the section on positionuncertainty in longitudeof _+300km. This range the Sierra Nevada path, geodetic information suggests in possiblepole positiondoes not conflictwith the style that the Great Basin extension rate may be under- of deformation observed within several hundred estimated. However, the southern continuations of the kilometerseast and west of the northernrift, which is not plateau-boundingfaults appear less active than they are noted as being significantly active. Thus our best in SW Utah, which is consistent with the occurrence of estimateof the pole positionis at the intersectionof the left-lateralstrike-slip activity on the WSW trendingLake Wyoming-Coloradoborder with the Rio Grande trend Mead fault system [Bohannon, 1979, 1984], a set of (this coincides with the best estimate of Golombek et al. faults that lie between the Colorado Plateau-bounding [1983]), and we use an uncertaintyin the pole positionof faults in Utah and thosein NW Arizona (Figure la). We 300 km to describethe uncertaintyin plateaumotion and choosea rate of 0.5* _+0.5*mm/yr to representthe rift opening. deformationrate of the plateau-boundingfaults where The Colorado Plateau-North America pole position, traversedby our path. This range allows the boundaryof together with the extension rate of 0.10-0.20 mm/yr the plateau south of the Lake Mead system to be across the central Rio Grande Rift that we discussed completelyinactive or to extend at a rate equal to the above, imply a plateau rotation rate of 0.01-0.02ø/m.y. estimatedextension rate in SW Utah. The net velocityat Consideringthe uncertaintyin pole position,and because this point is a relatively small 0.6 _+0.5mm/yr N81øW either the Colorado Plateau or Wyoming may not be +35ø (Figure 5). behaving exactly as we have assumed,we widen these Mojave faults. Very little deformationis encountered limits to 0.0-0.04ø/m.y. The domain of possiblepole in the Mojave Desert east of longitude 116øW. West of positionsand the • trend of the central Rio Grande longitude 116øW is an active set of faults that define a Rift imply a ratio of left-lateral rate to rift rate of NNW trending shear zone separatingthe Mojave block between 0.00 and 0.7, with a best estimate of 0.25. from crust to the east that moves nearly with North Palcomagnetic data also supply information on America [Dokka and Travis, 1990]. This eastern Colorado Plateau rotation. Steiner [1986] documents California shearzone is composedof NW strikingright- 9-14 ø of net clockwise rotation about a Colorado lateral faults, east trending left-lateral faults, and less Plateau-NorthAmerica Euler pole near the SW marginof significantdip-slip faults of diversetrends. The average the plateau. However, Steiner suggeststhat there have strike of the most active right-lateral faults is N41øW, been three episodesof rotation,with the Cenozoicbeing with deviationsof +10ø . This orientationis considerably the leastimportant, and that the SW plateaupole position more westerly than the trend of the zone as a whole, so is most consistent with pre-Cenozoic rotation. This that if the orientationof the right-lateralfaults alone is interpretationis consistentwith Laramide-age defor- used to infer Mojave motion, the accommodationzone mation east of the Colorado Plateau that may have would pull apart. Becausethis zone is neither dilating accommodate•plateau rotation during this time [Ham- nor contractinggreatly [Dol&a and Travis, 1990], Mojave ilton, 1988a, 1988b;Greis, 1983]. Thus we interpretthe block motion must approximatelyparallel the average palcomagnetic data to be consistent with the late trend of this shearzone, and the blocksbounded by the Cenozoic geologic data, but of lower resolution. We right-lateral faults must rotate clockwise slightly to thereforeuse the geologicallyinferred deformation of the accommodate the motion of the Mojave. This is Rio Grande Rift to describe the rotation of the Colorado consistentwith geologic [Dol&a and Travis, 1990] and Plateau. palcomagneticdata [Wells and Hillhouse, 1989] for the The TransverseRanges path continues-600 km across region. We estimate that the overall shear zone has an the Colorado Plateau. At the western margin of the orientation of about N15øW *. plateau, the inferred plateau rotation rate results in a Our best orientation estimate assumes that the shear velocityvector (relative to NorthAmerica) of 0.1* _+0.1* zone accommodateswestern Mojave motion without mm/yroriented north* _+30 ø*. Thenet velocity vector at involving net extensionor contractionwithin the zone. this point, which is the sum of Rio Grandeextension and However, the faults within the shear zone bound a series Colorado Plateau rotation, is 0.25 mm/yr (oriented of valleysthat may accommodatea componentof normal roughly in the Pacific Plate-North America relative faulting. We infer that the valleysare attributedto either direction). Comparedto deformationrates found to the local fight steppingof strike-slipmotion or the formation west, this vectoris negligible. of rhombochasmswhere conjugateleft-lateral and right- Basin and Range. Our path leavesthe SW Colorado lateral faults meet. Shortening(such as that reportedby Plateauand erosseaa portionof the Basin and Rangein BartIcy et al. [1990]) is similarly attributedto local fault 19,996 HUMPttREYS AND WELDON: WESTERN U.S. DEFORMATION geometry. We choosean uncertaintyin the orientation orientationwill accommodateMojave block rotation at estimatethat is wide enoughto encompassa range of ratesof +lø/m.y.; we considerrotation rates beyond this possibilitiesthat includesimple strike-slip faulting on the rangeas unlikely. dominantfight-lateral faults in the Mojave [Sauberet al., San Andreas fault. The path next crossesthe San 1986] and the possibilitythat contraction[Bartley et al., Andreas fault at a location where agreement exists 1990] and modestcrustal thickening occurs because the betweenthe short-termand long-termslip rates. Weldon westernMojave block moves in a directionthat is more [1986] proposeda slip rate of 37.5 +2 mm/yr basedon northerlythan the averagetrend of the shearzone. This offsetof early to middle Pleistocenefanglomerates, dated rangeof possibilitiesis bounded by +_25 ø*. by volcanic ash beds and paleomagneticinformation. A Dokka and Travis [1990] estimate a net late Cenozoic more conservativeteevaluation of an expandeddata set offset of 65 km across these faults, with deformation [Weldonet al., 1993] concludesa rate of 36 +8 mm/yr. initiating between -10 and 6 Ma. They report that This rate has been contimed for slip during the past deformationhas shifted recently from the easternto the several hundred years (36 +7 mm/yr [Salyards, 1988; central set of Mojave faults. Net late Cenozoic offset Salyardseta/., 1992]) as well as for the past 4-5 m.y. across the mid-Mojave fault zone is about 35 km (-35 mm/yr [Matti et al., 1985; Frizzell et al., 1986; [Dibblee, 1967; Powell, 1981; Dokka, 1983]. If western Weldon, 1986; Meisling and Weldon, 1989]), and we use Mojave faulting became active with the initiation of the this more conservativevalue as our rate estimate. Also, modem San Andreas fault activity -5 Ma, a rate of 7 this rate is consistent with the summed rates of the San mm/yr is implied (thoughno strongevidence exists for Andreasand San Jacintofaults to the south,as proposed the time of slip initiationon thesefaults). A similarrate by Weldonand Sieh [1985]. Despitethis evidence,there is obtainedby averagingthe net -65 km of offsetover 10 are workerswho believe the slip rate is either greateror m.y. Geodetic data yield similar results. Land-based less:rates as low as 20 mm/yr [Barrowset al., 1985] and data [Savage et al., 1990] suggestthat this shear zone 20-38 mm/yr [Schwartz and Weldon, 1986] have been trends about N35øW and accommodates-8 mm/yr of proposed,and rates of greaterthan 50 mm/yr have been simpleright-lateral shear. The VLBI station"Mojave" in proposedby Rust [1982] and Ramirez [1982]. Geodetic the northern Mojave moves with respect to North measurementssuggest San Andreasrates near 30 mm/yr. America at 9 mm/yr N20øW [Ward, 1990], with a Space-basedgeodetic results suggest26.9 +_2.6mm/yr negligibleformal uncertaintyin its site velocity. [Sauber, 1989], while land-based results have been We representactivity on the Mojave portion of the interpretedto yield far-field rates of 18.3 +1.2 [Prescott et al., 1987], 30 _+6[Eberhart-Phillips et al., 1990], and N15øWeastern*California and uncertaintiesshearzone of with +34* a vector mm/yrofand 8' +_25 mmlo•*. -36 mm/yr [Savage, 1990]. The differencesin geo- This rate is slightlygreater than the -7 mm/yr estimated detically estimatedrate reflect, primarily, differencesin geologicallyacross the right-lateralfaults so as to include the parameterizafionof the deformationmodel used in a contribution from the conjugate faulting, and the fitting the data. All of the above San Andreas rate uncertaintylimits encompassthe range in rate estimates. estimatesthat deviateby morethan a few millimetersper The uncertaintyassigned to this deformationis relatively year from 36 mm/yr are basedeither on geodetics(that great because we are not sure how the geodetic rate also can be interpretedat ratesof-36 mm/yr) or on the compareswith the geologicrate, and the geologicrate, offset of poorly definedor dated features. Becausethe while consistent,is poorly constrained.The upperlimit more rex•nt interpretationsof the geodetic data are allows the geologically estimated offset of 65 km to consistentwith the relatively well constrainedgeologic occur at rates slightly faster than average over the estimates,we considerthe lower geodeticrate estimates shortest interval allowed, i.e., since 6 Ma [Dokka and to be unlikely,and we do not includethem in our range Travis, 1990]. The lower limit allows 45 km of net of possibleslip rates. We use 36* +8* mm/yrto offset (our estimate of the minimum net offset, based on describethe San Andreasslip rate, in accordancewith the work of Dokka and Travis [1990]) to have occurred Weldon et al. [1993]. Orientation of the San Andreas steadilysince 10.6 Ma (the maximumduration considered faultis well constrainedin this region at N64øW* +2ø*. possibleby Dokka and Travis [1990]). Rotation of the San Gabriel block. Although the There has been considerabledebate on the subjectof strike of the San Andreasfault is well definedat any late Cenozoicblock rotationswithin the Mojave. Large- point in the vicinity of our path crossing,the fault is scale counterclockwise[e.g., Garfunkel, 1974; Calderone rather curved throughthe TransverseRanges. As noted and Butler, 1984] and clockwise[e.g., Luyendyket al., by Weldon and Humphreys[1986], strike-slipmotion 1980; Bird and Rosenstock, 1984] rotations have been acrossthe arcuateSan Andreasfault requiresa rotationof proposed. More rex•nt paleomagneticwork in the south- the San Gabriel block with respectto the Mojave. We ern [Weldon et al., 1984; Winston, 1985; Weldon, 1986], use11 ø* +4ø * to representthe vector contribution arising central [Wells and Hillhouse, 1989], and southeastern from San Gabriel rotation west of our crossing. This Mojave [Calderone et al., 1986; 1990] suggestthat -11 ø counterclockwise addition to the San Gabriel exceptwithin fault zonesthe Mojave has rotatedless than velocity (with respect to the Mojave at our crossing) 5ø since the late Miocene. In particular, the zones in representsthe correction in the velocity orientation whichLuyendyk et al. [1980] proposedclockwise rotation incurredby traversingfrom where our path crossesonto lie within the eastern California shear zone. Also, the the San Gabriel block to near the western end of this lack of an easternstructure to form a rotationalboundary block. The uncertaintyincludes both the possibleerror in for the Mojave block arguesagainst significant rotation of estimatingthe curvatureof the San Andreasfault (+_2ø), the Mojave as a whole. Our assigneduncertainty in and the possibilitythat the San Gabriel block moves HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION 19,997 relativeto the Mojavewith a smallcomponent of fault- estimate11 * mm/yr,a rate consistentwith the various normal motion (+2ø). We are aware that significant rate data. To include the great rangeof possiblerate, we amountsof fault-normaldeformation occurs near the San assignuncertainty limits of +10-3' mm/yr. We chose Andreasfault [e.g., Weldon, 1986; Weldon and Springer, N10øE * as the directionof shorteningbecause it is 1988]. However,this deformationis attributedto local, roughlyperpendicular to the trendsof the majorfolds and minor undulationsin the trace of the San Andreasfault in faults [Yeats et al., 1988] and •use it is similar to the conjunctionwith strike-slipmotion of the San Andreas orientationof N17øE estimatedby Jacksonand Molnar fault in an orientationthat is consistentwith the regional, [1990]. We are not aware of any piercing fines across curved trace [Saucier et al., 1992]. Assuminga San the zone to confirm the actual direction of shortening. Andreasslip rate of 36 mm/yr,slip alongthis curved For thisreason we assignlarge errors of +20ø* to the fault is equivalentto 4.2ø/m.y. of San Gabriel block deformationvector. This allows either true shortening rotation. With pure strike-slipmotion at our San Andreas betweenN10øW and N30øE, or, for shorteningdirected crossingand no rotationof the San Gabriel block along N10øE,up to 7 mm/yr of associatedstrike-slip shear. the San Andreas,a 36 mm/yr slip rate would resultin 7 Faultsnear the northernedge of the westernTransverse mm/yr of convergenceat the westernside of the Mojave. Ranges,including the Santa Ynez and Big Pine faults, This is considered unreasonable. Our limits allow 2 accommodateleft-lateral slip [Dibblee, 1982; Jennings, mm/yr of convergenceor extension. 