The role of low-angle normal faulting in active tectonics of the northern Owens Valley, California

The Great Basin constitutes a continental-scale topographic and hydrologic feature created by extensional tectonics. Extension began during the Miocene in the central Great Basin, but at present, most of the tectonic activity is concentrated on the western and eastern margins of the region (Wernicke, 1992). The western margin is particularly active. The mechanics of extension in this region are of interest for several reasons. One reason is to help understand the continuity, or lack thereof, between tectonic mechanisms in the highly extended terrain of the Las Vegas–Death Valley area (in considerable part accomplished during the Miocene) and the more limited Pliocene–Quaternary extension directly east of the Sierra Nevada block. Due to a long interval of extension and erosion, deep-seated low-angle structures are well exposed in, for example, the Las Vegas area, Death Valley, and Panamint Valley (Wernicke et al., 1988). In contrast, along the highly active southwestern boundary of the Great Basin, the structures facilitating extension are generally buried beneath subsiding and aggrading basins. Are the surficial neotectonic features there rooted in low-angle structures similar to the Miocene ones to the east, or is the geometry of faulting instead “steep and fairly penetrating,” as postulated by Wernicke et al. (1988, p. 1751)?

A second reason to study the mechanics of extension in this area is to investigate a discrepancy between contemporary geodetic estimates of displacement rate and rates inferred over geological time scales based on fault investigations. The geodetically based rates are quite high. Geodetic measurements indicate that at the latitude of the northern Owens Valley, approximately two-thirds of the total extension between the North American craton and the Pacific plate is concentrated in a zone of only 70 km between the Sierra Nevada crest and Fish Lake Valley (Dixon et al., 2000). However, in a comparison of geodetic and geological displacement rates using a kinematic block model, the modeled rates based on geodetic data are two to five times larger than the ones based on field studies of faults (McCaffrey, 2005). The fault-based rates necessarily involve assumptions regarding fault geometry, and thus elucidation of fault structures may shed light on this discrepancy.

Finally, seismic hazard analyses also require assumptions regarding fault geometry. A better understanding of fault geometry may help to improve hazard estimates, and such insights may be transferable to other portions of the Basin and Range Province.

The Owens Valley defines the western margin of the Great Basin and the eastern margin of the Sierra Nevada block (Fig. 1). The valley consists of a trough-like structure, 140 km in length and 25–10 km in width, striking 30° to 20° west of north. Along most of its length, the western side of the valley is formed by the dramatic eastern escarpment of the Sierra Nevada, with total relief varying from 1700 to 2700 m. The eastern side is formed by the White and Inyo Mountains, which are more gentle than the Sierra Nevada and which have crests generally rising 1500–1000 m above the valley floor, but which extend up to 2700 m above the valley floor in the northern portion of the White Mountains. This generally north-south–oriented graben-like structure is typical of valleys in the Great Basin to the east. The mechanics of tectonics in the northern Owens Valley since 760 ka can be traced by evaluation of the deformation of the Bishop tuff, created by the eruption of Long Valley caldera.

The Owens Valley is clearly a tectonically produced feature, and its structural characteristics have largely been a matter of consensus for the past 140 yr. It is bounded on the west side by the Sierra Nevada frontal fault and on the east by the White Mountains fault zone and the Inyo Mountains fault zone (Fig. 1). These are normal faults, with the exception of the northern portion of the White Mountains fault zone, which exhibits oblique slip. The Owens Valley fault zone runs down the center of most of the valley and is predominantly a right-lateral strike-slip fault. The earliest investigators (Gilbert, 1883, 1884; Le Conte, 1901; Lee, 1906; Knopf, 1918) considered the Owens Valley to be a graben, by which they meant a crustal block that has been lowered (in a relative sense) along high-angle normal faults that bound mountain ranges on either side (i.e., horsts). Although the realization that the Owens Valley fault zone is a strike-slip fault has somewhat complicated this model (Beanland and Clark, 1994), high-angle normal block faulting has continued to be accepted as the mechanism for the formation of the Owens Valley as a topographic feature (Matthes, 1937; Pakiser et al., 1964; Bateman, 1965; Bateman and Wahrhaftig, 1966; Hollett et al., 1991; Berry, 1997; Le et al., 2007).

