Rates of extension along the Fish Lake Valley fault and transtensional deformation in the Eastern California shear zone–Walker Lane belt

The Eastern California shear zone–Walker Lane belt is an evolving fault system east of the San Andreas fault that accommodates ∼20%–25% (9–10 mm/yr) of Pacific–North America plate-boundary deformation (e.g., Dokka and Travis, 1990b; Humphreys and Weldon, 1994; Hearn and Humphreys, 1998; Thatcher et al., 1999; Dixon et al., 2000, 2003; McClusky et al., 2001; Miller et al., 2001; Bennett et al., 2003; Faulds et al., 2005; Wesnousky, 2005a, 2005b; Frankel et al., 2007a). The Eastern California shear zone extends northward for almost 500 km from the San Andreas fault through the Mojave Desert and along the eastern side of the Sierra Nevada. In the Mojave section of the shear zone, fault motion is primarily right-lateral, with a subordinate component of north-south shortening, and slip is localized along several major north-northwest–striking faults (Dokka, 1983; Bartley et al., 1990; Dokka and Travis, 1990a, 1990b; Schermer et al., 1996; Walker and Glazner, 1999; Glazner et al., 2002; Oskin et al., 2007a, 2007b, 2008). North of the active left-lateral Garlock fault, motion is accommodated on four major fault systems: the Owens Valley, Panamint Valley–Saline Valley–Hunter Mountain, Death Valley–Fish Lake Valley, and Stateline fault zones (Fig. 1; e.g., Beanland and Clark, 1994; Lee et al., 2001, 2009a, 2009b; Oswald and Wesnousky, 2002; Kylander-Clark et al., 2005; Bacon and Pezzopane, 2007; Frankel et al., 2007a, 2007b, 2008a, 2008b; Guest et al., 2007; Kirby et al., 2008). In the northern part of the Eastern California shear zone, between latitude 37°N and 38°N, dextral motion between the stable Sierra Nevada block and North America is distributed from Long Valley caldera in the west to the Silver Peak–Lone Mountain extensional complex in the east (e.g., Burchfiel et al., 1987; Brogan et al., 1991; Berry, 1997; Reheis and Sawyer, 1997; Hearn and Humphreys, 1998; Dixon et al., 2000; Oldow et al., 2001; Petronis, 2005; Kirby et al., 2006; Frankel et al., 2007a, 2007b; Lee et al., 2009b). Multiple northeast-striking, down-to-the-northwest normal faults transfer slip between the right-lateral Owens Valley, Panamint Valley–Saline Valley–Hunter Mountain, and Death Valley–Fish Lake Valley fault systems (Burchfiel et al., 1987; Applegate, 1995; Dixon et al., 1995; Reheis and Dixon, 1996; Lee et al., 2001, 2009a; Oswald and Wesnousky, 2002; Walker et al., 2005; Andrew and Walker, 2009). North of the Mina Deflection, strain is accommodated by a series of right-lateral faults as part of the Walker Lane belt (e.g., Nielsen, 1965; Stewart, 1988; Oldow et al., 1989, 2001; Oldow, 2003; Wesnousky, 2005a, 2005b).

Although dextral shear accommodates most slip on the northwest-striking faults in Fish Lake Valley, fault segments that strike more northerly exhibit a relatively large extensional component of slip. Space-based geodetic studies demonstrate that the current strain field is consistent with transtensional deformation and indicate an extension rate of ∼1 mm/yr oriented normal (i.e., N65°E) to the predominant ∼N25°W strike of the Eastern California shear zone at this latitude (e.g., Savage et al., 2001; Bennett et al., 2003; Wesnousky, 2005a). This extension rate in the northern Eastern California shear zone is most likely distributed among the Sierra Nevada frontal fault, the faults of the Volcanic Tablelands, the White Mountains fault, and the Fish Lake Valley fault system (Fig. 1). The geologically determined component of the extension rates on the Sierra Nevada frontal fault (Berry, 1997; Le et al., 2006) and the White Mountains fault (Kirby et al., 2006) are both ∼0.2–0.3 mm/yr, and the rate across the distributed normal faults of the Volcanic Tablelands is ∼0.3 mm/yr (Sheehan, 2007). Thus, the three main fault systems to the west of the Fish Lake Valley fault collectively accommodate approximately half of the current rate of extension within the northern part of the Eastern California shear zone.

In this paper, we investigate how much of the remaining extension is accommodated by the Fish Lake Valley system. Specifically, herein we report new observations from our analysis of high-resolution digital topographic airborne laser swath mapping (light detection and ranging [LiDAR]) data, as well as terrestrial cosmogenic nuclide (TCN) dating of faulted landforms along the Fish Lake Valley fault. Our results provide new, geochronologically determined, late Pleistocene–Holocene extension rates on this fault system that clarify how motion in the Eastern California shear zone is accommodated and, ultimately, transferred northward to the Walker Lane belt. Moreover, these results allow us to assess the degree to which short-term geodetic and longer-term geologic rates are consistent, and the extent to which deformation is accommodated by slip on geomorphically well-defined faults versus distributed, off-fault deformation.

The Fish Lake Valley fault, which forms the northern 80 km of the Death Valley–Fish Lake Valley fault system, is marked by steep, east-facing fault scarps, ponded drainages, and shutter ridges indicative of recent fault activity (Sawyer, 1990; Brogan et al., 1991; Reheis, 1992; Reheis et al., 1993, 1995; Hooper et al., 2003; Frankel et al., 2007a, 2007b, 2008b). The southern and central sections of the Fish Lake Valley fault strike predominantly north-northwest, whereas the northern part of the fault is characterized by numerous north- and northeast-striking strands that splay out into Fish Lake Valley from the main north-northwest–trending, range-bounding fault (Fig. 2; Sawyer, 1991; Reheis and Sawyer, 1997).