1974; Luyendyket al., 1980; Hornafius, 1985; Clark et Rotation of the San Gabriel block has been inferred al., 1984]. Additionalunrecognized left-lateral shear may from paleomagneticdata by Terres [1984], Terres and occur acrossthe westernTransverse Ranges, particularly Luyendyk[1985], and Hornafius[1985], who report 16ø in areasundergoing strong shortening. We are not aware +30ø of counterclockwiserotation since 10 Ma. Greater of good slip rate estimates for these faults, so we constraintis given by Liu [1990],who finds 27.5 +4.5 ø of arbitrarilyassign a cumulativerate of 2' +_2'mm/yr to counterclockwiserotation since 7 Ma. Assumingthe San reflect the consensusthat they may accommodatea few Andreasfault definesan arc that itself is not rotatingwith millimetersper year of slip. We includethe possibility respect to North America (as suggestedby a lack of of no left-lateral slip becauserecent work suggeststhat Mojave rotation [Weldon, 1986; Wells and Hillhouse, the SantaYnez fault, a memberof thisgroup, is currently 1989]),and that the curvatureof the San Andreasfault is not slippingin a strike-slipmanner [Narnson and Davis, not changingwith time, thenthe currentSan Andreas slip 1988; R. S. Yeats, personalcommunication, 1989]. The rate producescounterclockwise rotation at 4ø +lø/m.y. averageorientation of thesefaults is N85øE*, with a This rotationrate since7 Ma wouldproduce-28 ø of rangeof +10ø-20ø*. Includingthis 2 mm/yrof left- counterclockwiserotation of the San Gabriel block, a lateral deformationis similar to rotatingthe directionof resultthat is in agreementwith the paleomagneticdata. convergence-12 ø clockwise from the N10øE direction Western Transverse Ranges. West of the San assumedabove (for our estimatedconvergence rate of 11 Gabriel block the path crossesthe western Transverse mm/yr). Ranges and enters the California borderland. Three The final componentof deformationin the western majordeformational components are activein the western TransverseRanges results from clockwiserotation, which TransverseRanges: shortening,left-lateral faulting and we infer is ongoing. Luyendyket al. [1980, 1985] and clockw• rotation. While these elementsprobably are Hornafius [1985] documentan averagerotation rate for interrelated,we considerthem separatelyin an attemptto the westernTransverse Ranges and the ChannelIslands establishconservative limits on deformationuncertainty. of --4ø/m.y.since 15 Ma, and apparentlylower ratessince The dominant deformation across the western -5 Ma. They accountfor this with a model in which the Transverse Ranges is shortening oriented essentially Transverse Ranges/Channel Islands blocks rotate north-south.Yeats [1983] proposeda rate of about 23 counterclockwisewithin a zone of NW trending, mm/yr, basedon crosssections across the VenturaBasin elongatedblocks accommodatingregional right-lateral region, which he believed representedmost of the shear. In this model the western Transverse convergenceacross the region. Using similar methods, Ranges/ChannelIslands blocks are separatedfrom the Narnsonand Davis [1988] estimate22 +4.5 mm/yr across NW trendingblocks by left-lateral faults. Jacksonand this zone. Rockwell[1983] and Rockwellet al. [1984] Molnar [1990] estimatea rotationrate of 6 +3ø/m.y. proposeda rate of 17 +4 mm/yr based on offset of usinga similar block model and VLBI data, suggesting Quaternarysurfaces, and recent work on individual that rotationis ongoing. However,as discussedin the structureswithin the zone [Yeats, 1988; Yeats et al., "kinematicsof Pacific-NorthAmerica interaction"section, 1988; Rockwell, 1988; Rockwellet al., 1988] supports we believe that the VLBI stationsthey use O/AND and similarlyhigh rates. However,Yeats [1993] suggests that JPL, Figure 12) are in regionsnot representativeof the ratescould be abouthalf (or less)of thosehe proposed kinematicsof the Ventura Basin region. Furthermore, earlierif the reversefaults that accommodateshortening while this kinematicdescription may have been valid do notflatten into a midcrustaldetachment surface, as whenright-lateral shear was concentrated in the outer assumedin earlier work [Yeats, 1981], and a criticalage continentalborderland, the low rates of currentactivity estimate is correctedto be consistentwith recent acrossleft-lateral faults within the westernTransverse evidence[Ponti et al., 1993]. Ratesof shorteningRanges and the paleomagneticdata for recently estimatedwith geodeticdata yield a rate of 1I +3 across decreasingrotation rates argue against high rates of the western TransverseRanges near the Ventura Basin ongoingrotation at the longitudeof the Ventura Basin. [Feigl et al., 1993; Donnellan et al., 1993], consistent We choose a rotation rate of 2 +2ø/m.y. to describe with the revised geologicestimate. We use as our best western TransverseRanges rotation. This allows the 19,998 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION possibilityof no rotation, and the averagepaleomagnetic Assuming 15% of net extension across the 750 km of rate since 15 Ma. This rotation'rate is associated with a southernBasin and Range since 13 Ma yields an average velocityvector of 2 * +2* mm/yrdirected N95øW * +30ø * rate of-7 mm/yr. One to 10% of this rate is -0-1 for a pole near the latitude of the TransverseRanges. mm/yr. The net relative velocity vector for the western Choosinga rate that is intermediatewithin this range TransverseRanges path is 13 +5-3 mm/yr N10øE +25ø of estimates,we use0.5 * mm/yras our best estimate for (at 90% confidence). the net extensionrate along the southernpath between The end-of-pathrelative velocity vector calculatedfor the Rio Grande Rift and California, with a range from the Transverse Ranges path is 46 +_5 mm/yr oriented 0.25* to 1.0* mm/yr. The mostactive faults in the N50øW _+7ø. The uncertainty bounds are given at the southernBasin and Range are normal faults that trend 90% confidence level. north to NNW [Menges and Pearthree, 1983]. Other faults that may contributea small amountof deformation trendNW to NNW. We takeN100øW • as our preferred Peninsular Ranges Path directionand assign an errorof +30ø • to accommodate Rio Grande. Like the TransverseRanges path, the the range of fault trends and the possibleoccurrence of first significantdeformation encountered along this path is oblique extension. associatedwith the Rio Grande Rift. Publishedslip rate California east of the San Andreas fault. We estimatesexist only for the centralportion of the rift, east include southern Basin and Range faulting near the of the Colorado Plateau, and have been discussedabove California border with the deformation estimate for with respectto the TransverseRanges path. The long- extensionacross the southernArizona Basin and Range. term averageextension rate there is -0.2 mm/yr, and the Deformationthere since -5 Ma is minor [Buising,1992; late Quaternary rate is inferred to be -0.14 mm/yr Richard et al., 1992] and is not likely to influenceour [Golombek, 1981; Golombek eta/., 1983]. path-integratedvelocity estimateby a significantamount. The Peninsular Ranges path crossesthe rift south of The region east of the San Andreas fault is an area the central rift, in a region where the rift and the major broken by east trending left-lateral faults and north to faults within it trend north. Therefore we assume that NW trending normal faults. It has been suggested extensionoccurs N90øW *, perpendicularto this trend. [Richardand Dokka, 1992] that this region is part of the As discussed under the Transverse Ranges path eastern California shear zone recognizedfarther north discussionabove, we assignan uncertaintyin orientation [Dokka and Travis, 1990]. Little evidence exists for the of _+30ø*. Giventhe southward increasing net extension occurrence of deformationin this region at rates similar and the estimated Quaternary extension rate for the to those found farther north, and we think that the eastern central rift, a simple extrapolationof rate to where the California shearzone is more closelyassociated with the PeninsularRanges path crossesthe rift yields0.25 t San Andreas fault in the vicinity of the eastern mm/yr, which we use for our best estimate. As bounds TransverseRanges (althoughearlier it may have been to the extension rate we use the lowest central rift concentratedfarther eas0. This idea is supportedby estimate(0.1 * mm/yr)and the maximumcentral rift geodetic observationsacross the southernMojave that estimatewith a linear extrapolationin rate to the south indicatemuch greater strain rates in the westernMojave (0.5* mm/yr). than in the eastern Mojave [Savage et al., 1990]. Southern Basin and Range. The southern path However, earthquakeactivity east of the San Andreas crossesapproximately 750 km of Basin and Range fault suggestssome right-lateral activity on the NW between the Rio Grande Rift and the San Andreas fault. trendingfaults and some left-lateral activity on the east The large number of poorly studied structuresthere trending faults. The east trending faults appear to be makes estimatingdeformation rate difficult. There have more important,and these have been used to describea been 6-7 scarp-producingearthquakes in southeastern regionaldeformation where the blocks they boundrotate Arizona during the past 15,(X}0-20,000 years [Menges clockwise,producing -40 ø of paleomagneticallyinferred and Pearthree, 1983; Pearthree et al., 1983; Pearthree, rotation since -12 Ma [Carter et al., 1987]. An 1986], suggestinga rate of a few tenthsof a millimeter alternative model, in which left-lateral faulting between per year acrossthat zone. While there are a few late theseblocks occurs as the blockspull away from Arizona Quaternary scarps and scattered seismicity along the with increasingdisplacement to the north, seemsunlikely southernpath in southcentral and southwestArizona, it becauseof •he nearly undeformed-5 Ma sedimentsthat is unlikely that those areas add significantlyto the total lie eastof the easttrending faults [Buising,1992; Richard deformation [Pearthree, 1986]. et al., 1992]. The east trendingfaults have an average Consideringthe region as a whole, most workersagree orientation of N100øW, and their rotation would result in that the southernBasin and Range is currently1-2 orders a velocity at their westernend oriented-N10øW (with of magnitude less active than the northern Basin and respectto their easternend). Any right-lateralactivity on Range [e.g., Stewart, 1978; Greensfelderet al., 1980; the NW trendingfaults would result in a more westerly Zoback, et al., 1981; Pearthree, 1986]. As discussed orientationfor the velocitltjust east of the SanAndreas with respect to the Sierra Nevada path, the northern fault• We chooseN20øW • as our preferredorientation so Basin and Range is extending at 5-15 mm/yr, which as to admit some componentof right-lateraldeformation suggestsa southern Basin and Range rate of between and so that deformation in this zone parallels the 0.05 (zero for our purposes)and 1.5 mm/yr. A similar proposedeastern California shear zone in this area. To estimate could be inferred from the consensus that current allow for a range of possible interpretationsfor the southernBasin and Range extensionoccurs at a rate that deformationin this region, we assigna wide uncertainty is 1-2 ordersof magnitudeless than the 13 Ma average. of +30o-20ø*. HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION 19,999