The faulting of the Owens Valley is one component of the regional-scale extension of the western Great Basin. This extension was initiated during the middle to late Miocene and has continued episodically to the present (Jones et al., 2004; Phillips, 2008; Jayko, 2009). The locus of extension has progressively migrated from east to west. As a result, the extensional basins exhibit decreasing maturity westward, with the floor of Death Valley now below sea level, that of Panamint Valley at ∼500 m, and Owens Valley at ∼1300 m. Death Valley and Panamint Valley are generally accepted to have extended along a combination of strike-slip and low-angle detachment faults (Burchfiel et al., 1987; Hamilton, 1988; Burchfiel et al., 1995). Due to the relatively advanced stage of extension, low-angle normal faults are clearly exposed in outcrop near the bottoms of these basins.

During the Paleozoic, the area now comprising the Owens Valley was in a marginal marine setting that produced thick sequences of sedimentary rock. Starting in the late Paleozoic, the continental boundary to the west changed to a subduction zone. Throughout the Mesozoic, arc volcanism dominated, and coalescence of plutons under that arc created the Sierra Nevada Batholith. Today, the Inyo and White Mountains that bound the Owens Valley on the east are dominated by Paleozoic sedimentary rock, while the Sierra Nevada range to the west has mostly been eroded down to the granitic rocks of the batholith (Bateman and Wahrhaftig, 1966; Saleeby, 1999).

Subsequent to the Late Cretaceous, the Sierra Nevada entered a passive phase (Bateman and Wahrhaftig, 1966; Wakabayashi and Sawyer, 2001). Thermochronologic data over the Cenozoic (Clark et al., 2005) and modern cosmogenic nuclide data (Riebe et al., 2001) both indicate that the uplands of the southern Sierra have eroded downward at an average rate of ∼40 m/m.y. This material was removed from the highlands of the range and deposited in the San Joaquin Valley, on the toe of the Sierra Nevada block. This unloading of the eastern part of the block and loading onto the western portion produced an isostatic response of very gradual westward tilting of the block (Small and Anderson, 1995). The net result was presumably a slow decrease in the average elevation of the range.

The middle Neogene paleotopography of the area now forming the Owens Valley has not been well established. Extensive basalt flows that have been dated to ca. 12 Ma rest on bedrock surfaces along the crest of the White Mountains and the Inyo Mountains (Krauskopf, 1971; Jayko, 2004, 2009), close to the crest of the Sierra Nevada west of Bishop, and in the foothills of the Owens Valley (Phillips et al., 2011). At ca. 12 Ma, the northern White Mountains underwent a major cooling episode (Stockli et al., 2003), and similar cooling ages are also found near Mount Whitney (Maheo et al., 2004). Phillips (2008) evaluated this information in the context of additional paleoenvironmental data and concluded that, prior to 12 Ma, the area between the crests of the Sierra Nevada and White/Inyo Mountains was probably a broad plateau-like summit region of subdued topography between 2000 and 3000 m elevation. An initial westward step of Basin and Range extension between 12 and 11 Ma (Henry and Perkins, 2001; Surpless et al., 2002) probably produced a relatively narrow and shallow proto–Owens Valley. Long, relatively flat ridge crests extending eastward from the southern Sierra Nevada crest, such as Lone Pine Peak, may be remnants of the floor of this valley (Matthes, 1937), suggesting a valley depth of 750–500 m (Phillips et al., 2011). Matthes (1937) speculated that the Alabama Hills were a subsequently downfaulted continuation of this valley floor, indicating that Miocene faulting may have been toward the eastern side of the present Owens Valley. At that time, extension was directed approximately east-west (Wernicke and Snow, 1998; McQuarrie and Wernicke, 2005), and thus normal faulting was initiated along an approximately north-south strike.

Middle to late Neogene cooling ages in the Owens Valley region exhibit a distinctly bimodal distribution, with one mode close to 12 Ma and the other close to 3 Ma (Surpless et al., 2002; Stockli et al., 2003; Maheo et al., 2004). This is paralleled by a similar bimodal distribution of basaltic volcanism, with groupings between 12 and 9 Ma and 4 and 3 Ma (Moore and Dodge, 1980; Manley et al., 2000; Phillips et al., 2011). Phillips (2008) and Jayko (2009) have reviewed evidence indicating relative tectonic quiescence between these episodes, and this is supported by similar evidence from portions of the Sierra Nevada to the north of the Owens Valley (Gilbert et al., 1968; Henry and Perkins, 2001; Surpless et al., 2002).