Right-lateral motion on the Fish Lake Valley fault is thought to have begun ∼10 m.y. ago (Reheis and Sawyer, 1997), and the strike-slip rate averaged over late Pleistocene–Holocene time is 2.5–3 mm/yr (Frankel et al., 2007b). The extensional component of oblique-normal-dextral motion, responsible for the opening of Fish Lake Valley, most likely began ∼5 m.y. ago, as suggested by the observation that the bounding faults in the northern section of the valley cut across sedimentary rocks of the Miocene Esmeralda basin (Reheis and Sawyer, 1997; Petronis et al., 2002, 2009; Petronis, 2005). On the west side of the White Mountains, the right-lateral White Mountains fault originated later, at ca. 3 Ma (Stockli et al., 2003). The late Pleistocene–Holocene White Mountains fault oblique-slip rate is ∼0.9 mm/yr, parallel to a net slip vector plunging ∼20° toward N10°–20°W (Kirby et al., 2006).

The focus areas of our study are normal fault scarps formed in four alluvial fans along the Chiatovich Creek, Dyer, and Oasis sections of the fault, which have been mapped in detail by a number of researchers (Fig. 2; Brogan et al., 1991; Reheis, 1992; Reheis et al., 1993, 1995). These four alluvial fans were deposited along the eastern White Mountains piedmont at the mouths of (from south to north) Furnace Creek, Wildhorse Creek, Perry Aiken Creek, and Indian Creek. We refer to each of our study sites relative to the respective creek that formed them. At two of the sites, Furnace Creek and Indian Creek, Frankel et al. (2007b) measured large (180–290 m) dextral offsets of alluvial surfaces that they dated with cosmogenic ¹⁰Be surface exposure geochronology at 94 ± 11 ka and 71 ± 8 ka, respectively. Their resulting right-lateral strike-slip rates at these two locations are 3.1 ± 0.4 mm/yr at Furnace Creek and 2.5 ± 0.4 mm/yr at Indian Creek. Multiple normal fault scarps are also present in both of these alluvial fans, as we discuss in the following sections.

Our other two study sites, Wildhorse Creek and Perry Aiken Creek, are located between the Furnace Creek and Indian Creek fans (Fig. 1). Both the Wildhorse Creek and Perry Aiken Creek sites have numerous normal fault scarps. The tallest single normal fault scarp (minimum vertical component of displacement of 73.5 m) along the entire Death Valley–Fish Lake Valley fault zone is found just north of Perry Aiken Creek (Reheis et al., 1993; Reheis and Sawyer, 1997). Reheis et al. (1993) used tephrochronology to suggest that alluvial fans at Wildhorse Creek and Perry Aiken Creek were deposited during middle to late Pleistocene time.

We used terrestrial cosmogenic nuclide geochronology (TCN) to date the offset alluvial fans along the Fish Lake Valley fault. TCN geochronology allows us to determine the age of abandonment of an alluvial surface (e.g., Gosse and Phillips, 2001). TCN geochronology measures the concentration of nuclides produced in a rock by the interaction between cosmic rays and minerals at Earth's surface (e.g., Lal, 1991; Gosse and Phillips, 2001). In order to obtain an accurate age for abandonment of each sampled geomorphic surface, several criteria must be satisfied: (1) the sampled boulder must be in the same geometry as it was at the time of deposition; (2) the sampled boulder should not have prior exposure history (inheritance); and (3) boulders with evidence of erosion should not be sampled since they will provide an attenuated concentration of cosmogenic isotopes and, hence, an anomalously young age (Gosse and Phillips, 2001).

The isotope of interest for this study is ¹⁰Be, which is produced through spallation and muon-induced reactions with Si and O in quartz. Beryllium-10 is well retained in quartz, so we collected quartz-bearing samples from granitic boulders embedded in the surface of the faulted alluvial fans. Fifteen samples were collected from the top 1–5 cm of large granitic boulders (Table 1). These boulders came from the stable parts of fan surfaces mapped by Reheis et al. (1993, 1995) as Qfm (alluvium of McAfee Creek) at Wildhorse Creek and Qfi (alluvium of Indian Creek) at Perry Aiken Creek (Fig. 3). We carefully selected well-varnished boulders that lacked evidence of erosion (for example, we did not sample “sombrero-shaped” boulders). If, however, any erosion has occurred, which we think is probably unlikely due to the arid climate in the study area, the boulders would yield anomalously young ages, which would in turn make our calculated slip rates too fast.

Quartz was purified by standard techniques, and Be was extracted using ion-exchange chromatography, precipitated as BeOH, and converted to BeO at the Georgia Institute of Technology cosmogenic nuclide geochronology laboratory (e.g., Kohl and Nishiizumi, 1992; Bierman and Caffee, 2002). The ¹⁰Be/⁹Be ratio for each sample was measured at the Center for Accelerator Mass Spectrometry at Lawrence Livermore National Laboratory, and model ¹⁰Be ages were calculated using the CRONUS-Earth ¹⁰Be-²⁶Al exposure age calculator (version 2.1; http://hess.ess.washington.edu) and constant ¹⁰Be production rates (Lal, 1991; Stone, 2000; Balco et al., 2008) (Table 1). In Table 1, we report the ages using a constant production rate, as well as time-varying production rates calculated using four different models (Lal, 1991; Stone, 2000; Dunai, 2001; Desilets and Zreda, 2003; Lifton et al., 2005; Desilets et al., 2006). In reporting ages and slip rates herein, we use the constant-production-rate model developed by Lal (1991) and Stone (2000) to ensure consistency in our comparisons with the previously determined ages of Frankel et al. (2007b), who used the same model for the Furnace Creek and Indian Creek sites. Time-varying production rates could result in younger ages by up to 10%, which would yield commensurately faster rates of extension.