The length of the east trendingfaults (and the blocks strain associated with the San Jacinto fault makes it they bound)decrease to the south;at the latitude of our difficult to use the geodetic data to constrain the San path they are --70 km long. If these blocks have been Andreas fault strain-rate estimate. rotating at a constant rate over the last 12 m.y., the We use N48øW * with uncertaintiesof +7o-3ø* for the velocity at the west end relative to the east end resulting vector representingSan Andreas slip orientation. This in 40ø of rotation is -4 mm/yr. It seems more likely, assumespure right-lateral slip acrossthe averagetrend of however, that these faults are most active when they lie the zone, with uncertaintiesrepresenting the range of east of a pronouncedleft bend in the San Andreasfault trendsfound for the fault nearby. North and southof our and permit the crust there to negotiatethe bend; in this crossingpoint there are recently uplifted hills that have case, the rate of fault activity diminishes as this crust been interpretedas indicating transpressivemotion across moves south of the bend [Powell, 1981]. In support of the San Andreas fault [e.g., Sylvesterand Smith, 1976]. this belief, deformation rates east of the northern Salton We infer that these structuresresult from geometric Trough appear to be greatestwithin 10 km of the San complexitiesin the local fault zone structurerather from Andreas fault across a zone of north to NW trending regional transpression.A potentially more seriousprob- normal faults that displaceby only a few metersthe same lem may be the occurrenceof extensionacross the Salton late Quaternarysurfaces offset hundredsof meters by the Trough as a whole. Our best slip vector is basedon the San Andreas fault [Clark et al., 1984]. This observation assumptionthat the depressionis due entirely to a pull- implies that thesefaults contributeless than a few tenths apart mechanism,as is widely believed [e.g., Elders et of a millimeter per year. By consideringthe set of data al., 1972; Crowell and Sylvester, 1979], but if a for deformation in California east of the San Andreas componentof extensionperpendicular to the trend of the fault, a net velocity of 2-3 mm/yr would be a San Andreas fault occurs, our vector should trend more conservative maximal estimate for deformation over this westerly. The range of trends allowed by our assigned broader region. For the velocity incurred by crossing uncertaintyfor the San Andreasslip vector allows nearly easternCalifornia in the vicinity of our path, we assigna 2 mm/yr of fault-normal contractionand 4 mm/yr of bestestimate of 1* mm/yr,and limits that includeno* extension. motionand 3 * mm/yr. San Jacinto fault. The next fault encounteredalong Each of the three deformationregions encountered east our path is the San Jacinto,which is crossedat a location of the San Andreas fault (the Rio Grande Rift, the where the fault zone is relatively simple (consisting southernBasin and Range, and easternCalifornia) •s essent_i_a_llyof one branch), and there are a number of probably deforming at less than ! mm/yr. For the relatively consistent slip-rate estimates. Sharp [1981] combinedregion, we calculatethat between0.3 and 2.0 proposeda minimum rate of 8-12 mm/yr based on the mm/yr of net velocity is accumulatedoriented N44øW offset of gravelscontaining the 0.7 Ma Bishopash. We +45ø, with a best estimate of slighfiy over 1 mm/yr. infer from his discussionof the relationshipsthat the While it is difficult to quantify this small amount of actual slip rate value is likely to be close to this range. deformation,it should be clear that the overall style is More recent work in the area [Rockwell et al., 1986b; extensionin the direction of our path at rates very low Merifield et al., 1987] suggestsan offset rate of 12-14 comparedto the other rates encounteredalong the path mm/yr for many latestPleistocene and Holocenedeposits farther west. and surfaces. Because of the consistencyamong these San Andreas fault. The Peninsular Ranges path ratesand our interpretationof Sharp'slong-term rate, we crossesthe San Andreasat a place where the fault zone use12' mm/yras our bestrate estimate. Sharp [1981] is fairly simple,where the bestslip rate dataexist and the argues that the short-term slip rate varies, and he fault trend is well constrained.Sieh [1986; also personal estimatesrates as low as severalmillimeters per year for communication,1988] has inferred a slip rate of • 30 the pastseveral thousand years. We do not includethese low numbersin our estimate becausethey were derived mm/yr based on 21 m of offset causedby at least four from other portionsof the fault, where the fault zone is earthquakesbetween 1000 and 1700 A.D. The actual more complex and where we believe additional dis- averageslip rate may be greaterthan 30 mm/yr because tributed deformationis occurring. (That is, we infer that displacementacross at leastone strandof the fault is not rate variability across San Jacinto fault zone is spatial included in this estimate [$ieh, 1986]. However, because rather than temporal, and probably representsvarying $ieh [1986] infers that an earthquakeoccurred at both the fractions of the total deformation accommodated off the beginningand end of the measuredinterval, it is also main trace.) In either case, we follow our overall possiblethat the averagerate is overestimatedby an philosophyof using the longer-termrate at the simplest amount associated with the displacement of one site. We assigna rateuncertainty of +4' mm/yrto our earthquake.We useSieh's estimate of 30' mm/yras the rate estimateof 12 mm/yr. bestestimate and includean errorof +7* mm/yrto The San Jacinto fault strikes N53øW * where crossed accommodatethese counterbalancingpossibilities. This by our path. Offset featuresnearby are consistentwith rate estimateis supportedby the work of Keller et al. pure strike-slip there [Sharp, 1981; Rockwell et al., [1982], who, on the basis of an offset fan crossingthe 1986b; Merifield et al., 1987]. The San Jacinto fault fault just north of where our path crosses the San zone consistsof a seriesof what appear to be pull-apart Andreas fault, determine a rate of 10-35 mm/yr, with valleysbetween overlapping fight-stepping fault segments 23-35 considered"most likely." Geodeticmeasurements that trend from about N43øW to N58øW. This range of the slip rate fall in this range as well [e.g., Savage, yieldslimits on our preferredtrend of +5ø-10ø*. The 1983]; however, the short time interval representedby trendat at our crossingis 5ø more westerlythan the trend these measurementsand the difficulty of separatingthe of the zone as a whole [N48øW, Sharp, 1967], which is 20,000 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION consistentwith the observedslightly transtensionalnature decreaseas the Euler pole is approached,and where the of the fault zone as a whole. path has a componentthat is parallel to curved San Elsinore fault and nearby faults to the east. Andreasfault, this vector will rotate as the path is Between the San Jacinto and Elsinore fault zones lie a traversed. number of poorly characterizedNW trendingfault zones As discussedunder the TransverseRange path, the San (not shown on our figures), including the Hot Springs, Gabrielblock probablyrotates at a rate impliedby slip Agua Caliente, San Felipe, and EarthquakeValley fault along the curved San Andreas fault. If the blocks south zones. Although we know of no publishedslip rate of the San Gabriel block were not rotatingwith the San estimatesfor these zones, portionsare consideredactive Gabrielblock, a "jaw-like"effect would occur,whereby [Jennings,1974; Rockwell et al., 1989]. As a set these northdirected shortening would occur with an increasing faults form a broad zone trending N58øW ñ20ø rate toward the west. This results in a greater connectingthe southern San Jacinto with the northern convergencerate north of the Los Angeles Basin and Elsinore fault zones and perhapstransferring slip from across at least the eastern portion of the western the Salton Trough to the northern Elsinore fault zone. Transverse Ranges, relative to the central Transverse Given the similar trend of thesefaults with known right- Ranges, as is observed. Complicating the situation, lateral faults in the area, we expect that these faults are however,is the similar effect producedby right-lateral predominanfiy right-lateral. Rockwell et al. [1989], slip on the San Jacinto,Elsinore, and Newport-Inglewood however,discuss left-lateral slip on at least one strandof faults,which also causesincreasing rates of thrustingto this fault zone. By comparisonwith the morphologyof occurtoward the west. In lieu of a thoroughanalysis of the Elsinore and San Jacinto faults we infer that this zone the combination of effects, we simply assume that contributesless than 1 mm/yr of slip. To include the rotationof the entireregion north of the San Miguel and possibilityof either left- or right-lateralslip, we assigna Rose Canyon fault systemsis occurringat half the rate slip rate of 0 +1 mm/yr. For presentationpurposes, we implied by the curvatureand rate of slip on the San includethis slip with the Elsinorefault. Andreas fault, whereas the San Gabriel block rotates at Our path crosses the Elsinore fault near where the full rate implied by the San Andreaskinematics. As Vaughan and Rockwell [1986] have proposeda slip rate limits we choose a possible range of rotation that of 5 mm/yr with uncertaintiesof +1 -2 mm/yr. This rate includesboth no rotationand full rotationof this region. is comparablewith other slip rate estimatesalong the To completethe path west of the Elsinorefault, we fault producedby this group [e.g., Pinault and Rockwell, estimate the velocity resulting from two independent 1984; Rockwell et al., 1986a]. These estimatesare on the paths (Figures 2 and 4) that end in the southern high end of the range of earlier published estimates, California borderland. One path heads south into which range from 1 mm/yr [Ziony and Yerkes,1984] to 7 northernBaja before heading offshore and back to the mm/yr [Kennedy, 1977]. In every study, however, the southernCalifornia borderland, and the otherpath heads fault zone was found to be complex,and few absolute directlytoward the assumedsouthern California pole of dates have been reported. We use the best estimateof rotationand into the borderland(see Figure 2). The first Vaughan and Rockwell [1986] from their site near our path usesthe rate informationdetermined for the northern path(i.e., 5' mm/yr),but extend the uncertainty to +24 Baja faultsbut has poor orientationconstraint owing to mm/yr to accommodatethe range in publishedrates. uncertainties in the rotation rate of these blocks. The Where the southernpath crossesthe Elsinore, the fault second path has better orientation control because it undergoesseveral changesin trend. The average trend trendstoward the pole, so block rotation does not affect thereis aboutN49øW *, whichis comparableto the the orientation.However, rate informationis pooralong average trend of the zone as a whole. Individual faults this path. When convolvedtogether, these two paths deviate+18 ø -10ø from this orientation. Convolvingthe result in a well constrained estimate. Elsinore velocity probability field with that of the NW The velocitycontributed to a path integralby crossing trending faults between the San Jacinto and Elsinore a region experiencingrotation dependson the locations faults mentioned above results in slightly greater where the path entersand leavesthe rotatingblock. The uncertainty:rate uncertainty limits of +3-5* mm/yrand vector associated with the block rotation for the southern orientationuncertainty limits of +20ø-12ø*. California option, including contributionsfor radial and Rotation of southern California blocks. The trace of tangentialpath segments, is 3.5' +3.5* mm/yroriented the San Andreasfault zone is distinctlycurved southof N132øE*+10 ø*. Thisorientation is very near to thatof and through the Transverse Ranges. Discounting the southernCalifornia faults and hencedoes not degrade stepoversand other local geometric complexities,the seriouslythe uncertaintyin orientation. The correction curvature is remarkably constant from the western vectorfor the northernBaja optionis 4' +4' mm/yr termination of the Garlock fault to the Mexican border, orientedN83øE * +20 ø*. includingthe Imperial fault (Figure lb). As pointedout Northern Baja faults. The San Miguel-Vallicitos and by Weldon and Humphreys [1986], pure strike-slip Agua Bianca fault zones (Figure lb) are the two main motion along this curved trace produces a onshorezones crossedby the southernpath in Mexico. counterclockwise rotation of the crest SW of the fault, The San Miguel-Vallicitosis a zone of seismicactivity and motions associated with this rotation should be that has received very little field work. On the basis of estimatedand includedin our path integration. Taking its morphologyand trend, which is parallel to the Agua the San Andreasfault as the northernboundary of the Bianca fault, it is likely to be a slightly active right- rotatingregion, once the integrationpath crossesthe San lateral fault. Bemuse little is known about this fault, we Andreas fault the San Andreas-relatedvelocity will arbitrarilyassign a rateof 1* ñ1* mm/yr,based solely on HUMP•YS AND WELDON: V•rESTERN U.S. DEFORMATION 20,001 its seismicityand morphology. It trendsN55øW t _+8øt. accumulating deformation on the faults thought to be The Agua Blanca fault is a major right-lateralfault [Allen kinematiccontinuations of the northernBaja faults [Legg