The 4 to 3 Ma eruptive pulse has been linked to delamination of the dense eclogitic root of the Sierra Nevada and its subsequent convection downward into the mantle (Wernicke et al., 1996; Manley et al., 2000; Farmer et al., 2002; Zandt et al., 2004). The delamination event has, in turn, been implicated as the initiator of a major increase in the rate of extension at that time in the area immediately east of the Sierra Nevada (Farmer et al., 2002; Jones et al., 2004). A wide range of evidence indicates that the rate of tilting of the Sierra Nevada increased markedly at about the same time (Le Conte, 1886; Lindgren, 1911; Huber, 1981; Unruh, 1991; Stock et al., 2005).

There is good reason to think that the origin of the modern Owens Valley dates to this late Pliocene interval. Wernicke and Snow (1998), Jones et al. (2004), and Phillips (2008) have reviewed evidence that points to initiation at that time of the current regime of relatively rapid extension. This regional evidence can be supported by local data. The geomorphology of the Sierra Nevada front along Owens Valley is consistent with formation of the range front in a single tectonic pulse, which is ongoing. Aside from the high ridge crests mentioned already as a possible remnant of Miocene tectonism, the front of the range along the northern Owens Valley generally presents a relatively smooth and uniform profile. Phillips et al. (2011) obtained a 3.4 Ma ⁴⁰Ar-³⁹Ar date from basaltic tephra contained in glacial fill of a doubly beheaded paleovalley near the summit of Mt. Humphreys (west of Bishop; Fig. 1). The glacial valley was thus present during the 4 to 3 Ma volcanic pulse. The paleovalley is evidently a relic of the gentle Miocene–early Pliocene upland landscape. The paleovalley is currently beheaded to the east by erosion encroaching westward due to down drop of the Owens Valley along the Sierra Nevada frontal fault. The preservation of the Pliocene tephra and till in the paleovalley implies that, at the time of the eruption, a low-relief upland landscape containing the valley extended to the west and had not yet been disrupted by normal displacement along the Sierra Nevada frontal fault. Initiation of significant movement on the Sierra Nevada frontal fault in the northern Owens Valley thus presumably postdates 3.4 Ma.

The sedimentation history of the Owens Valley also supports initiation of major downfaulting at this time. The Bishop tuff has been mapped in the subsurface within the fill of the northern Owens Valley (Bateman, 1965; Hollett et al., 1991). For example, beneath the town of Bishop, the tuff is encountered at a depth of 200 m. The total thickness of fill at this location is ∼950 m, based on gravity modeling (Saltus and Jachens, 1995). The age of the Bishop tuff is very well established at 760 ka (Sarna-Wojcicki et al., 2000). These values imply a sedimentation rate of ∼260 m/m.y. If the sedimentation rate has remained approximately constant, the total thickness of ∼950 m of fill would take ∼3.6 m.y. to accumulate. Similar calculations at other locations in the northern Owens Valley yield estimated basin-fill ages ranging from 3.5 to 6 Ma. The assumption of constant deposition rate is clearly only an approximation. It seems likely that, if deposition rates were not constant, they were probably larger in the early stages of basin filling, rendering the estimated fill ages maxima. Although the analysis is only semiquantitative, the results appear consistent with basin initiation at ca. 3.5 Ma. Pinter and Keller (1995) obtained a similar result from an independent approach involving basin tilting.

One important difference between the Miocene and Pliocene episodes of extension is that at ca. 8 Ma, the direction of extension shifted from approximately east-west to northwest-southeast (Wernicke and Snow, 1998; McQuarrie and Wernicke, 2005). However, the normal component of extension was accommodated mostly along north-south faults, resulting in a transtensive strain regime. This was likely due to reactivation of preexisting north-south faults formed during the Miocene, but it could also have resulted from the formation of new north-south faults, the orientations of which were controlled either by the edge of the Sierra Nevada Batholith or by the Owens Valley fault zone, a feature that dates back to at least the Cretaceous (Bartley et al., 2007).