Nine TCN samples were analyzed from the Qfi alluvial-fan surface at Perry Aiken Creek. The Qfi surface is characterized by subdued to moderately incised channels, well-developed desert pavement, and a continuous, thick desert varnish on clasts. Clasts are tightly packed and well sorted in a continuous pavement, and prominent solifluction treads and risers are also present (Reheis and Sawyer, 1997). In addition, the Qfi deposit is distinguished by a well-developed soil with a 5- to 10-cm-thick silty vesicular A horizon, a weak argillic B horizon with thin clay films, and stage II to III carbonate development (e.g., Reheis and Sawyer, 1997).

Our Qfi surface exposure ages exhibit a pronounced, single peak distribution with a mean age and standard deviation of 71 ± 8 ka (Fig. 4A). The ages range from 54 ± 5 ka to 79 ± 7 ka (Table 1). We take the tight clustering of ages as an indication that the Qfi surface has remained relatively stable and that there is negligible inheritance. The age we obtained for Qfi at Perry Aiken Creek is indistinguishable from the previously determined ¹⁰Be age of 71 ± 8 ka for the Qfi deposit at Indian Creek by Frankel et al. (2007b), and falls within the 50–130 ka age range estimated by Reheis and Sawyer (1997) on the basis of soil development and surface morphology. The presence of an extremely tight subcluster of six ¹⁰Be ages centered at ca. 75 ka suggests that these ages may most accurately reflect the age of abandonment of the Qfi surface at Perry Aiken Creek. These six ages yield an age for the surface of 75.4 ± 2.2 ka. In this interpretation, the three younger ages outside of the very tight 75 ka cluster would reflect either slight, unrecognized erosion of the sampled tops of these boulders and/or minor exhumation of these three clasts. Although we suspect that the slightly older 75.4 ± 2.2 ka age based on the tight subcluster of six samples may more accurately reflect actual age of the surface, in the following discussion, we use the 71 ± 8 ka age based on the entire suite of nine samples for consistency with our other measurements and those of Frankel et al. (2007b). We note, however, that if the tight ca. 75 ka cluster of six ages does represent the true age of abandonment for the Qfi surface, then the slip rate we calculate from the 71 ± 8 ka age range for all nine samples would yield values (both normal and strike-slip) that are ∼5% too fast.

We analyzed six TCN samples collected from the Qfm alluvial-fan deposits at Wildhorse Creek. The Qfm surface, referred to as “alluvium of McAfee Creek” by Reheis and Sawyer (1997), is characterized by subdued to moderately incised channels, well-developed desert pavement, and a continuous, thick desert varnish on clasts. The Qfm deposit exhibits a thick, silty vesicular A horizon, strong argillic B horizon, and stage IV laminar carbonate development (e.g., Reheis and Sawyer, 1997). The pavement is locally moderately well packed and well sorted, and abundant carbonate rubble is present throughout (Reheis and Sawyer, 1997).

The six ¹⁰Be samples we dated from this surface range in age from 100 ± 9 ka to 140 ± 13 ka (Table 1), forming a relatively tight cluster at 121 ± 14 ka (Fig. 4B). As at Perry Aiken Creek, we take the clustered distribution of these ages as evidence that the Qfm surface has remained relatively stable since the time of deposition and that inheritance is minimal. The age we determined for the Qfm surface is significantly younger than the 600?–760 ka age range proposed by Reheis and Sawyer (1997) on the basis of soil development and surface morphology.

The ages of the alluvial fans at Furnace Creek and Indian Creek were recently determined by Frankel et al. (2007b) using cosmogenic ¹⁰Be geochronology, and we utilized their results to determine the extension rates at these two locations. At Furnace Creek, Frankel et al. (2007b) reported an age for surface Qfio (modified from Qfi of Reheis and Sawyer, 1997) of 94 ± 11 ka, whereas at Indian Creek, they determined the age of surface Qfiy (modified from Qfi of Reheis and Sawyer, 1997) to be 71 ± 8 ka. Both of these ages are in agreement with soil and fan morphology data reported previously from those sites by Reheis and Sawyer (1997).

The analysis of LiDAR digital topographic data is an integral component of our study. The LiDAR data were collected in the fall of 2005 by the National Center for Airborne Laser Mapping (NCALM) using an Optech Inc. model ALTM 2033 laser mapping system. The laser was installed on a Cessna 337 twin-engine airplane, which flew over the fault trace at an average elevation of 600 m above ground level and an average speed of 60 m/s. The pulse rate frequency of the Optech ALTM 2033 was set at 33 kHz, and it recorded the first and last returns of each pulse, plus the relative intensity of each return. The average shot density for the LiDAR data was ∼3 points/m². The aircraft was equipped with a dual-frequency global positioning system (GPS) receiver and a real-time display of the flight path and area coverage. High-resolution digital elevation models (DEMs) with 5–10 cm vertical accuracy and 1 m horizontal resolution were produced using a kriging algorithm in SURFER software (version 8.04; Carter et al., 2007; Sartori, 2005).