et al., 1960], connectingthe California borderlandto the et al., 1991]. Good orientation information exists for Gulf of California. Preliminary work on the late thesefaults [Legg, 1991; Legg et al., 1989, 1991], but no Quaternaryslip rate of the main trace suggestsa rate of reliable estimatesexist for the slip rates of these faults. about4 mm/yr [Schuget al., 1987; Hatch and Rockwell, We include velocity information for the Rose Canyon, 1986], with the possibilityof 1-2 mm/yr on the other Coronado Bank, and San Diego Trough fault systems, strandsthought to be important(T. K. Rockwell,personal which Legg describesas kinematic continuationsof the communication,1987). On the basis of their work we northern Baja faults. Because there is very lime rate assigna rateof 5 • _+2* mm/yrto thenorthern and main control, we allow a wide range of possible rate: a strand,and include1 * mm/yrto the upperuncertainty preferred rate of 1* mm/yron eachfault system,a estimateto accountfor possibledeformation beyond that minimumrate of 0.5* mm/yr,and a maximumrate of 3 * measuredon these traces. In the region where the fault mm/yr on the Rose Canyon and CoronadoBank systems is a simplezone the averagetrend is N67øW*, though and4 * mm/yron the SanDiego Trough (a higherrate individualsegments trend +10ø-5 ø* fromthis. because of the relatively pronounced physiography Our path continues west across the continental associatedwith this fau10. Orientationsare: N37øW* borderlandjust south of the Mexican border. Once +7ø * for theRose Canyon fault zone, N33øW* _+7 ø* for offshore there is litfie information other than the theCoronado Bank fault zone, and N32øW* _+6 ø* for the seismicityand morphologyof the faults with which to San Diego Troughfault zone. We alsoinclude a velocity estimate styles and rates of activity. Using detailed vector for the San Clemente fault, which is known to be topography determined from Seabeam surveys, Legg seismically active, but which is inferred to be kine- [1991] and Legg et al. [1989] have documentedon matically independentof the northernBaja and inner offshorefaults many of the classicmorphologic features borderland faults. As with the other borderland faults, its associatedwith active strike-slipfaulting, includingpull- orientationis well constrained,N46øW* _+5 ø*, andits rate aparts,rift valleys,scarps, and shutterridges. Legg et al. is poorlyconstrained. The rate of seismicityof the San [1991] note that the Agua Blanca fault zone joins the Clementefault appearsto be greaterthan any of the inner offshore inner borderland faults near the coast and that borderlandfaults but significantlyless than the seismicity there is another independentsystem of active faults, of the combined set of inner borderland faults. To admit associatedwith the San Isidro fault, connectingCalifornia this wide range, we representSan Clementeactivity with borderlandfaults farther to the west with faults along the 2' mm/yr+5-1.5' mm/yr. Thisallows a rangeof slip Pacific margin of Baja. Seismicity and documented from 0.5 mm/yr to 7 mm/yr. faulted offsetsof late Quaternarydeposits along the Baja The velocity estimatefor faults west of the Elsinore coast [Yeatsand Haq, 1981] provide additionalevidence providedby this path is 4.4 +3.0 mm/yr orientedN32øW for the occurrenceof deformationwest of Baja. From +9 ø-18 ø (at 90% confidence). Legg'swork we infer that this outersystem is activein a By determiningeach PeninsularRange-to-borderland right-lateral sense, and on the basis of its modest subrouteindependently and taking the productof their seismicitywe assign2' mm/yrof slip. Becausethis is uncertaintydistributions (Figure 4) we have an estimate little more than a guess,we assigna large uncertaintyof for the rate of deformationover this portion of the plate ß+2' mm/yr,which includes, as a limit,no motion.The boundaryof 5.3 +1.7 mm/yr at N29øW +7o-9ø (at 90% trendof thefaults west of theAgua Blanca is N32øW* confidence),where a minor 1.4ø clockwiserotation of the +8ø . If theseare pure right-lateralstrike-slip faults, this northern Baja vector has been made. This is the is the trend of the velocity vector representingtheir correctionneeded to bring the end of the Baja path deformation. From Legg's work we infer that motion on option to the southern California borderland along a these faults is nearly pure strike-slip. However, to NUVEL-1 or RM2 Pacific-North America small circle. includethe possibilityof a normal componentof slip to Our estimateis comparablewith both our borderlandand producethe distinctiveextensional morphology of the northernBaja faults slip rate estimates.Anderson [1979] borderland, we add an additional 7 ø to the western limit infers a combinedrate of <10 mm/yr for the same region, of the directionof motion(for a total of 15ø*), and to basedon seismicity. include the possibility of a thrust component,as sug- By convolving the estimate for motion west of the gested by thrust mechanismsfound in the borderland Elsinore fault with the estimate for motion across and farther north [e.g., Pacheco and Nabelek, 1988], we east of the Elsinore fault we arrive at a net end-of-path include an additional 7 ø to the northern limit (for a total estimatefor the southernpath of 52 +6 mm/yr oriented of 15ø*). Effectively,our best vector for motionon this N48øW +2 ø (uncertaintiesgiven at the 90% confidence fault systemallows pure right-lateral strike-slipmotion level). This result is representedgraphically in Figure 3. acrossthe average trend of the faults, and the range of uncertaintyallows up to 25% convergenceand extension. Sierra Nevada Path The velocity estimate for faults west of the Elsinore In this section, Pacific-North America relative plate prodded by this path is 5.5 +1.9 mm/yr orientedN32øW motion is estimatedby consideringdeformation along a +20o-30ø . The uncertaintybounds are given at the 90% path that crossesthe active southernGreat Basin, the San confidence level. Andreasfault where its behavioris relatively simple, and California faulting west of the Elsinore fault. We a complex zone of shorteningand lateral sheareast and also determinea velocity estimatefor a path option that west of the San Andreas fault (Figure 2). This is heads SW from the point of Elsinore fault crossing, kinematically equivalent to the route considered by 20,002 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION

Minster and Jordan [1984]. To help assessthe error whole), the presenceof NNW to NW trending faults associatedwith this path where it crossesthe SW Great which accommodaterelatively high rates of right-lateral Basin, we also considera path that leads south out of shear[Stewart, 1988; Burchfiel eta/., 1987;Lubetkin and southernNevada and into the Mojave and then crosses Clark, 1988], and a kinematicrelationship to the Garlock the westernGarlock fault, joining the main path in the fault and deformationoccurring in the Mojave Desert Sierra Nevada. [Sauber, 1989; Dokka and Travis, 1990]. East of California. At the northerly latitude of this The Death Valley region is the first zone of ongoing path, Rio Grande Rift extensionis nearly absent. By deformation encountered along our path within extrapolatingthe ratesdiscussed in the previouspaths to California. De•th Valley graben formation has been the location in south central Colorado where this path attributedto extensionacross a fight step between the crossesthe rift, we infera rateof 0.05' _+0.05* mm/yr right-lateralFurnace Creek and Death Valley fault zones orientedto the west* +_30ø*. The velocitycontribution [Burchfiel and Stewart,1966; Wrightet al., 1974;Troxel providedby ColoradoPlateau rotation is the sameas that and Wright, 1987], thoughthe southernDeath Valley discussedunder the Transverse Rang[es path: 0.1 * +0.1* fault zone may accommodateonly a fractionof the slip mm/yrdirected to the north*+30 ø*. The net velocity experiencedby the northernDeath Valley/Furn'aceCreek resultingfrom Rio Granderifting and ColoradoPlateau fault system[Stewart, 1988]. Unfortunately,the bestrate rotationis 0.11 _-_+0.10mm/yr orientedN30øW +30 ø. estimate is associated with the southern Death Valley Deformationwithin the Great Basin is heterogeneously fault system. Butler et al. [1988] estimate• 3.5 mm/yr distributed. Northern Great Basin extension is currently across this zone for the duration -11-1 Ma, consistent in concentrated in three zones: the Wasatch and Sierra rate with the -80 km of late Cenozoic displacement Nevada fronts and the central Nevada seismic zone inferred for the Panamint Mountains block on the [Wallace,1984; Eaton et al., 1978]. Seismicityoccurring Furnace Creek fault zone [Stewart, 1988]. However, near the WasatchFront is part of a seismicallyactive belt Butler et al. [1988] find the southernDeath Valley fault that coincideswith the Great Basin margins to the east zone to be nearly inactive since-1 Ma. Across the (the southernintermountain seismic belt [Smithand Sbar, northernDeath Valley and Furnace Creek fault zones, 1974]) and south [Smith, 1978]. Schwartz and Clark et al. [1984] estimate a Quaternaryslip rate of-1 Coppersmith[1984] estimatean extensionrate of-1 mm/yr, which is consistentwith the geomorphology mm/yron the WasatchFront. However,the VLBI station there. ELY in NE Nevadaappears to be movingN98øW at 4.9 We usethe 1* mm/yrslip rate estimate of Clarket al. +1.3 mm/yr Dixon et al. (submittedmanuscript, 1994), [1984] to characterizethe currentslip rate of the Furnace and the WasatchFront representsthe only kinematically Creek-Death Valley fault system. This rate is importantfeature known between stable North America intermediateto the range of possiblevalues and is and ELY. Hence additional deformation may be suggestedby the Quaternaryestimate. To representthe occurringnear the WasatchFront or in the easternGreat widerange of uncertaintyin the sliprate estimate, we use Basin. Our path crossesthe easternmargin of the Great valuesthat include both no * slipand slip at 4 * mm/yr(to Basin in SW Utah. Deformation there occurson a family admit Butler's -10 m.y. average,where we must assume of predominantlynormal faults near the westernmargin that since 1 Ma, nearbyfeatures presently accommodate of the Colorado Plateau, and the most active of these the deformation).Orientation of the major strike-slip faults is the Hurricane [Hamblin et al., 1981; Mayer, elementscomprising the Death Valley-FurnaceCreek 1986]. From this work we infer a rate of about 1.0 +0.5 systemisN47øW* +8 ø *. mm/yr for the entirezone bounding the westernColorado The next active documentedzone encounteredalong Plateau. Motivatedby the geodeticallyinferred motion our path is the HunterMountain fault system.This of ELY and its apparentcontradiction with the geologic systemis associatedwith Salineand Panamint Valleys, data, we includethe possibilityof additionaldeformation both of whichare consideredby Burchfielet al. [1987] to near this fxontalsystem of faultsor in the southernGreat be pull-apartbasins similar in kinematicorigin to Death Basinby choosinga rate of 2.0' +3.0-1.5'mm/yr. This Valley. Theseresearchers estimate a displacementof 8- allowsrates as greatas the inferredmotion of ELY and 10km since -3 Ma, yieldingan average rate of 3* +1* as low as 0.5 mm/yr. Faultstrend essentially to the north mm/yr. Wherethe system is strike-slipin natureit trends where our path crossesthis fault zone, suggestinga N57øW * +6ø*. This rate is similar to the Holoceneslip westward* extensiondirection. As we have done above rate of 2.4 (uncertaintyof _+0.8mm/yr) foundby Zhang in characterizingextensional zones, we assign+30 ø* et al. [1990] for the southernPanamint Valley fault zone, from this direction to permit for oblique extension. Net which is the southern continuation of the Hunter deformationto this point in southernNevada is 2.1 +2.1- Mountain fault zone. The trend of the Panamint Valley 1.1 mm/yr directedN86øW +22ø (at 90% confidence). fault zoneis nearlyto the north,though local topography Great Basin option. The westernportion of the Great suggeststhat nearbynormal faults are active (as is Basin, where crossedby our path, consistsof an actively requiredby the pull-apartmodel for the originof the deforming, triangular-shaped region (Figure 1b) valley). delineatedby the SierraNevada on the west sideand the Lubetkin and Clark [1988] have documented a Death Valley fault systemon the eastside. The Garlock Holoceneright-lateral slip rate of 1-2 mm/yracross the fault commonlyis believedto definethe southernlimit of OwensValley fault zone,and they suggest that this zone this active extensional domain [Davis and Burchfiel, is the mostactive of the strike-slipOwens Valley faults. 1973]. This region is characterizedby relatively high This value is less than the 3 to 7 mm/yr rate inferred rates of extension (compared to the Great Basin as a from geodeticstudies [Savage et al., 1975;Savage and HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION 20,003