In summary, prior to the middle Miocene, the area now occupied by the Owens Valley was probably a broad, low-relief summit upland stretching between the current crests of the Sierra Nevada and White/Inyo Mountains, and perhaps further east. An initial pulse of east-west extension in the interval 12–9 Ma produced approximately north-south normal faults, probably the ancestor of the current White Mountains fault zone. The throw across these faults was probably in most places limited to less than 1 or 2 km. After an interval of relative quiescence, extension was renewed starting at ca. 3.5 Ma, but with a northwest-southeast orientation. This Pliocene episode of extension has been linked to delamination of the eclogitic root of the Sierra Nevada. This extension has continued to the present day, resulting in the formation of the modern Owens Valley. The change in orientation required preexisting Miocene faults to accommodate transtensive stresses.

Numerous geodetic campaigns utilizing global positioning system (GPS) technology have quantified the modern displacement rates in the Owens Valley region. Based on evaluation of a large GPS data set, Dixon et al. (2000) estimated that the Sierra Nevada block is moving 13 ± 1.2 mm yr⁻¹ toward 312°, relative to stable North America. This can be compared to a high-quality displacement measurement of 10.6 ± 0.1 mm yr⁻¹ toward 310° at the Owens Valley Radio-Astronomy Observatory (OVRO) using very long baseline array interferometry (data compiled by Dixon et al., 2000). The OVRO site, however, is in the floor of the Owens Valley, and its displacement should differ from that of the Sierra block by the amount of displacement across the Sierra Nevada frontal fault. Dixon et al. (2000), indeed, found that a large proportion of the strain across the Great Basin at this latitude was absorbed in the region between the Sierra Nevada block and Death Valley.

In this paper, we analyze mechanisms of deformation across the northern Owens Valley, which reflects displacement between the Sierra Nevada and White/Inyo Mountains blocks. One approach to estimating this displacement rate is to compute the differences between stations in the two mountain blocks from the compilation of GPS-measured displacement rate vectors in Oldow (2003). Six GPS station velocities in the southern Sierra Nevada block were averaged and subtracted from the average of five station velocities in the White/Inyo Mountains block. The difference was 4.1 ± 0.8 mm yr⁻¹ directed toward 305° ± 6°. This direction is the same, within uncertainties, as that of the Sierra Nevada block and the entire southwestern portion of the Great Basin, and the azimuth of 312° from Dixon et al. (2000) is probably preferable. If this displacement rate (4.1 ± 0.8 mm yr⁻¹ toward 312°) is decomposed into components parallel to the Owens Valley fault zone and perpendicular to it, the parallel component is 3.7 ± 0.8 mm yr⁻¹ toward 340° and the perpendicular component is 1.5 ± 0.3 mm yr⁻¹ toward 250°. For comparison, the displacement rate parallel to the Owens Valley fault zone was estimated, using an elastic half-space model incorporating terrestrial geodetic data, to be 7 ± 1 mm yr⁻¹ (Savage and Lisowski, 1995). Also using an elastic half-space model, in this case incorporating GPS data, Dixon et al. (2000) obtained a fault-parallel displacement rate of 6 ± 2 mm yr⁻¹. Using a viscoelastic coupling model specifically parameterized to account for the 1872 Owens Valley earthquake, Dixon et al. (2003) obtained a displacement rate of 2.1 ± 0.7 mm yr⁻¹. Our empirical estimate lies between those from the elastic half-space and viscoelastic coupling models.