ArcGIS (version 9.2) was used to produce hill-shaded relief and topographic maps to aid in the identification and mapping of all normal fault scarps at each site. We analyzed a total of 25 profiles perpendicular to the strike of each set of scarps at the four study sites. The vertical displacement components for each set of faults were measured by multiple topographic profiles, and the means of these measurements were calculated (Fig. 5; Fig. DR1 and Table DR1¹). The total vertical component of displacement for each site was determined by summation of the mean values. The horizontal component of each displacement was subsequently calculated using simple trigonometric relationships by assuming a 60° dip of the fault plane for each scarp (e.g., Kirby et al., 2006; Le et al., 2006, 2009a). Uncertainties associated with measurements of the vertical components of each displacement include surface roughness (∼20 cm) and the vertical accuracy (∼10 cm) of the LiDAR data. Combined, these two uncertainties are <5% of the total mean vertical component of displacement at each site, and thus we report a conservative displacement error of 5%. This 5% error is also carried over to the extensional component of displacement. We report the horizontal (extensional) component of displacement across the fault zone at each of the four sites, rather than the total vertical separation, due to the presence of multiple scarps associated with antithetic faults. Antithetic faults would lead to a reduction in the net vertical displacement across the fault zone. Therefore, we feel that the total extension represents a more robust measure of the net fault slip. At each site, we determined an extension direction that is normal to the average strike of the faults. Due to the presence of multiple sets of faults with different orientations at Furnace Creek, Wildhorse Creek, and Indian Creek, we resolved the cumulative extension at these three sites as the sum of the extension vectors for each fault set (Fig. 6).

One concern in this type of analysis is the oblique nature of slip on the Fish Lake Valley fault (Figs. 7–10,⁸⁹¹⁰). In the absence of direct field evidence for the rake of oblique slip (e.g., slickensides), we analyzed the difference between horizontal extension and right-lateral slip along the main right-lateral fault strand at each of our four study sites. Specifically, at each site, we used restorations of total strike-slip motion to define the pre-offset geometry of the dated alluvial-fan surfaces. Any scarps that exhibit a vertical component of slip on the restored surfaces must be caused by true dip-slip motion, and we used measurements of those scarp heights in our analysis of extension. We used the strike-slip restorations of Frankel et al. (2007b) for this analysis at the Furnace Creek and Indian Creek sites. As discussed later, we provide a new strike-slip restoration and strike-slip rate at the Perry Aiken site. In marked contrast to the other three study sites, we could not determine a unique strike-slip restoration at the Wildhorse Creek site, so we used the strike-slip rate of the fault determined by Frankel et al. (2007b) and the age of the offset alluvial surface to provide an approximate restoration. Given the overall similarity of the three well-constrained strike-slip rates that are now available along the fault (Furnace Creek and Indian Creek from Frankel et al. [2007b]; Perry Aiken Creek as documented herein), we think that this restoration is likely to be approximately correct.

Two fault sets are prominent at the Furnace Creek site: north-northwest–trending faults of the main, predominantly right-lateral strand of the Fish Lake Valley fault, and a northeast-trending set of distributed normal faults to the east. In order to account for the difference in strike between the two sets of faults at this site, we resolved the extension measured by the profiles from the two fault sets as a vector sum oriented toward N83°E (Fig. 7). Transects PP′-P1P1′ and QQ′-Q1Q1′, and R-R′ through W-W′ are superimposed in order to capture the vertical component of displacement across two scarps that juxtapose the right-laterally offset alluvial fan (Figs. 5A and 7). Inasmuch as right-lateral offset of the alluvial fan at this location will result in apparently smaller scarp heights due to juxtaposition of laterally variable topography, these profiles provide us with the minimum vertical displacement. In order to determine a more accurate measurement of potential dip-slip motion on the mountain-front strand, we used the strike-slip restoration of Frankel et al. (2007b). These authors showed that restoration of 290 ± 20 m of right-lateral slip realigns the cone-shaped geometry of the Qfi fan as well as a major incised drainage (their Fig. 2). Our measurement of residual scarp heights on the Frankel et al. (2007b) restoration indicates a vertical component of displacement of 8 ± 1 m. This small amount of oblique-normal displacement indicates that the main mountain-front fault strand at Furnace Creek is a near-pure right-lateral strike-slip fault (strike-slip/normal: 290/8), with a slip direction oriented N48°W, plunging at 3° toward the northwest.

In marked contrast to the predominantly right-lateral oblique-slip mountain-front strand, the scarps we mapped to the east do not exhibit any apparent strike slip, suggesting that they are nearly pure dip-slip faults. We used profiles X-X′ and Y-Y′ to capture the horizontal component of extensional displacement across these northeast-oriented scarps. By combining the extension we measured across these faults with that on the main, mountain-front fault strand, we get a cumulative mean horizontal extension across all faults at Furnace Creek of 13.0 ± 0.7 m toward N83°E (Fig. 6A).

Unlike the other three study sites, we could not determine a well-defined strike-slip offset along the mountain-front strand (Fig. 8) at Wildhorse Creek. We therefore used the well-constrained 3.1 ± 0.4 mm/yr of Frankel et al. (2007b) from the nearby Furnace Creek site (5 km south) and the 121 ± 14 ka age of the Wildhorse Creek Qfm surface to estimate total strike-slip offset of the Qfm surface of ∼350 m. This restoration matches a shutter ridge to the east with the Qfm deposits from the main alluvial fan formed out of Wildhorse Creek (Fig. 11). This admittedly crude strike-slip restoration indicates that there is some oblique displacement along the main fault strand, which is best observed on the restored surface north of the creek. We estimate the height of the scarp on the restored fan surface north of Wildhorse Creek to be less than 15 m. We note that this observation is relatively insensitive to the details of our approximate strike-slip restorations, as changing the restoration by 40 m does not result in a significant change in scarp height. This suggests that motion along the main mountain-front fault strand at Wildhorse Creek is predominantly strike slip.

The two profiles (N-N′ and O-O′) we analyzed north of Wildhorse Creek were oriented to most effectively capture all of the identifiable fault scarps (Figs. 5B and 8). Profile N-N′ measured four smaller scarps striking approximately N5°W, whereas profile O-O′ measured the scarp along the main mountain-front fault strand, which strikes N40°W. Although some of the scarps are antithetic, which would lead to a reduction in the net vertical displacement across the fault zone, it does not affect our calculations in terms of horizontal displacement. The cumulative mean horizontal component of extensional displacement at Wildhorse Creek is 52.3 ± 2.6 m toward N53°E (Fig. 5B).