Lisowski,1980]. We use the Holoceneslip rate estimate, the two active zones, giving slight preference to the consistentwith our convention of choosing such data intermountainbelt •use of its greater seismicity. In when they are available. The orientationof this fault the maximum limit, this rate is as great as our best zone is N12øW +8ø. While the Owens Valley earthquake estimate for extensionin the vicinity of the Hurricane of 1872 createdsignificant vertical relief [Lubetkinand and fault system(discussed above in this section). This Clark, 1988], this fault is essentiallystrike-slip in nature uncertainty range also allows the possibility of no [Beanlandand Clark, 1987]. The southernSierra Nevada activity. The next significantactivity encounteredis that frontal systemis dominatedby normal slip [Gillespie, in the easternMojave, which is discussedin the section 1982]. However, this deformationis documentedto be abovethat discussesthe TransverseRanges path (8.0* less active than the Owens Valley strike-slipactivity by +3-4* mm/yrN15øW * +25ø at least an order of magnitude[Clark et al., 1984; The Garlock fault slip rate near the SW comer of the Zoback, 1989]. Thus deformation in the Owens Valley Great Basin is estimatedat -7 mm/yr [Clark and Lajoie, region is dominatedby the Owens Valley fault, and, 1974; Astiz and Allen, 1983; Pampeyanet al., 1988]. We accordingly,we assigna velocity to the deformation [Weldon and Humphreys, 1986] had inferred from the acrossthis regionof 1.5' mm/yr+1.0-0.5' mm/yr, work of Carter [1980; 1982] that the rate was about 11 N17øW* +20ø-13ø*. The five additionalwesterly degrees mm/yr. However, more recent work has led B. A. Carter accommodatepossible normal extensionat a rate -10% (personal communication, 1988) to conclude that the that of the Owens Valley fault strike-slip rate, and offset units upon which his estimateis based are older orientationlimits have been widenedto allow a range in than he had thought earlier, requiring a lower slip rate frontal fault extensionrate from zero to 0.5 mm/yr. The consistentwith other estimates. We use 7 mm/yr as our resultant estimated motion of the southern Sierra Nevada best estimate of Garlock slip rate, which is used to with respectto the point where this path crossesthe describe the relative motion between the southern Sierra California-Nevadaborder (i.e., the Great Basin option, Nevada and the Mojave block. However, this velocity is Figure 7b) is 8.2 +1.8 mm/yr N57øW +7ø (at 90% derived from the location where slip rate estimatesare confidence).Including the deformationto the east yields available,and slip there dependsupon the local kinematic an estimatedvelocity for the Sierra Nevada with respect role played by the Garlock fault. Two important to North Americaof-10 mm/yr N60øW. kinematic processesmay contributeto this velocity: (1) An estimateof Sierra Nevada motion averagedover 15 The Garlock fault may accommodateextension in the m.y. is availableby reconstructinglate Cenozoicoffsets Great Basin, allowing the Sierra Nevada to move relative in the regionbetween the southernSierra Nevada and the to the Mojave in a direction that is parallel to the trend Colorado Plateau [Wernicke et al., 1989]. This of the Garlock fault [Davis and Burchfiel, 1973]. calculationyields an averageextension rate of 16.7 +4.5 Because neither the rotation of the Mojave [Wells and mm/yr N73øW +12ø, which Wemicke et al. use to argue Hillhouse, 1989] nor the Sierra Nevada [Argus and that the extension rate has slowed and reoriented, with Gordon, 1990, 1991] is thought to occur at rates great extensionin the last -4 m.y. concentratedin the region enoughto contributesignificantly to Garlock fault slip, west of the California-Nevada border and in a more the slip rate of the Garlock fault where it separatesthe northerlydirection. Sierra Nevada from the Mojave is thought to represent Mojave/Gariock option. An alternativeSierra Nevada this role. Figure 13 showsthis in a North America frame block velocityestimate is obtainedby leavingthe original of reference: the displacement across the eastern Sierra Nevada path in southernNevada, crossingthe California shearzone causesthe westernMojave to move southernNevada portionof the intermontainseismic belt approximately to the north, while the addition of Great and the NE trending,left-lateral Lake Mead fault system, Basin extensionnorth of the Garlock fault producesa following the ColoradoRiver to California, headingwest northwest direction of motion of the Sierra Nevada acrossthe Mojave, and finally headingnorth acrossthe (Figure 13). (2) Garlock fault activity may be involved Garlock fault to the Sierra Nevada. The active fault with deformationoccun•ng across the easternCalifornia systemsin southernNevada are complex, expressing shearzone, which trendsthrough the regionof the eastern normal, strike-slip and thrust styles of deformation Garlock fault [Dokka and Travis, 1990]. In this role, the [Andersonand Barnhard, 1993]. This region is one of Garlock fault would be one of several left-lateral west transitionbetween the active Great Basin and the nearly trending faults in the NW Mojave that accommodate inactive southernBasin and Range, and these faults clockwise rotation of blocks they bound. Hence this probablyhelp accommodatethis difference. Anderson componentof Garlock slip would be confined to the and Barnhard [1993] find a N105øW direction of region where importantblock rotationis occurring,which. extension in the Mormon Mountains portion of the presumablyis defined by the longitudinalrange of the intermountainseismic belt (in eastern Nevada, west of west trending faults in the NW Mojave that lie south of the Utah-Arizona border), which is similar to the overall the SW Great Basin. trend in seismicitythat definesthis belt [Smith, 1978]. The Garlock fault slip rate estimatethat we use (i.e., The Lake Mead fault systemhas been active in the last 5 the values of-7 mm/yr) comes from an area that lies m.y., offsettingthe now inactive I_as Vegas shear zone west of the mapped west trending faults in the NW [Andersonet al., 1972; Bohannon,1984] and producing Mojave that is near the section of fault that boundsthe scarpsin Pleistocenealluvium. Its trendis approximately southernSierra Nevada. We thereforethink that this slip N130øW. To allow for the possibility of significant rate estimatedfor the Garlock fault receives relatively activityon thesezones, we assigna velocityof 1* +1* little contributionfrom deformationoccurring across the mm/yrN65øE * +_25ø*. Thisorientation is intermediate to eastern California shear zone and probably represents 20,004 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION

Sierra Nevada-Mojave relative motion well. We use the Garlock estimatedslip rate (of 7 mm/yr) is a result of trend of the Garlock south of the southern Sierra Nevada easternCalifornia shearzone activity. Alternatively,our (N125øW) to give the orientation of the relative motion assumption(in the Great Basin path) that the orientation vector. However, because the location from which this of deformationacross the CaliforniaGreat Basin is given estimate is derived lies east of the Sierra Nevada and by whatare inferredto be purelyslrike-slip elements may becausethe fault trend there is more westtrending than it be in error. Kinematic information on the strike-slip is in the region where the Garlock fault more clearly elements is well documented, but additional, west representsSierra Nevada-Mojave relative motion, we are directed extension may possibly have been neglected, not certain that some of the estimatedslip is not related thusomitting a westerlycontribution to the SierraNevada to activity associatedwith the eastern California shear velocity estimate. If such extension occurs, and the zone. westernGatlock fault slip rate is near the estimatedrate To handle this problem, we first assume that our (Table 1), then the deformation field is more similar to velocity estimateof 7 mm/yr N125øW describessouthern the "intercontinentaltransform" kinematics described by Sierra Nevada-Mojave relative motion, and we assign Davis and Burchfiel [1973] than the "pull-apart"kine- uncertainty limits under this assumption. We then matics presentedby Burchfiel and Stewart [1966] and consider that some componentof Garlock slip at the Burchfiel et al. [1987]. It is also possiblethat some of location of the rate estimate is due to eastern California the difference between the two estimates is accom- shearzone activity, and becausethis correctionis small, modatedby complex deformationin the southernSierra we simply modify our previousestimate to accommodate Nevadathat we have not considered.If this complication this. Uncertainty limits on the rate are made wide is occurring, then the Great Basin estimate of Sierra enoughto includea minimum rate of 3 mm/yr (given by Nevada motion would be more representativeof Sierra the minimum creep rate estimatefor the westernGarlock Nevada/Great Valley motion as a whole becausethe fault [Louie et al., 1985]) and a maximum of 11 mm/yr Great Basin path entersthe Sierra Nevada north of this (inferred from the initial work of Carter [1980, 1982]). possiblezone of deformation. The observed trend of the Garlock fault where it We take the product of the two estimatesto obtain a separatesthe southernSierra Nevada from the Mojave best velocity estimatefor southernSierra Nevada motion varies by +3ø and -8ø. This defines our uncorrected (relative to North America), yielding 8.6 +2.0 mm/yr velocity estimate. N51øW _+9ø (Figure 7c). This estimateis similar to the If eastern California shear zone deformation contributes southern Great Basin velocity estimate because the uncertaintiesassociated with this path are smaller than to the velocity, the rate shouldsimply be subtractedfrom the existing rate estimate so the deformation is not those associatedwith the Mojave/Garlockpath. The included twice (the orientation should not be affected estimatedSierra Nevada velocityis inconsistentwith that because eastern California shear zone deformation is determinedgeodetically by Argus and Gordon [1990, accommodated by faulting and rotation yielding a 1991] (11 +2 mm/yr oriented N28øW +6% with velocity estimatedand applied above; slip on the Gatlock uncertaintiesat two standarderrors), is marginally fault is assumedhere to be a part of this faulting). To consistentwith the estimatesof Dixon et al. (submitted include this possibility, we reduce our best rate estimate manuscript,1994) (-10 _+0.6mm/yr orientedN41øW +5ø) of 7 mm/yr by 2 mm/yr. There is no changein either the and Gordon et al. [1993] (12.3 mm/yr _+0.6mm/yr upper rate limit (becauseeastern California shear zone orientedN46øW +_3ø for station OVRO, which is close to deformationmay not contributeto slip at the site where the Sierra Nevada). The principal differencesamong the velocity estimate is derived) or the lower rate limit these models are in the assumptionsof what stations define stable North America. Coecauseit represents data derived from where the Sierra Nevada rotation. In order to cross the Sierra Gatlock bounds the southern Sierra Nevada). Dimin- ishing the most likely rate estimateby 2 mm/yr without Nevada block, we first must estimate its rotation rate. changingthe uncertaintylimits reducesthe centroid of Without well-constrainedgeologically based velocity the rate probability distributionby less than 1 mm/yr. estimatesfor the northernand southernSierra Nevada, we The final velocity estimate for the relative motion obtainonly roughlimits. Perhapsa moreuseful approach between the southern Sierra Nevada and the Mojave is to usethe rotationrate estimatedby Argusand Gordon block is 5 t mm/yr +6-2 t mm/yr orientedN125 o W t +3 o- [1990, 1991], based on VLBI geodetic results. At 8øt. 0.6ø/m.y.counterclockwise, their rate is greaterthan our g•logically inferredrate (whichis indistinguishablefrom Estimating the southern Sierra Nevada velocity zero) yet still contributesrelatively little influenceon the (relative to point where the path crossesthe California- net Pacific-NorthAmerica velocity estimate. A rotation Nevada border) by using features found along this rate of 0.6ø/m.y.yields a southdirected velocity of 1.5 alternatepath yields a southernSierra Nevada velocity mm/yr for a path crossingthe SierraNevada-Great Valley (for the Mojave/Garlock option, Figure 7a) of 9.7 +_2.5 block. Becausethis estimatedepends on the velocityof mm/yr N62øW _+19ø (at 90% confidence). This result is a geodetic site within the Owens Valley (OVRO) that about 1.7 mm/yr greaterin a westerlydirection than that excludes(low) rates of normal activity on the frontal obtainedby the Great Basin option, i.e., the path that lies range of the Sierra Nevada, we think that this estimate north of the Gatlock fault yields a more westerly represents an upper limit of the rotation rate. To velocity. This difference suggests that the actual representthe effects of Sierra Nevada rotation, we choose contributionto the Mojave path option by the Gatlock a vector1.0' +0.5-1.0* mm/yroriented south * +35 ø *. fault may be toward the lower limit of its uncertainty Contraction near the San Andreas fault. We infer distribution,which in turn suggeststhat most of the that little deformationis encounteredin crossingthe HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION 20,005