Numerous studies in the Owens Valley area have employed paleoseismic and longer-term geological methods to estimate fault displacement rates over periods ranging up to several million years. For the problem of determining a displacement rate between the White/Inyo Mountains block and the Sierra Nevada, the task is considerably complicated by the partitioning of strain among the Sierra Nevada frontal fault, the Owens Valley fault zone, and the White/Inyo Mountains fault zone (Le et al., 2007). In addition to these major fault zones, there exist numerous minor faults, and additional unmapped faults may exist beneath the young alluvium of the Owens Valley floor and beneath the bed of Owens Lake, which expanded far up the Owens Valley in the late Quaternary and may have obliterated evidence for older fault displacements (Bacon and Pezzopane, 2007). These prior studies are too numerous to evaluate here, but Bacon and Pezzopane (2007) and Le et al. (2007) have recently provided critical reviews. Le et al. (2007) employed displacement estimates based on their own data and a large number of prior studies to construct displacement-vector diagrams across the southern Owens Valley. These resulted in estimates of displacement azimuths ranging from 306° to 331° and displacement rates ranging from 2.0 to 3.2 mm yr⁻¹. The range of azimuth overlaps the GPS-based estimate of 312°. The range of rates also overlaps the GPS-based estimate of 4.1 ± 0.8 mm yr⁻¹, but the GPS value is clearly at the upper end of their range. However, their displacement-rate estimate does not include movement on the southern Inyo Mountains fault (Bacon et al., 2005) nor displacements on likely unmapped or concealed faults in the valley floor (Bacon and Pezzopane, 2007). Given a modest amount of additional displacement on these faults, the geological and GPS-based estimates of the total displacement vector between the White/Inyo and Sierra Ranges would appear to be in reasonable agreement.

Two factors, in particular, motivate a reexamination of the mechanics of faulting in the northern Owens Valley. The first is structural. The Owens Valley is a relatively shallow graben. Only small portions of the basin exceed 3 km of basin fill. A large proportion, especially of the southern Owens Valley, contains <1.5 km of fill (Saltus and Jachens, 1995). Given an average topographic difference of 2.7 km between the crest of the Sierra Nevada and the valley floor, this indicates typical total vertical displacement of ∼4 km. This is surprisingly shallow if the Owens Valley has been the locus of strong east-west extension (∼1.5 mm yr⁻¹) over the past 3.5 m.y. If the bounding faults dip at 60°, 8–10 km of vertical displacement would be produced. This discrepancy between observed and calculated displacement can, of course, be reconciled by calling on temporal variations in strain rate or orientation, but it nevertheless raises the question of whether other structural configurations might produce better agreement between observed and inferred vertical displacement.

A second motivation is provided by comparison with the observed fault structure in the extensional basins to the east. Cumulative extension much larger than in the Owens Valley, combined with lower rates of basin sedimentation, has revealed that the opening of Death Valley and Panamint Valley was largely accomplished along west-dipping low-angle faults (Burchfiel et al., 1987; Wernicke et al., 1988; Hayman et al., 2003). Similar mechanisms can be less directly inferred for Eureka Valley and Deep Spring Valley (Peltzer and Rosen, 1995; Lee et al., 2001). If Owens Valley is instead opening along high-angle faults, the difference should be explicable within a regional tectonic framework. One possibility is that Owens Valley and the basins to the east all initially opened on high-angle faults, but that the older eastern ones have subsequently rotated to lower dips (Proffett, 1977; Wernicke and Axen, 1988). A second is that the Owens Valley and the basins to the east actually do have fundamentally different styles of tectonics, for reasons that have not yet been explained. A third hypothesis is that the traditional interpretation that the Owens Valley is bounded by high-angle normal faults is not valid and that both it and the basins to the east extended along moderate- to low-angle normal faults.

The principal objective of this study is therefore to evaluate structural interpretations of faulting in the Owens Valley area in the context of tectonic forcing over the past ∼4 m.y., in order to determine the fault geometries that are most consistent with both the inferred tectonic history and the neotectonic evidence from the area. We have employed direct observations on fault outcrops, tectonic geomorphology, deformation of the Bishop tuff surface, and three-dimensional visualization of earthquake hypocenters to evaluate alternative interpretations.

As described in the previous sections, most geological researchers in the northern Owens Valley area have accepted high-angle normal faulting as the mechanism for relative displacement of the Owens Valley floor and the bounding mountain ranges. However, actual data on fault geometry are sparse. We have therefore attempted to find localities where direct measurement of fault-plane orientation on bedrock outcrops can be performed. Such sites, however, are scarce, so we have supplemented them with orientations determined from three-point solutions on the horizontal and vertical coordinates of points picked from fault scarps that disrupt topographic features. As fault orientations become steep, the fault-dip estimates become less robust, unless the fault runs across large vertical topographic features, and so for these we report only limiting minimum dip estimates and fault strike rather than direction of dip.