The two profiles (NN-NN′ and OO-OO′) that we measured across the normal fault scarps south of Wildhorse Creek provide additional information about the distribution of slip on this section of the Fish Lake Valley fault. Specifically, although we could not effectively measure the amount of extension across the main, mountain-front strand south of Wildhorse Creek because of the presence of a prominent pressure ridge along this part of the fault system, we were able to measure the amount of extension accommodated to the west of the mountain-front strand along several north-northwest–trending normal faults that add slip to the mountain-front system from the southwest. Specifically, profiles NN-NN′ and OO-OO′ indicate that these faults have accommodated a total of 19.9 ± 1.0 m of extension oriented toward N77°E since abandonment of the Qfm surface (Fig. 6B). As discussed later, northward addition of this extra component of extension from distributed extension within the White Mountains partially accounts for the northward increase in extension rates that we observe along the Fish Lake Valley fault system.

The fault zone at Perry Aiken Creek exhibits a complex pattern of subparallel, anastomosing fault strands (Fig. 9). This section of the fault includes one of the tallest single fault scarps in Fish Lake Valley, located just north of the mouth of Perry Aiken Creek. This 73.5-m-tall fault scarp provides clear evidence for significant normal displacement. However, the fault strand also exhibits evidence for large right-lateral strike-slip displacement. Specifically, back-slipping the main, mountain-front fault strand by 250 m restores the right-laterally offset Qfi terrace riser along the margins of Perry Aiken Creek (Fig. 11). Combining the 250 m offset with the 71 ± 8 ka age for the nine ¹⁰Be samples from the Qfi surface yields a strike-slip rate of ∼3.2–4.0 mm/yr, similar to, but slightly faster than, the 3.1 ± 0.4 mm/yr rate calculated by Frankel et al. (2007b) from the Furnace Creek site to the south. Alternatively, if, as we suspect, the 75.4 ± 2.2 ka age for the very tight cluster of six ages from the Perry Aiken fan surface more accurately reflects the age of abandonment of the Qfi surface, then the strike-slip rate would be slightly slower at ∼3.2–3.4 mm/yr. We also note that on the restored image (Figs. 11C–11D; Fig. DR2 [see ^{footnote 1}]), there remains a pronounced right deflection of the Perry Aiken Creek canyon ∼300 m west of (i.e., upstream from) the main strike-slip fault strand. It is possible that this represents an additional strike-slip strand, which would increase the strike-slip rate at this site. However, the westernmost main fault strand that extends southward to near this right deflection does not appear to cross the canyon, suggesting that the deflection may be purely of fluvial origin, perhaps at a bend in the stream localized at one of the minor normal fault scarps that extend through this area.

The presence of a significant component of strike slip at this site raises the possibility that the very tall scarp just north of Perry Aiken Creek is due partially to strike-slip offset of laterally varying topography. However, measurements of four topographic profiles across all strands of the fault system, including across the tallest part of the scarp north of Perry Aiken Creek, yield strikingly similar amounts of total vertical separation on the Qfi surface of 85.1 ± 4.3 m (Figs. 5C and 9). The similarity of these measurements along different profiles suggests that this is a robust observation of the total vertical component of oblique slip at the Perry Aiken Creek site. Moreover, along profiles L-L′ and M-M′, the Qfi surface is buried beneath the gently north-sloping, northern shoulder of a younger fan east of the fault zone, indicating that the vertical measurements on those profiles are minima. The vertical component of total displacement is in agreement with previous measurements by Reheis and Sawyer (1997), who suggested 85.5 m of vertical separation on the Qfi surface. Based on our 85.1 ± 4.3 m measurement of total vertical separation, the cumulative mean horizontal extensional component of displacement at Perry Aiken Creek is 49.1 ± 2.5 m toward N68°E (Fig. 6C). A comparison of the extensional and strike-slip components of slip indicates that the Fish Lake Valley fault at Perry Aiken Creek is primarily a right-lateral strike-slip fault, with a strike-slip to dip-slip ratio of ∼5:1 (i.e., 250 m:49 m).

At the Indian Creek site, we analyzed nine topographic profiles across two different sets of fault scarps (Figs. 5D and 10). Profile A-A′ extends across the two normal fault scarps, one facing east and one facing west, to the west of the range-front dextral fault strand. We used profile B-B′ to capture the horizontal extensional displacement across the main right-lateral strand of the fault. This profile provides us with a minimum vertical displacement, as right-lateral offset of the alluvial fan at this location will result in apparently smaller scarp heights due to juxtaposition of laterally variable topography. As at Furnace Creek, we used the strike-slip restoration of Frankel et al. (2007b) to determine a more accurate measurement of potential dip-slip motion on the mountain-front strand. Frankel et al. (2007b) showed that restoration of 178 ± 20 m of right-lateral slip realigns numerous incised drainages, as well as the overall limits of the Qfi fan (their Fig. 3). Our measurement of residual scarp heights on the Frankel et al. (2007b) restoration indicates a vertical component of displacement of 23 ± 1 m. By combining the horizontal and vertical components of slip, we find a total oblique displacement of the Qfi surface of 179 ± 20 m, with a slip direction plunging 7° toward N30°W. Profiles C-C′ through I-I′ were used to measure the displacement across a set of distributed, north-northeast–trending normal faults to the east. In order to account for the difference in strike between the main strand and the eastern fault set, we resolved the extension measured by the profiles from the two fault sets as a vector sum. The cumulative mean horizontal extensional displacement at Indian Creek is 43.7 ± 2.1 m toward S88°E (Fig. 6D). The cumulative vertical displacement across all fault strands at this site (75.4 ± 3.8 m) is almost twice as large as the previously reported preferred vertical component of displacement of 40 m by Reheis and Sawyer (1997).