Sierra Nevada and the San Joaquin Valley until the geodetic estimates of the rate of strain accumulation westernmargin of the San JoaquinValley is encountered. [Thacher, 1979; Savage, 1983; King et al., 1987]. The Contractionnorth of the TransverseRanges essentially averagetrend of the San Andreasfault throughthis part perpendicularto the San Andreasfault is expressedby of its courseis N4 løW* +4 ø*. the presence of folds and earthquakeswith reverse Of the strike-slip faults to the west, only the San mechanismslike the 1983 Coalingaevent and by offshore Simeon strand of the Hosgri fault system has well- thrusting[Crouch et al., 1984]. The best evidencefor the documentedlate Quaternaryslip rates.Weber and Lajoie magnitude of shortening comes from retrodeformable [1977] reporteda rate of 6 to 13 mm/yr. More recent crosssections through the entire California coastalmargin studiesof the sameoffset marine terraces suggest a lower [Namson and Davis, 1988]. For this reason we include all rate of 4 mm/yr +6-2 mm/yr [Hansonet al., 1992; Hall of the San Andreasnormal contractionin a single value et al., 1987]. Becausethe San Simeonfault is only one derived from cross sectionsacross the entire region. strandof the Hosgri zone and the Hosgri is only the most Namsonand Davis [1988] reportapproximately 33 km of active of the faults west of the San Andreas, we infer a contractionalong our path between the San Joaquin rateof 6' +6-4' mm/yr. The limitsare basedon the Valley and the Pacific plate, normal to the San Andreas lowest and highest proposedrates for the San Simeon fault. It is difficult for us to evaluatean uncertaintyin fault alone,but appearto be wide enoughto encompass net shorteningfrom the reconstructionsof Namson and the entire deformation west of the San Andreas fault. Davis, but we infer it is minor relative to the associated The orientationof the Hosgrifault systemat San Simeon large age uncertainty. Namsonand Davis [1988] refer to isN40øW *, thoughindividual traces deviate by +15 ø *. the shorteningas "late Cenozoic"but appearto prefer an Convolvingthe probabilitydistributions associated with initiation to the major phaseof shorteningnear 2.3 Ma. this Sierra Nevada path resultsin an end-of-pathvector This gives a rate that we believe is too great (-14 of 48 +4 mm/yr N41øW +4ø (at 90% confidence),which mm/yr). We infer from their sectionsthat the defor- is shownin Figure 6. mationbegan in the late Miocene;this is supportedby a late Miocene onsetfor shorteningnear the coastalong a Acknowledgments. We thank Andrew Oriscom and James similar line [Davis and Mcintosh, 1987]. Using an Savagefor thoughtfuland carefulreviews. Don Argus,Tim initiationdate of 10 Ma, we get an estimatedshortening Dixon, Roy Dokka, Ken Dueker, Richard Gordon, Randy rate of 3.3 mm/yr. However,it has also been suggested Palmer, and Bob Powell are thanked for discussions about the nature and implicationsof deformationin the southwestU.S. that much of the late Cenozoicshortening observed in a We thankDon Argusand Tim Dixon for providingresults on broadregion of the CoastRanges where our path crosses NUVEL-1A and SierraNevada motion (respectively)prior to occurredwhen this countrywas in the vicinityof the Big publication.Beverly Effmgerand Ann Blanchardconverted our Bend [Weldonand Humphreys,1989]. Consideringthis, sketchesinto quality illustrations. This work occurred over we choosea lowerrate of 2' mm/yras ourmost likely many years, supportedby grants 14-08-0001-G1780, 1434-92- rate, and •use of the uncertaintyin the timing of G-2211 and 1434-93-G-2288 fxom the USGS, EAR-905714 deformationand the possibilitythat the Quaternaryrate is fxom the NSF, and NAG 5-755 and NAG 5-1903 from NASA. greater the late Cenozoic average, we assign limits of +5* and-1.5' mm/yr.This allows, for instance,for 21 km of shorteningprogressing at a steadyrate since3 Ma. References At the lower limit, shorteningoccurs at trivial rates. Allen, C. R., L. T. Silver, and F. G. Stehli, Agua Blanca fault: Argus and Gordon [1990, 1991] consideredVLBI data A major transversestructure of northern Baja California' from eastern California and the Pacific and concluded Mexico, Geol. Soc. Am. Bull., 71, 457-482, 1960. that only a few millimeters per year of contraction Anderson,J. G., Estimatingthe seismicityfrom geological currentlyoccurs in California west of the Sierra Nevada. structurefor seismic-risk studies,Bull. Seismol. Soc. Am., 69, Mount and Suppe[1987], Springer[1987], and Zoback 135-158, 1979. et al. [1987] reporta directionof maximumcompressive Anderson,R. E., T. P. Barnhard,Aspects of three-dimensional stresstypically in the range of N40øE and N60øE. The strain at the margin of the extensionalorogen, Virgin River orientationof shorteninginferred from geologicstructures depressionarea, Nevada, Utah, and Arizona, Geol. Soc. Am. in this region is-N40øE [Mountand Suppe,1987]. We BulL, 105, 1019-1052, 1993. Anderson,R. E., C. R. Longwell, R. L. Armstrong,and R. F. assignan averagevalue of N45øE* for the shortening Marvin, Significanceof K-Ar agesof Tertiary rocks from the direction. We includethe rangeof estimatedorientations Lake Mead region, Nevada/Arizona,Geol. Soc. Am. Bull., 83, from N40øE* and N60øEt. 273-288, 1972. Right-lateral shear on and near the San Andreas Argus, D. F., and R. G. Gordon, Pacific-NorthAmerica plate fault. Right-lateralshear across this portion of the path motion from very long baselineinterferometry compared with is dominatedby the San Andreasand the Hosgri fault motion inferred from magnetic anomalies,transform faults, zones. However, there are two other zones of active and earthquakeslip vectors, J. Geophys.Res., 95, 17,315- seismicitybetween these zones [Poley et al., 1987]. We 17,324, 1990. divide the activity between the San Andreas, which is Argus, D. F., and R. G. Gordon, Current Sierra Nevada-North America motion from very long baseline interferometry: well defined,and an estimateof the net right-lateralshear Implicationsfor the kinematicsof the westernUnited States, to the west. Our path crossesthe San Andreasfault at a Geology,19, 1085-1088, 1991. place wherethe best availableslip rate existsfor the San Astiz, L., and C. R. Allen, Seismicity of the Garlock fault, Andreas fault in central California. Sieh and Jahns California, Bull. Seismol. Soc. Am., 73, 1721-1734, 1983. [1984]have documented a late Holocene slip rate of 34' Atwater, T., Implicationsof plate tectonicsfor the Cenozoic +3* mm/yr.This value is consistentwith creep rates to tectonic evolution of western North America, Geol. Soc. Am. the north [Burford and Harsh, 1980; Schultz,1989] and Bull., 81, 3513-3536, 1970. 20,006 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION

Barrows, A. O., J. E. Kahle, and D. J. Beebe, Earthquake Chapin, C. E., Evolutionof the Rio GrandeRift-A summary,in hazard and tectonic history of the San Andreas fault zone, Rio Grande Riff: Tectonicsand Magmatism, edited by R. E. Los AngelesCounty, California, Open File Rep. Calif. Div. Reicker,pp. 1-5, AGU, Washington,D.C., 1979. Mines Geol., 851 85-10 LA, 236 pp., 1985. C*arlc, M. M., and K. R. Lajoie, Holocene behavior of the Bartley, J. M., A. F. Glazner, and E. R. Schermer,North-south Garlock fault, in Geological Survey Research 1974, U•. contractionof the Mojave block and strike-slip tectonicsin Geol. Surv.Prof., Pap., 900, 184-185, 1974. southern California, Nature, 248, 1398-1401, 1990. Clark, M. M., et al., Preliminary slip rate table and map of late Beanland, S., and M. M. Clark, The Owens Valley fault zone, Quaternaryfaults of California, U•. Geol. Surv. Open File eastern California, and surface rupture associatedwith the Rep., 84-106, 1984. 1872 earthquake,Seisrnol. Res. Lett., 32, 32, 1987. Crouch, J. K., S. B. Bachman,and J. T. Shay, Post Miocene Biasi, G. P., and E. D. Humphreys,P-wave image of the upper compressionaltectonics along the central California margin, manfie smacture of central California and southern Nevada, in Tectonicsand SedimentationAlong the California Margin, Geophys.Res. Lett., 19, 1161-1164, 1992. edited by J. K. Crouch and 5;. B. Bachman, pp. 37-54, Bird, P., and J. Baumgardner,Fault friction, regionalstress, and Society of Economic Paleontologistsand Mineralogists, crust-mantle coupling in southern California from finite PacificSection, Los Angeles,Calif., 1984. elementmodels, J. Geophys.Res., 89, 1932-1944, 1984. Crowell, J. C., and A. G. Sylvester(Eds.), Tectonicsof the Bird, P., and R. W. Rosenstock,Kinematics of presentcrust and Juncture Between the San Andreas Fault System and the mantle flow in southern California, Geol. Soc. Am. Bull., 95, Salton Trough, SoutheasternCalifornia, A Guidebookfor 946-957, 1984. Fieldtrips, Geological Societyof America,Boulder, Colorado, Bohannon,R. G., Strike-slipfaults in the Lake Mead regionof 1979. southernNevada, in CenozoicPaleogeography of the Western Davis, G. A. and B.C. Burchfiel, Gatlock fault: An intra- United States,Pac. CoastPaleogr. Syrup.3, editedby J. M. continental transform smacture, southern California, Geol. Armentrout et al., pp. 129-139, Society of Economic Soc. Am. Bull., 84, 1407-1422, 1973. Paleontologistsand Mineralogists, Pacific Section, Los Davis, T. L., and K. D. Mcintosh, A retrodeformable structural Angeles,Calif., 1979. solution acrossthe southernCoast Ranges and implications Bohannon,G. G., Nonmarinesedimentary rocks of Tertiary age for seismically active smactures,Geol. Soc. Am. Abstr. in the Lake Mead region, southeasternNevada and Programs,19, 385, 1987. northwe.stem Arizona, U.S. Geol. Surv. Prof. Pap., 1259, 69 DeMets, C., R. G. Gordon, D. F. Argus, and S. Stein, Current pp., 1984. platemotions, Geophys. J. lnt., 101, 425-478, 1990. Buising, A. V., Small-scale late Cenozoic faulting in the Dibblee, T. W., Jr., Aerial geology of the westernMojave Colorado River extensional corridor, western Arizona and Desert,California, U.S. Geol. Surv. Prof. Pap., 522, 153 pp., southeasternCalifornia, in Deformation AssociatedWith the 1967. Neogene Eastern California Shear Zone, Southeastern Dibblee, T. W., Jr., Regionalgeology of the TransverseRanges California and Southwestern Arizona, San Bernardino province of Southern California, in Geology and Mineral County,Spec. Publ. 92-1, pp. 11-15, editedby S. M. Richard, Wealth of the California TransverseRanges, Mason Hill San BernardinoCounty Museum, Redlands, Calif., 1992. Volume,edited by D. L. Fife and J. A. Minch, pp.7-26, South Burchfiel,B.C., and J. H. Stewart,The "pull-apart"origin of CoastGeological Society, Santa Ana, Calif., 1982. Death Valley, Geol. Soc.Am. Bull., 77, 439-442, 1966. Dokka, R. K., Displacementson late Cenozoicstrike slip faults Burchfiel,B.C., K. V. Hodges,and L. H. Roy&n, Geologyof of the central Mojave Desert, California, Geology, 11, 305- 308, 1983. Panamint Valley-Saline Valley pull-apart system,California, J. Geophys.Res., 92, 10,422-10,426,1987. Dokka, R. K., and C. J. Travis, Late Cenozoic strike-slip faulting in the Mojave Desert, California, Tectonics,6, 311- Burford, R. O., and P. Harsh, Slip on the San Andreasfault in 240, 1990. central California from alignment array surveys, Bull. Seisrnol.Soc. Am., 70, 1233-1261, 1980. Donnellan,A., B. H. Hager, R. W. King, and T. A. H•g, Geodetic measurementof deformation in the Ventura Basin, Buffer, P. R., B. W. Troxel, and K. L. Verosub, Late Cenozoic southern California, J. Geophys. Res., 98, 21,727-21,739, history and styles of deformationalong the southernDeath 1993. Valley fault zone,Geol. Soc.Am. Bull., 100, 402-410, 1988. Eaton, 13. P., R. R. Wahl, H. J. Prostka,D. R. Mabey, and M. Calderone,G. J., and R. F. Butler, Paleomagnetismof Miocene D. Kleinkopf, Regional gravity and tectonicpatterns: their volcanic rocks from southwestern Arizona: Tectonic relationto late Cenozoicepeirogeny and lateral spreadingin implications,Geology, 12, 627-630, 1984. the westernCordillera, in CenozoicTectonics and Regional Calderone, G. J., (3. D. Acton, and R. F. Buffer, Concordant Geophysicsof the Western Cordillera,edited by R. B. Smith paleomagnefi½directions from Miocene volcanic rocks, and 13. P. Eaton, Mere. Geol. Soc. Am., 152, 51-92, 1978. Colorade River region, Arizona and California (abstract),Eos Eaton, J. E., Decline of the Great Basin, southwestern United Trans. AGU, 67, 922, 1986. States,AAPG Bull., 16, 1-49, 1932. Calderone, G. J., R. F. Buffer, and G. D. Acton, Eberhart-Phillips,D., M. Lisowski, and M.D. Zoback, Crustal Paleomagnefismof middle Miocene volcanic rocks in the strainnear the Big Bend of the San Andreasfault: Analysis Mojave-Sonoran Desert region of western Arizona and of the Los Padres-Tehaehapitrilaterafion networks, California, southeasternCalifornia, J. Geophys.Res., 95, 625-647, 1990. J. Geophys.Res., 95, 12,139-12,153, 1990. Carter, B. A., Quaternarydisplacement on the 13arlockfault, Elders, W. A., R. W. Rex, T. Meidav, P. T. Robinson, and S. California, in Geology and Mineral Wealth of the California Biehler, Crustal spreading in southern California: The Desert, Dibblee Volume, edited by D. L. Fife and A. R. Imperial Valley and the Gulf of Californiaformed by rifting Brown, pp. 457-465, South Coast GeologicalSociety, Santa apartof a continentalplate, Science, 178, 15-24, 1972. Aria, Calif., 1980. Feigl, K. L., R. W. King, and T. H. Jotdon, Geodetic Carter, B. A., Neogene displacementon the Garlock fault, measurement of tectonic deformation in the Santa Maria fold California (abstract),Eos Trans. AGU, 63, 1124, 1982. and thrustbelt, California,J Geophys.Res., 95, 2679-2699, Carter, J. N, B. P. Luyendyk, and R. R. Terres, Neogene 1990. clockwise rotation of the eastern Transverse Ranges, Feigl, K. L, et al., Space geodeticmeasurement of crustal California, suggestedby paleomagneficvectors, Geol. Soc. deformation in central and southernCalifornia, 1984-1992, J. Am. Bull., 98, 199-206, 1987. Geophys.Res., 12, 21,677-21,712, 1993. HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION 20,007