The results are given in Table 1 and illustrated in Figure 1. The data show a wide range of dip estimates, even along individual faults. Along the Round Valley fault, for example, dips are lowest in the center of the fault, at ∼31°. They steepen toward the south and especially the north, where, at the north end of Wheeler Crest, the fault becomes near-vertical. The Round Valley fault has previously been considered to be characterized by only normal displacement (Bryant, 1984, 2005; dePolo et al., 1993), but we found evidence of at least 150 m of dextral displacement of shutter ridges and streams (Fig. 2) between Swall Meadow and the north end of Red Mountain west of Rock Creek. The increase in fault dip correlates with increase in elevation of the fault outcrop while moving from the center of the fault on the face of Mount Tom toward the northern and southern ends of the fault. This suggests that the dip of the exposed portion of the fault is related to the amount of displacement across that portion of the fault (i.e., decreasing from the center of the fault toward its tips). This, in turn, suggests that the fault may have a listric geometry.

Fault-exposure dips on the White Mountains fault zone are more complex. At the southern end of the study area (“southern section”), where fault displacement is apparently only normal (Kirby et al., 2006), the dip is steep (>60°). Toward the north, the pattern of faulting becomes more complex, with numerous fault strands and overall dextral-oblique slip. Slip partitioning between fault strands is not readily apparent. In this section, all measured dips were quite steep (>60°). Starting at about Gunter Creek (Fig. 1), a pattern of slip partitioning begins to develop, with strike-slip displacement apparently focused on linear faults that run parallel to the mountain front on the upper portions of the alluvial fans and dip-slip concentrated on discontinuous fault segments at the mountain front to the east. This pattern grows more pronounced toward the north. In this portion of the fault, the strike-slip strands appear to be essentially vertical in orientation, while limited measurements on the mountain-front normal segments appear to indicate westward dips of ∼45°. We refer to this section as the “central section.”

Between Birch and Falls Canyons, the White Mountain fault zone transitions to a third regime. The strike-slip strand to the west dies out, while the zone of mountain-front dip-slip faulting dissolves into several separate strands that step eastward up the face of the range while maintaining a dip of ∼45° (Stockli et al., 2003). Farther northward (“northern section”), as Rock Creek is approached, this zone of normal faults coalesces into a single, normal, mountain-front fault. The measured dip angles decrease from 45° at Rock Creek to ∼35° north of Queen Dicks Canyon. The dips we obtained south of Rock Creek are similar to those measured by Stockli et al. (2003), but those north of this point are significantly less. One possible explanation is that Stockli et al. (2003) measured kinematic indicators on small-displacement faults subsidiary to the main range-bounding fault, whereas our hillslope-scale measurements were on the main fault. Displacement may be partitioned, with oblique slip focused on steeper faults behind or in front of the range front and dip slip on a shallower range-front fault.

The most important finding from the fault outcrop study is that significant portions of the valley-bounding faults show low-angle (<30°) dips or dips close to low angle. Low-angle faulting clearly plays at least some role in the tectonics of the Owens Valley.

We have in Table 2 compiled data on drainage areas, fan areas, and fan/drainage area ratios for the entire west face of the White Mountains. The geographical distribution of the fan/drainage ratios is shown in Figure 1. Across the study area, the fan/drainage area ratio varies by a factor of 20, which is surprisingly large. The ratios can be divided into three spatial groupings. From Morris Creek to Rock Creek, near the northern end of the range, the ratios are close to one. From Falls Canyon to Milner Creek, in the central section, the ratios vary from 0.5 to 0.2, with most in the range 0.5–0.35. South of Milner Creek, the ratios decrease and average 0.2 between Sabies and Silver Canyons, then decrease further to values less than 0.1 for the three southernmost canyons.

As previously noted by Whipple and Trayler (1996), these groupings are associated with structural features. They also correspond to the fault-style divisions we described previously. The area of large fan/drainage ratios is located where thin fill (<0.5 km) forms a veneer over an extensive pediment (Fig. 1). The rather abrupt transition from ratios approximately equal to one to those less than 0.5 corresponds to the point where a single fault strand along the mountain front transitions southward into multiple strands stepping back into the range. In the zone of distributed mountain-slope faulting (between Falls and Milner Canyons), the broad, shallow piedmont to the west of the mountain front narrows to a small shelf that falls off westward to a deep basin (3–4 km depth).