We calculated extension rates at each of our four study sites by combining vertical components of displacement, fan surface ages, and an assumed 60° fault dip. Specifically, our extension rates were computed by combining probability density functions of the measured displacements and TCN ages. These extension rates were first calculated in their original extension direction, and then resolved to an extension direction of N65°E to facilitate comparison with geodetic measurements of extension rates (Table 2). We used a Gaussian uncertainty model (e.g., Bird, 2007; Kozaci et al., 2009; McGill et al., 2009; Zechar and Frankel, 2009), and the uncertainties in the extension rates are reported at the 2σ confidence interval. The resulting extension rates from south to north are: Furnace Creek = 0.2 ± 0.1 mm/yr toward N83°E, Wildhorse Creek = 0.4 ± 0.2 mm/yr toward N53°E, Perry Aiken Creek = 0.7 +0.3/–0.1 mm/yr toward N68°E, and Indian Creek = 0.6 +0.2/–0.1 mm/yr toward S88°E.

The ¹⁰Be dates from all four of our sites should be considered maximum ages for calculating the extension rates because the normal fault scarps must have developed after the deposition and abandonment of the Qfi and Qfm alluvial-fan deposits. However, inasmuch as extension has been active at similar rates along the Fish Lake Valley fault system for at least 760 k.y. (Reheis and Sawyer, 1997), any time lag between fan abandonment and initial extensional deformation of the fan surface is likely to have been brief compared with the late Pleistocene ages of the measured fans. Thus, this source of potential error will have a negligible effect on our calculated extension rates. However, although we are confident that we have captured all of the main fault strands that exhibit a normal component of slip, some distributed deformation that does not manifest itself as recognizable fault scarps could be present. Furthermore, in our analysis, we do not include minor fault strands that are visible on the LiDAR data, but which exhibit scarp heights of <0.5 m, which we consider to be close to the resolution limit of this technique. In addition, assuming a steeper fault dip angle, for example, 75°, would lower each of the extension rates by ∼50%. The oblique, predominantly strike-slip motion along the Fish Lake Valley fault might at first suggest that the dip of the faults may indeed be somewhat steeper than 60°. As noted already, however, the fault system is largely strain partitioned into strike-slip and normal faults, lending confidence to our assumption of a 60° dip angle for the extensional faults. If steeper faults do exist, they would most likely be the mountain-front strands at Wildhorse Creek and Perry Aiken Creek, where they exhibit oblique motion. Thus, the extension rates we calculate for these sites may be maxima. We emphasize, however, that we do not have any direct measurements of the dip of these faults.

The new rate data described here allow us to place quantitative constraints on the style and location of northward strain transfer through the Mina Deflection from faults of the Eastern California shear zone to structures in the Walker Lane belt. The extension rates we obtained on the Fish Lake Valley fault increase northward from 0.2 ± 0.1 mm/yr at Furnace Creek, to 0.4 ± 0.2 and Wildhorse Creek, and to 0.7 +0.3/–0.1 mm/yr and 0.6 +0.2/–0.1 mm/yr at the Perry Aiken Creek and Indian Creek sites, respectively. These extension rates are similar to those estimated by Reheis and Sawyer (1997), who reported a preferred late Pleistocene vertical component of oblique slip at Furnace Creek of 0.3 mm/yr and 0.8 mm/yr at Indian Creek, on the basis of tephrochronology. If we assume a 60° dip for the fault plane, their preferred extension rates at Furnace Creek and Indian Creek would be 0.2 mm/yr and 0.5 mm/yr, respectively. No preferred vertical component of total slip rate was reported by Reheis and Sawyer (1997) for Wildhorse Creek and Perry Aiken Creek.

At our Wildhorse Creek site, we were able to quantify the amount of extension that has been added to the Fish Lake Valley fault system between that site and the Furnace Creek site to the south. Specifically, we used two profiles (NN-NN′ and OO-OO′) just south of Wildhorse Creek that capture normal fault scarps that provide additional extension to the fault system. When resolved to N65°E, the extension rate on the fault set west of the main strand is 0.2 ± 0.1 mm/yr. Profiles N-N′ and O-O′, located to the north of Wildhorse Creek, demonstrate that the cumulative extension rate oriented toward N65°E at the northern part of the Wildhorse Creek site is 0.3 ± 0.2 mm/yr. Thus, the addition of the extension accommodated by the western fault set increases the overall extension rate at Wildhorse Creek by 0.1 ± 0.1 mm/yr. This is consistent with our observation that the extension rate oriented toward the geodetically defined current extension direction of N65°E increases northward from 0.1 ± 0.1 mm/yr at Furnace Creek to 0.3 ± 0.2 mm/yr at Wildhorse Creek, and it suggests that significant extension may be accommodated by distributed normal faulting within the White Mountains to the southwest of the Wildhorse Creek site.

It is possible that the northward increase in extension rates that we measured along the Fish Lake Valley fault is a manifestation of a temporally variable slip rate, rather than along-strike variations in slip, inasmuch as we compared offset alluvial fans that span a wide range of ages between 71 ± 8 ka and 121 ± 14 ka. If this change is due to temporal variation in slip rate, then the extension rate must have accelerated since ca. 94 ka (the age of Furnace Creek alluvial fan) by a factor of 2–3 times to account for the more rapid extension observed in the younger Perry Aiken Creek and Indian Creek fans. Moreover, the overall consistency between our late Pleistocene extension rates (averaged over 71 k.y. to 121 k.y., depending on the site) and the longer-term rates of Reheis and Sawyer (1997; averaged over 760 k.y.) suggests that the extension rate may be relatively constant over a wide range of time scales. Thus, although we cannot rule out a temporal change in extension rates, we think that a more likely explanation lies in the overall geometry of the fault system and the manner in which slip is transferred northward from the Eastern California shear zone into the Walker Lane belt across the Mina Deflection.