Frizzell, V. A., Jr., J. M. Matrison, and J. C. Matti, Distinctive Hirth, J.P., and J. Lothe, Theory of D/s/ocat/ons,p. 293, John Triassicmegaporphyritc monzogranite: Evidence for only 160 Wiley, New York, 1982. kilometersof offset along the San Andreas fatfit, southern Homafius, J. S., Neogene tectonicrotation of the Santa Ynez California,J. Geophys.Res., 91, 14,080-14,088, 1986. Range, westernTransverse Ranges, California, suggestedby GaffunkeL Z., Model for the late Cenozoic tectonichistory of paloemagneticinvestigation of the Monterey Formation, J. the Mojave Desert, California, Geol. $oc. Am. Bull., 85, Geophys.Res., 90, 12,503-12,522,1985. 1931-1944, 1974. Humphreys, E. D., and K. G. Ducker, We.stem U.S. upper Gillespie, A. R., Quaternaryglacialion and tectonismin the mantlestructure, J. Geophys.Res., 99, 9615-0634, 1994. southeasternSierra Nevada, Inyo County, Cal., Ph.D. Humphreys,E. D., and B. H. Hager, A kinematicmodel for the dissertation,695 pp., Calif. Inst. of Technol.,Pasadena, 1982. late Cenozoic developmentof southernCalifornia crest and Golombek, M.P., Geometry and rate of extensionacross the uppermantle, J. Geophys.Res., 95, 19,747-19,762,1990. Pajaritofault zone, Espanolabasin, Rio Granderift, northern Humphreys,E. D., and R. J. Weldon, Kinematicconstraints on New Mexico, Geology,9, 21-24, 1981. the rifting of Baja California, Chapter 12, in The Gulf and Golombek, M.P., G. E. McGill, and L. Brown, Tectonic and Peninsular Provinces of the Califomias, edited by J.P. geologicevolution of the Espanolabasin, Rio Grande Rift: Dauphin and B. R. T. Simoneit'AAPG Mere., 47, 217-229, Structure, rate of extension, and relation to the state of stress 1991. in the western United States,Tectonophysics, 94, 483-507, Jackson,J., and P. Molnar, Active faulting and block rotations 1983. in the western TransverseRanges, California, J. Geophys. Gordon, D., C. Ma, and J. W. Ryan, Results from the CDP Res., 95, 22,073-22,087, 1990. mobile VLBI program in the western United States, in Jennings,C. W., Fault map of California, with locationsof Contributionsof Space Geodesyto Geodynamics,Geodyn. volcanos,thermal springs, and thermalwells, scale1:750,000, Ser., vol. 23, edited by D. E. Smith and D. L. Turcotte,pp. Calif. Div. of Mines and Geol., Sacramento, 1974. 131-138,AGU, Washington,D.C., 1993. Keller, E. A., M. S. Bonkowski, R. J. Korsch, and R. J. Greensfelder, R. W., F. C. Kintzer, and M. R. Somerville, Shlemon,Tectonic geomorphologyof the San Andreasfault Seismotectonicregionalization of the Great Basin, and zone in the southernlndio Hills, CoachellaValley, California, comparisonof momentrams computedfrom Holocenestrain Geol. Soc. Am. Bull., 93, 46-56, 1982. and historicseismicity, US. Geol. Surv. Open File Rep., 80- Kelley, V. C., Tectonics,middle Rio Granderift, New Mexico, 801, 433-493, 1980. in Rio Grande Riff: Tectonicsand Magmatism,edited by R. Greis, R., North-south compression of Rocky Mountain E. Reicker,pp. 57-70, AGU, Washington,D.C., 1979. Forelandstructures, in RockyMountains Foreland Basins and Kennedy,M.P., Recencyand characterof faulting along the Uplifts, edited by J. D. Lowall, pp. 9-32, Rocky Mountain Elsinore fault zone in southernRiverside County, California, Associationof Geologists,Denver, Colo., 1983. Calif. Div. Mines Geol. Spec.Rep. 131, 12 pp., 1977. Hall, N. T., T. D. Hunt, P. A. Vaughan, and F. R. Bickxer, King, N. E., P. Segall,and W. Prescott,Geodetic measurements Trenchingof the San Simeonfatfit, San Luis ObispoCounty, near Parkfield, California, 1959-1984, J. Geophys.Res., 92, California,Geol. Soc.Am. Abstr.Programs, 19, 385, 1987. 2747-2766, 1987. Hamblin, W. K., P. E. Damon, and W. B. Bull, Estimates of [,egg, M. R., Fault characterand seismotectonicsalong the verticalcrustal strain rates along the westernmargins of the western boundary of the southernCalifornia microplate ColoradoPlateau, Geology, 9, 293-298, 1981. (abstract),Eos Trans. AGU, 69, 1419, 1988. Hamilton,W., Laramidecrustal shortening, in Interactionof the [,egg, M. R., Sea Beam evidenceof recenttectonic activity in Rocky MountainForeland and the CordilleranThrust Belt, the California continental borderland, in The Gulf and editedby C. J. Schmidtand W. J. Perry,Mem. Geol. Soc. Peninsular Provinces of the Califomias, edited by J.P. Am., 171, 27-39, 1988a. Dauphin and B. R. T. Simoneit'AAPG Mere., 47, 179-196, Hamilton, W., Tectonic setting and variationswith depth of 1991. some Cretaceous and Cenozoic structural and magmatic [,egg, M. R., J. Mammerickx,B. P. Luyendyk,C. Demoustier, systemsof the westernUnited States,in Metamorphicand and R. C. Tyce,,Seabeam survey of an activestrike slip fault: CrustalEvolution of the WesternIgnited States, Ruby Volume The San Clemente fault in the California continental VII, edited by W. G. Ernst, pp. 1-40, Prentice-Hall, Borderland,J. Geophys.Res., 94, 1727-1744,1989. EnglewoodCliffs, N.J., 1988b. [,egg, M. R., O. V. Wong, and F. Suarez,Geologic structure Hanson,K. L., W. R. Lettis, J. R. Wesling,K. I. Kelson, and L. and tectonics of the inner continental borderland of northem Mezger, Quammary marine terraces,south-central coastal Baja California,in The Gulf and PeninsularProvinces of the California:Implications for crustaldeformation and coastal Californias,edited by J.P. Dauphin and B. R. T. Simoneit, evolution,in QuaternaryCoasts of the United States:Marine AAPG Mem., 47, 145-177, 1991. and LacustrineSystems, edited by C. H. Fletcherand J. F. Liu, W., Palcomagnetismof Miocene sedimentaryrocks in the Wehmiller, Spec. Publ. Soc. Econ. Paleontol.Mineral., 48, Transverse Ranges; the implications for tectonic history, 323-332, 1992. Ph.D. dissertation, 233 pp., Calif. Inst. of Technol., Hatch, M. E., and T. K. Rockwell, Neotectonicsof the Agua Pasadena, 1990. Blanea fault, Agua Blanca valley, northernBaja California, Louie, J. N., C. R. Allen' D. C Johnson,P. C. Hasse, and S. N. Mexico, Geol. Soc.Am. Abstr.Programs, 18, 114, 1986. Cohn, Fault slip in southernCalifornia, Bull. Seismol.Soc. Hauksson,E., and L. M. Jones,The July 1986 Oceanside(ML Am., 75, 811-833, 1985. 5.3) earfiquake sequencein the continentalborderland, Lubetkin, K. K. C., and M. M. Clark, Late Quaternaryactivity southernCalifornia, Bull. Seismol. Soc. Am., 78, 1885-1906, alongthe Lone Pine fatfit, easternCalifornia, Geol. Soc.Am. 1988. Bull., 100, 755-766, 1988. Hauksson, E., and G. F. Saldivar, Seismicity and active Luyendyk,B. P., M. J. Kamerling, and R. Terms, Geometric compressionaltectonics in Santa Monica Bay, southern model for Neogene crustal rotationsin southernCalifornia, California,J. Geophys.Res., 94, 9591-9606, 1989. Geol. Soc. Am. Bull., 91,211-217, 1980. Hill, M. L., Anomalous trends of the San Andreas fault in the Luyendyk, B. P., M. J. Kamerling, R. Terres, and J. S. Transverse Ranges, California, in Geology and mineral Hornafius, Simple shear of southern California during wealth of the California TransverseRanges, Annu. Symp. Neogene time suggestedby palcomagneticdeclinations, J. Guideb.,I0, editedby D. L. Fife and J. A. Minch, pp. 367- Geophys.Res., 90, 12,454-12,466,1985. 369, SouthCoast Geological Society, Santa .Ana, Calif., 1982. Matti, J. C., D. M. Morton, and B. F. Cox, Distribution and 20,008 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION

geologicrelations of fault systemsin the vicinity of the Powell, R. E., and R. J. Weldon, Evolution of the San Andreas central TransverseRanges, southernCalifornia, U•. Geol. fault, Annu. Rev. Earth Planet. Sci., 20, 431-468, 1992. Surv.Open File Rep.,85-865, 27 pp., 1985. PrescottW. H., M. Lisowski, and J. C. Savage, Velocity field Mayer, L., Topographicconstraints on modelsof lithospheric alongthe San Andreasfault in southernCalifornia (abstract), stretchingof the Basin and Rangeprovince, western United Eos Trans. AGU, 68, 1506, 1987. States, in Extensional Tectonics of the SouthwesternUnited Ramirez,V. R., HungryValley Formation:Evidence for 220 km States:A Perspectiveon Processesand Kinematics,edited by of postMiocene offset on the San Andreasfault, in Tectonics L. Mayer, Spec.Pap. Geol. Soc.Am., 208, 1-14, 1986. and SedimentationAlong the San Andreas Fault System, Meisling, K. E., and R. J. Weldon, Late Cenozoictectonics of edited by D. W. Anderson and M. J. Rymer, pp. 33-44, the northwestern San Bernardino Mountains, southern Pacific Section, Society of Economic Paleontologistsand California, Geol. Soc. Am., Bull., 101, 106-128, 1989. Mineralogists,Los Angeles,1982. Menges,C. M., and P. A. Pearthree,Map of neotectonic(latest Richard,S. M, and R. K. Dokka, Geologyof the PackardWells Pliocene-Quatemary)deformation in Arizona,Open File Rep. fault zone, southeasternCalifornia, in DeformationAssociated 83-22, Ariz. Bur. of Geol. and Min. Technol., Tucson, 1983. With the Neogene Eastern California Shear Zone, Merifielck P.M., T. K. Rockwell, and C. C. Loughman,Slip SoutheasternCalifornia and SouthwesternArizona, San rate on the San Jacintofault zone in the Anza seismicgap, BernardinoCounty Spec. Publ. 92-1, pp. 71-74, editedby S. southernCalifornia, Geol. Soc. Am. Abstr. Programs, 19, M. Richard, San Bernardino County Museum, Redlands, 431432, 1987. Calif., 1992. Mills, G. and P. J. Fischer,Thrusting associatedwith offshore Richard, S. M., and D. R. Shenod, and R. M. Tosdal, Cibola Newport-Inglewood-RoseCanyon zone, California, Geol. Soc. Passfault, southwesternArizona, in DeformationAssociated Am. Abstr. Programs,23, 198, 1991. With the Neogene Eastern California Shear Zone, Minster, J. B., and T. J. Jordan,Present-day plate motions,J. SoutheasternCalifornia and SouthwesternArizona, San Geophys.Res., 83, 5331-5354,1978. BernardinoCounty Spec. Publ. 92-1, pp. 66-70, editedby S. Minster, J. B., and T. J. Jordan, Vector constraints on M. Richard, San Bernardino County Musere, Redlands, Quaternarydeformation of the westernUnited Stateseast and Calif., 1992. west of the San Andreas fault, in Tectonicsand Sedimentation Rockwell, T. K., Soil chronology,geology and neotectonicsof Alongthe CaliforniaMargin editedby J. K. Crouchand S. B. the north-central Ventura basin, California, Ph.D. dissertation, Bachman, pp. 1-16, Society of Economic Paleontologists 424 pp., Univ. of Calif., SantaBarbara, 1983. Mineralogists,Pacific Section, Los Angeles,Calif. 1984. Rockwell, T. K., Neotectonics of the San Cayetano fault, Mount, V. S. and J. Suppe,State of stressnear the San Andreas TransverseRanges, California, Geol. Soc. Am. Bull., 100, fault: Implicationsfor wrenchtectonics, Geology, 15, 1143- 500-513, 1988. 1146, 1987. Rockwell, T. K., E. A. Keler, M. M. Clark, and D. L. Johnston, Namson, J., and T. L. Davis, Seismicallyactive fold and thrust Chronologyand rates of faulting of Ventura River terraces, belt in the San JoaquinValley, centralCalifornia, Geol. Soc. California, Geol. Soc. Am. Bull., 95, 1406-1474, 1984. Am. Bull., 100, 257-273, 1988. Rockwell,T. K., E. A., Keller, and G. R. Dembroff,Quaternary Pacheco, J., and J. Nabelek, Source mechanisms of three rate of folding of the Ventura Avenue anticline, western moderateCalifornia earthquakesof July, 1986, Bull. Seismol. TransverseRanges, southern California, Geol. Soc. Am. Bull., Soc. Am., 78, 1907-1929, 1988. 