In order to compare short-term geodetic and longer-term geologic extension rates at the latitude of Fish Lake Valley, we resolved the geologic extension-rate vectors from our four study sites toward the extension direction of N65°E defined by the geodetically constrained models of Bennett et al. (2003) and Wesnousky (2005a). The resulting N65°E extension rates at our four sites are 0.1 ± 0.1 mm/yr at Furnace Creek, 0.3 ± 0.2 mm/yr at Wildhorse Creek, 0.7 +0.3/–0.1 mm/yr at Perry Aiken Creek, and 0.5 +0.2/–0.1 at Indian Creek (Table 2).

As noted already, at the latitude of Fish Lake Valley, over geologic time scales, approximately half of the current 1.0 mm/yr (Wesnousky, 2005b) extension rate defined by short-term geodetic data is accommodated by faults to the west, including the White Mountains fault (Kirby et al., 2006), distributed normal faulting in the Volcanic Tablelands (Kirby et al., 2006; Greene et al., 2007; Sheehan, 2007; data of Greene and Kirby inFrankel et al., 2008a), and the Sierra Nevada frontal fault system (Le et al., 2006), including the Round Valley and Hilton Creek faults north of Owens Valley (Fig. 12; Berry, 1997). The oblique-normal-dextral White Mountains fault exhibits a late Pleistocene extension rate at the latitude of our Furnace Creek site of ∼0.2 mm/yr (Kirby et al., 2006), whereas at approximately the same latitude, there is clear evidence for distributed normal faulting across the Volcanic Tablelands (e.g., Sheehan, 2007; Pinter and Keller, 1995). Sheehan (2007) reported an extension rate across the Volcanic Tablelands of ∼0.3 mm/yr. Further west at the same latitude, Berry (1997) reported a 0.5–0.6 mm/yr late Pleistocene vertical component of slip on the Round Valley fault. This is equivalent to an extension rate on a 60° dipping fault of ∼0.3 mm/yr.

Combining all of these extension rates with our rates from the Fish Lake Valley fault, the N65°E component of extension for all of the major fault systems in the northernmost Eastern California shear zone, from the Sierra Nevada fault to the west and the Fish Lake Valley fault to the east, is approximately equal to the geodetically determined extension rate at the latitude of central/northern Fish Lake Valley. This comparison suggests that the rate of extension at the latitude of northern Fish Lake Valley may have remained relatively constant over the past 10⁴–10⁵ yr, although it is also possible that temporal variations in slip rate have occurred over shorter time scales on the various faults that accommodate extension across the region. Additional slip rate calculations on all of these faults at a wider span of time scales are necessary to test whether such temporal variations in rate have occurred.

The Fish Lake Valley fault terminates just north of Indian Creek, and slip from this fault, as well as the White Mountains–Queen Valley and Tablelands fault systems to the west, is transferred northeastward across the Mina Deflection onto oblique-normal right-lateral faults of the Walker Lane belt. Thus, the Mina Deflection can be thought of as a major (∼80-km-wide), east-trending right step in a dominantly right-lateral, north-northeast–trending fault system (Stewart, 1988; Dixon et al., 1995; Reheis and Dixon, 1996; Oldow et al., 1989, 1994, 2001, 2009; Lee et al., 2001, 2006, 2009a, 2009b; Petronis et al., 2002, 2009; Oldow, 2003; Stockli et al., 2003; Wesnousky, 2005a, 2005b; Kirby et al., 2006; Frankel et al., 2007b; Sheehan, 2007). Within the Mina Deflection, deformation is accommodated by east-trending left-lateral faults and clockwise block rotations (Stewart, 1985; Cashman and Fontaine, 2000; Faulds et al., 2005; Wesnousky, 2005a).

As documented by Frankel et al. (2007b) and the Perry Aiken strike-slip rate we present herein, the strike-slip rate decreases northward along the Fish Lake Valley fault, from 3 to 3.5 mm/yr at Furnace Creek and Perry Aiken Creek, to 2.5 mm/yr at our northernmost study site at Indian Creek. The northward decrease in right-lateral slip rate is even more pronounced in the northernmost part of Fish Lake Valley, where the surface expression of the fault zone ends abruptly ∼10 km north of Indian Creek. The observation that extension rates increase northward along the Fish Lake Valley fault, whereas dextral rates decrease, has important implications for the distribution of strain along this section of the Pacific–North America plate boundary, and more generally for mechanisms of slip transfer along evolving, structurally complex fault systems (Fig. 12).

In general, both the northward increase in extension rate that we document and the northward decrease in dextral slip rate documented by Frankel et al. (2007b) reflect transfer of slip off the predominantly right-lateral Fish Lake Valley fault and onto north- and northeast-trending normal faults as part of a distributed zone of slip transfer located in the ∼40-km-long by 30-km-wide, triangular area east of the Fish Lake Valley fault between the Emigrant Peak fault and the east-trending left-lateral faults of the Mina Deflection (Fig. 2). For example, the north-northeast–trending normal faults that cut the fan to the east of the main range-front fault strands at the Furnace Creek site appear to “pull” slip off the Fish Lake Valley fault system and transfer it northeastward onto the Emigrant Peak fault system (Fig. 6). Similarly, the north-northeast–trending normal faults at Indian Creek serve to transfer slip off the Fish Lake Valley fault and into the zone of distributed normal faulting in this corner of Fish Lake Valley, leaving only 2.5 mm/yr of right-lateral strike-slip motion on the Fish Lake Valley fault at this site (Fig. 9; Frankel et al., 2007b). This diffuse normal faulting along with normal displacements on the prominent Emigrant Peak fault system account for the development of the deep basin that defines the northeast-trending part of northern Fish Lake Valley (including the dry “Fish Lake” proper; Fig. 12). The most pronounced decrease in right-lateral strike-slip rate along the Fish Lake Valley fault occurs just north of the Indian Creek site, where the geomorphic expression of the fault system dies out completely within a zone of extensive recent lava flows in the Volcanic Hills (Fig. 12). Between Indian Creek and the northwestern limit of faulting observable on the LiDAR data within the southern part of the Volcanic Hills, the Fish Lake Valley fault system splays northward into an increasingly distributed set of numerous small-scale normal faults. Thus, between the north end of the geomorphically well-defined Fish Lake Valley fault at Indian Creek, and the left-lateral faults of the Mina Deflection to the north, it appears that distributed normal faulting may accommodate as much as 2.5 mm/yr of dextral motion. The coincidence of this zone of apparent distributed normal faulting and the extensive volcanism in the Volcanic Hills suggests that the volcanism may be localized by this slip transfer zone. The young volcanism in this area may also serve to obscure geomorphic evidence for recent normal faulting.

Ultimately, much of this motion must be accommodated along the east-west–trending left-lateral faults of the Mina Deflection (Fig. 1; e.g., the Coaldale, Excelsior Mountains, and Rattlesnake Flat faults) (Stewart, 1985; Wesnousky, 2005a; Lee et al., 2009a). However, the manner in which this slip transfers northward onto the left-lateral faults remains unclear because there are no geomorphically well-expressed faults in the 10-km-wide zone between the northern end of the Fish Lake Valley fault and the Coaldale fault (Fig. 12). If, as we suspect, this slip is transferred northward into the Mina Deflection along a diffuse set of highly distributed normal faults beneath the Volcanic Hills, then this would imply that clockwise rotations and/or left-lateral slip rates on the Mina Deflection faults would increase eastward.

One possibility that we consider is that the northeast-trending normal faults that characterize the northern part of Fish Lake Valley represent an early stage in the evolution of faults similar to the east-trending left-lateral faults of the Mina Deflection. In such a scenario, these faults would develop as northeast-trending normal faults that act to transfer strain across the major right step of the Mina Deflection. As noted in Wesnousky's (2005a) earlier model for the structural evolution of the Mina Deflection, in response to ongoing right-lateral shear, these northeast-trending normal faults would gradually rotate clockwise into a more east-west orientation, switching to left-lateral strike-slip motion as a result of this reorientation. However, both the well-established nature of the northeast-trending basin along the north side of the Silver Peak Range and long-term activity of the north- to northeast-trending Emigrant Peak fault system (Petronis et al., 2002, 2009; Petronis, 2005) argue that these are well-established, long-lived features that do not appear to be actively rotating. In either case, slip transfer across the northern end of Fish Lake Valley into and across the Mina Deflection appears to involve a large component of distributed normal faulting, as well as left-lateral strike-slip faulting, perhaps quite distributed at the northwest corner of the valley, and clockwise rotations (Fig. 13).

New LiDAR topographic data and cosmogenic ¹⁰Be geochronology of offset alluvial-fan deposits on the dextral-oblique Fish Lake Valley fault yield well-determined late Pleistocene–Holocene extension rates on this major oblique-normal-dextral fault system. The surface exposure ages of four sites, Furnace Creek, Wildhorse Creek, Perry Aiken Creek, and Indian Creek (from south to north), range from 71 ± 8 ka to 121 ± 14 ka, and the mean horizontal extensional components of displacement at these sites range from 12.4 ± 0.6 m to 49.0 ± 2.5 m toward N65°E. By combining probability density functions of these displacements and ages, we find that extension rates averaged over late Pleistocene–Holocene time vary from 0.1 ± 0.1 mm/yr at Furnace Creek and 0.3 ± 0.2 mm/yr at Wildhorse Creek in the south, to 0.7 +0.3/–0.1 and 0.5 +0.2/–0.1 mm/yr at Perry Aiken Creek and Indian Creek, respectively, to the north.

These rates suggest that the Fish Lake Valley fault accommodates approximately half of the regionwide current rate of extension measured geodetically. When summed with extension rates on faults along the western White Mountains piedmont, the Sierra Nevada frontal fault, and distributed deformation across the Volcanic Tablelands, the long-term geologic rates of extension are commensurate with the short-term rates determined from GPS data.

The increase in the east-northeast–west-southwest extensional component of slip toward the northern end of the Eastern California shear zone reflects a gradual northeastward transfer of slip off the predominantly right-lateral Fish Lake Valley fault and across the Mina Deflection as part of a distributed zone of northeast-trending normal faulting. Further north, in the Mina Deflection proper, deformation is accommodated predominantly by east-west left-lateral faults (e.g., Wesnousky, 2005a). Collectively, the distributed normal faulting in northern Fish Lake Valley, together with clockwise rotations and motion on the east-west left-lateral faults of the Mina Deflection, serves to transfer deformation through this major right step in the Eastern California shear zone–Walker Lane belt.

We thank Dylan Rood and Alicia Nobles for assistance with sample preparation and analysis, and Trevor Thomas for his assistance in the field. Jeff Lee, Lewis Owen, and Michael Oskin provided thoughtful reviews that significantly improved the manuscript. The LiDAR data were collected by the National Center for Airborne Laser Mapping, and we are indebted to Michael Sartori and Ionut Iordache for help with data processing. This research was made possible by the support of National Science Foundation (NSF) grants EAR-0537901 and EAR-0538009, a National Aeronautics and Space Administration (NASA) Earth System Science Fellowship, and the Georgia Tech Research Foundation.