100, 850-858, 1988. Pampeyan,E. H., T. L. Holzer, and M. M. Clark, Modem Rockwell, T. K., D. L. Lamar, R. S. McElwain, and D. E. ground failure in the Garlock fault zone, Fremont Valley, Millman, Late Holocene recurrent faulting on the Glen Ivy California, Geol. Soc. Am. Bull., 100, 677-691, 1988. North strand of the Elsinore fault, southern California, Geol. Pearthree,P. A., Late Quaternaryfaulting and seismichazard in Soc.Am. Abstr. Programs,17, 177, 1986a. southeasternArizona and adjacentpotions of New Mexico Rockwell, T. K., P.M. Merifield, and C. C. Loughman, and Sonora, Mexico, Open File Rep. 86-8, Adz. Bur. of Holoceneactivity of the San Jacintofault in the Anza seismic Geol. and Min. Technol., Tucson, 1986. gap, southernCalifornia' Geol. Soc. Am. Abstr. Programs,18, Pearthree,P. A., C. M. Menges, and L. Mayer, Distribution, 177, 1986b. recurrence, and possible tectonic implications of late Rockwell, T. K., A. P. Thomas, A. L. Stenson, and R. Crisman, Quaternaryfaulting in Arizona, Open File Rep. 83-20, Adz. Active northeast trending left-lateral faults between the Bur. of Geol. and Min. Technol., Tucson, 1983. Elsinore and San Jacinto fault zones, southern California' Pezzopane,S., and R. J. Weldon H, Tectonic role of active Geol. Soc.Am. Abstr. Programs,21, 135, 1989. faultingin centralOregon, Tectonics, 12, 1140-1169, 1993. Rust, D. J., Radiocarbon dates for the most recent large Pinault, C. T., and T. K. Rockwell, Rates and sense of prehistoricearthquake and for late Holocene slip rams: San Holocenefaulting on the Elsinorefault: furtherconstraints on Andreasfault in part of the TransverseRanges north of Los the distribution of dextral shear between the Pacific and North Angeles,Geol. Soc.Am. Abstr. Programs,14, 229, 1982. American plates, Geol. Soc. Am. Abstr. Programs, 16, 624, Ryan, J. W., C. Ma, and D. S. Caprette,NASA spacegeodesy 1984. program-GSFCdata analysis-1992, NASA Tech. Memo. TM- Poley, C. M., J.P. Eaton, and A. G. Lindh, Recent seismicityof 1O4572, 1993. the central California region from San Francisco to the Salyards,S. L., Dating and characterizinglate Holocene TransverseRanges, Geol. Soc.Am. Abstr.Programs, 19, 440, earthquakesusing palcomagnetics, Ph.D. dissertation,225 pp., 1987. Calif. Inst. of Technol., Pasadena' 1988. Ponti, D. J., J. A. Barron, A.M. Sama-Wojcicki,M. L. Cotton, Salyards,S. L., K. E Sieh, and J. L. Kirschrink,Palcomagnetic C. Hummon, and C. L. Schneider, Benthie foram measurement of nonbrittle coseismic deformation across the biGstratigraphyand the age of the Pico formation in the San Andreas fault at Pallett Creek, J. Geophys.Res., 97, northernLos AngelesBasin: Potentialproblems for evaluating 12,457-12,470, 1992. activity of blind thrust faults (abstrac0,Eos Trans. AGU, 74 Sauber J., Geodetic measurementsof deformationin California' (43), Fall Meeting suppl.,434, 1993. NASA Tech.Memo. TM-100732, 212 pp., 1989. Powell, R. E., Geology of the crystallinebasement complex, Sauber, J., W. Thacher, and S. C. Solomon, Geodetic easternTransverse Ranges, southern California: constraints on measurementsof deformationin the centralMojave Desert,J. regional tectonic interpretation,Ph.D. dissertation,441 pp., Geophys.Res., 91, 12,683-12,693,1986. Calif. Inst. of Technol., Pasadena, 1981. Saucier,F., and E. Humphreys,Horizontal crustal deformation HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION 20,009

in southernCalifornia from joint modelsof geologicand very the Western Cordillera, edited by R. B. Smith and G. P. long baselineinterferometry measurements, in Contributions Eaton, Mere. Geol. Soc. Am. 152, 1-13, 1978. of Space Geodesyto Geodynamics,Geodyn. Ser., vol. 23, Stewart, J. H., Tectonics of the Walker Lane belt, western Great editedby D. E. Smith and D. L. Turcotte,pp. 139-176, AGU, Basin: Mesozoic and Cenozoic deformation in a zone of Washington,D.C., 1993. shear,in Metamorphismand CrustalEvolution of the Western Saucier, F., E. Humphreys, and R. Weldon, Stress near UnitedStates, Rubey Volume VII, editedby W. 13.Ernst, pp. geometricallycomplex faults: Application to the San Andreas 683-713, Prentice-Hall,Englewood Cliffs, N.J., 1988. fault at Cajon Pass,southern California, J. Geophys.Res., 97, Sylvester,A. G., and R. R. Smith, Tectonictranspression and 5081-5094, 1992. basement-controlled deformation in the San Andreas fault Savage, J. C., Strain accumulationin western United States, zone, SaltonTrough, California,AAPG Bull., 60, 2081-2102, Annu. Rev. Earth Planet. Sci., 11, 11-43, 1983. 1976. Savage,J. C., Equivalentstrike-slip earthquake cycles in half- Terres, R. R., Palcomagnetismand tectonicsof the centraland space and lithosphere-asthenosphereEarth models, J. eastern Transverse Ranges, southern California, Ph.D. Geophys.Res., 91, 4873-4879, 1990. dissertation, Univ. of Calif., Santa Barbara, 1984. Savage, J. C., and M. Lisowski, Deformation in the Owens Terres,R. R., and B. P. Luyendyk,Neogene tectonic rotation of Valley, California, Bull. Seismol.Soc. Am., 70, 1225-1232, the San Gabriel region, California, suggested by 1980. paloemagnetievectors, J. Geophys.Res., 90, 12,467-12,484, Savage, J. C., J.P. Church, and W. H. Prescott,Geodetic 1985. measurementof deformation in Owens Valley, California, Thacher, W., Systematicinversion of geodeticdata in central Bull. Seismol. Soc. Am., 65, 865-874, 1975. California,J. Geophys.Res., 84, 2283-2295, 1979. Savage,J. C., M. Lisowski, and W. H. Prescott,An apparent Troxel, B. W., and L. A. Wright, Tertiary extensionalfeatures• shearzone trendingnorth-northwest across the Mojave Desert Death Valley region, eastem California, in CentennialField into Owens Valley, easternCalifornia, Geophys.Res. Lett., Guide, 1, edited by M. L. Hill, pp. 121-132, Geological 17, 2113-2116, 1990. Societyof America, Boulder, Colo., 1987. Schug, D. L., T. K. Rockwell, and M. E. Hatch, Slip rate Tweto, O., The Rio Grande Rift systemin Colorado,in R/o estimatefor the westernreach of the Agua Blanca fault, Baja Grande Riff: Tectonicsand Magmatism, edited by R. E. California,Mexico, Geol. Soc.Am. Abstr. Programs,19, 448, Reicker,pp. 33-56, AGU, Washington,D.C., 1979. 1987. Vaughan, P., and T. K. Rockwell, Alluvial stratigraphyand Schultz, S.S., Catalog of creep meter measurementsin neotectonicsof the Elsinore fault zone at Agua Tibia California from 1966 through 1988, U.S. Geol. Surv. Open Mountain,southern California, in Neotectonicsand Faulting File Rep., 89-650, 1989. in Southern California, Field Trip Guides 10, 12, 18, Schwartz, D. P., and K. Coppersmith,Fault behavior and compiledby P. L. Ehlig, pp. 177-191, CordilleranSection characteristicearthquakes: Examples from the Wasatch and GeologicalSociety of America,Boulder, Colo., 1986. San Andreas fault zones, J. Geophys.Res., 89, 5681-5698, Wallace,R. E., Patternsand timing of late Quaternaryfaulting 1984. in the Great Basin Province and relation to some regional Schwartz,D. P., and R. J. Weldon, Late Holocene slip rate on tectonicfeatures, J. Geophys.Res., 89, 5763-5769, 1984. the Mojave segment of the San Andreas fault zone, Ward, S. N., Pacific-NorthAmerica plate motions'new results California: Preliminaxyresults (abstract),Eos Trans. AGU, from very long baselineinterferometry, J. Geophys.Res., 95, 67, 906, 1986. 21,965-21,981, 1990. Sharp,R. V., San Jacintofault zone in the PeninsularRanges of Weber, (3. E., and K. R. Lajoie, Late Pleistoceneand Holocene southernCalifornia, Geol. Soc. Am. Bull., 78, 705-729, 1967. tectonicsof the San Gregoriofault zone betweenMoss Beach Sharp,R. V., Variable ratesof late Quaternarystrike slip on the and Point Ano Nuevo, San Mateo County, California,Geol. San Jacintofault zone, SouthernCalifornia, J. Geophys.Res., Soc.Am., Abstr.Programs, 9, 524, 1977. 86, 1754-1762, 1981. Weldon, R. J., The late Cenozoic geology of Cajon Pass; Sheffels, B., and M. McNutt, Role of subsurface loads and implicationsfor tectonicsand sedimentationalong the San regional compensation in the isostatic balance of the Andreasfatfit, Ph.D. dissertation,382 pp., Calif, Inst. of TransverseRanges, California: Evidence for intracontinental Technol., Pasadena, 1986. subduction, J. Geophys. Res., 91, 6419-6431, 1986. Weldon, R. J., and E. D. Humphreys,A kinematicmodel of (Correction,J. Geophys.Res., 92, 6444, 1987.) southernCalifornia, Tectonics,5, 33-48, 1986. Sieh, K., Slip rate acrossthe San Andreasfault and prehistoric Weldon, R. J., and E. D. Humphreys,Comment on "Structural earthquakesat Indio, California (abstrac0,Eos Trans. AGU, transectof the westernTransverse Ranges" by J. Namsonand 67, 1200, 1986. T. Davis, Geology,17, 769-770, 1989. Sieh, K. E., and R. Jahns,Holocene activity of the San Andreas Weldon, R. J., and K. E. Sieh, Holocenerate of slip and fattit at Wallace Creek, California, Geol. Soc. Am. Bull., 95, tentativerecurrence interval for large earthquakeson the San 883-896, 1984. Andreasfault in Cajon Pass, southernCalifornia, Geol. Soc. Smith, R. B., Seismicity, crustal structure, and intraplate Am. Bull., 96, 793-812, 1985. tectonicsof the interior of the western Cordillera, in Cenozoic Weldon, R. J., and J. E. Springer,Active fatfiringnear the Tectonicsand RegionalGeophysics of the WesternCordillera, Cajon Pass well SouthernCalifornia: Implications for the editedby R. B. Smith and G. P. Eaton,Mere. Geol. Soc.Am., stressorientation near the San Andreasfault, Geophys.Res. 152, 111-144, 1978. Lea., 15, 993-996, 1988. Smith, R. B., and M. Sbar, Contemporarytectonics and Weldon, R. J., D. S. Winston, J. L. Kirschvink, and D. W. seismicityof the westernUnited Stateswith emphasison the Burbank,Magnetic stratigraphyof the Crowder Formation, Intermontain seismic belt, Geol. Soc. Am. Bull., 85, 1205- Cajon Pass, Southern California, Geol. Soc. Am. Abstr. 1218, 1974. Programs,16, 689, 1984. Springer,J. E., Stressorientations from wellbore breakoutsin Weldon,R. J., K. E. Meisling,and J. Alexander,A speculative the Coalingaregion, Tectonics, 6, 667-676, 1987. history of the San Andreas fault in the central Transverse Steiner, M. B., Rotation of the Colorado Plateau, Tectonics, 5, Ranges, California, in The San Andreas Fault System: 649-660, 1986. Displacement,Palinspasfic Reconstruction, and Geologic Stewart,J. H., Basin-rangestructure in westernNorth America: Evolution, edited by R. E. Powell et al. , Mere. Geol. Soc. A review, in CenozoicTectonics and RegionalGeophysics of Am., 178, 161-198, 1993. 20,010 HUMPHREYS AND WELDON: WESTERN U.S. DEFORMATION

Wells, R. E., and J. W. Hillhouse,Palcomagnetism and tectonic Zhang, P., M. Ellis, D. B. Slemmons,and F. Mao, Right-lateral rotation of the lower Miocene Peach SpringsTuff, Colorado displacementsand the Holocene slip rate associatedwith Plateau, Arizona, to Barstow, California, Geol. Soc. Am., prehistoricearthquakes along the southernPanamint Valley Bull., 101, 846-863, 1989. fault zone: Implications for southern Basin and Range Wernicke, B., 13. J. Axen, and J. K. Snow, Basin and Range tectonics and coastal California deformation,J. Geophys. extensionaltectonics at the latitude of Las Vegas, Nevada, Res., 95, 4857-4872, 1990. Geol. Soc. Am., Bull., I00, 1738-1757, 1989. Ziony, I. I., and R. F. Yerkes, Fault slip-rateestimation for the Winston, D. S., Magnetic stratigraphy of the Crowder Los Angeles region: Challenges and opportunities, Formation,southern California M.S. thesis,100 pp., Univ. of EarthquakeNotes, 55, 8, 1984. South.Calif., Los Angeles,1985. Zoback, M. L., State of stress and modem deformation of the Withjack, M. O., and W. R. Jamison,Deformation produced by northern Basin and Range province, J. Geophys.Res., 94, obliquerifting, Tectonophysics,126, 99-124, 1986. 7105-7128, 1989. Wright, L. A., J. K. Otton, and B. W. Troxel, Turtleback Zoback,M. L., R. E. Anderson,and G. A. Thompson,Cenozoic surfacesof Death Valley viewed as phenomenaof extension, evolutionof the state of stressand style of tectonismof the Geology,2, 53-54, 1974. Basin and Range provinceof westernUnited States,Philos. Yeats, R. S., Quaternary flake tectonics of the California Trans. R. Soc. London A, 300, 407-434, 1981. TransverseRanges, Geology, 9, 16-20, 1981. Zoback, M.D., M. L. Zoback, V. S. Mount, I. Suppe,I. P. Yeats, R. S., Large-scaleQuaternary detachments in Ventura Eaton,J. H. Healy, D. Oppenheimer,P. Reasenberg,L. Jones, Basin, southernCalifornia, J. Geophys.Res., 88, 569-583, C. B. Raleigh, I. 13. Wong, O. Scotti, and C. Wentworth, 1983. New evidence on the state of stress of the San Andreas fault Yeats, R. S., Late Quaternaryslip rate on the Oak Ridge fault, system,Science, 238, 1105-1111, 1987. TransverseRanges, California: Implicationsfor seismicrisk, J. Geophys.Res., 93, 12,137-12,149, 1988. E. D. Humphreys and R. I. Weldon 11, Departmentof Yeats, R. S., Convergingmore slowly, Nature, 366, 299-301, 1993. GeologicalSciences, University of Oregon,Eugene, OR 97403. (e-mail: gene• newben3,.uoregon.edu) Yeats, R. S., and B. U. Haq, Deep-sea drilling off the Californias: Implications of leg 63, Initial Rep. Deep Sea Drill. Proj., 63, 949-961, 1981. Yeats, R. S., G. J., Huftile, and F. B. Grigsby,Oak Ridge fault, Ventura fold belt, and the Sisar decollement, Ventura Basin, (ReceivedJuly 29, 1993; revisedMarch 28, 1994; California, Geology, 16, 1112-1116, 1988. acceptedMarch 31, 1994.)