South of Milner Creek (where fan/drainage area ratios drop to ∼0.2), the narrow shelf pinches out and in the subsurface the bedrock drops directly from the range front into a deep (5–6 km depth) basin. In this section, the range-front faulting is partitioned into strike-slip displacement (often with minor antithetic vertical displacement) on the upper alluvial fans west of the mountain front and normal faulting along the mountain front. Farther south (approximately at Piute Canyon), the deep basin to the west is replaced by a wide, shallow shelf. However, this transition does not appear to be reflected in the fan/drainage ratios.

Finally, south of approximately Gunter Canyon, the broad shelf begins to slope downward into the very deep (∼5 km) basin of the main Owens Valley, and the fan/drainage area ratios drop to values as low as 0.06. In this section, the clear slip partitioning is replaced by complex dextral-oblique motion on multiple fault strands.

Assuming that the fan/drainage area ratio is principally controlled by basin subsidence rate (Whipple and Trayler, 1996), the spatial distribution of the ratios appears to at least loosely correspond to basin depth west of the mountain front. However, there are some significant discrepancies. For example, the greatest basin depth and steepest descent of the bedrock from the mountain front is between Milner and Piute Canyons, but the fan/basin ratios are only moderate in this area. Conversely, the broad, shallow bench between Piute and Gunter Canyons does not correspond to especially large fan/drainage ratios. The similarity of fan/drainage ratios between these two sections may indicate that the deep basin to the west of Milner/Piute Canyons is a product of a past tectonic regime and is no longer subsiding at an unusually large rate.

Another discrepancy arises from comparison of tectonic geomorphology of the mountain front with the fan/drainage ratios. The southern mountain front between Silver and Black Canyons does not appear highly active. The mountain front is sinuous and does not exhibit pronounced facets. The floors of canyons draining the range are somewhat “U” shaped. Fan heads, however, are not entrenched. This mountain front would probably fall in the range of tectonic activity classes 2 (active) and 3 (slow) of Bull (2007). In contrast, the northern section of the White Mountains is characterized by a linear mountain front, planar, very steep interfluves, steep V-shaped canyons, and no entrenchment. It would probably fall in tectonic activity class 1A or 1B (maximal activity). However, this maximally active mountain front has very high fan/drainage ratios, while the moderate-activity southern section has exceptionally low ratios.

At least part of this discrepancy may be explained by the fault-geometry data. The southern section of the White Mountains fault zone is characterized by very steep (>60° and probably >70°) fault planes. The northern section has low dips (∼35°). For a unit of horizontal extension, the steep dip will produce two or more units of vertical displacement. For the same extension, the shallow-dip northern faults will produce only ∼0.5 units of vertical displacement. A highly active mountain front is thus not necessarily incompatible with only a moderate subsidence rate. We note that the transition from very large fan/drainage ratios (∼1) to moderate ratios (<0.5) coincides closely with the transition from ∼35° to ∼45° mountain-front faults between Rock Creek and Falls Canyon.

This observation regarding fault-plane geometry may also help to explain an apparent inconsistency between low unroofing rates inferred from thermochronology and the high activity of the northern segment of the White Mountain front. The U/Th-He cooling ages of Stockli et al. (2003) imply denudation rates on the order of 0.3 mm/yr, or less, and yet, as described previously, evidence from tectonic geomorphology indicates a highly active mountain front. If the mountain-front fault in this segment dips at a relatively low angle, it could maintain a high rate of displacement (high tectonic activity class) with only a moderate or low unroofing rate.

A second important factor is clearly basin sedimentation rate. For the northern section of the White Mountains fault zone, the only significant source of sediment is the White Mountains themselves. In contrast, for the southern section, the Owens River transports in large volumes of sediment from the Sierra Nevada to the north and west. Sedimentation has apparently nearly kept pace with basin subsidence, resulting in a relatively stable base level and thus only a moderately active mountain front. The canyons along this section are partially backfilled with alluvial and colluvial deposits dating to 2.5–2.0 Ma (Bateman, 1965; Lueddecke et al., 1998; Kirby et al., 2006). These deposits are found starting at only ∼100 m above the floor of the Owens Valley, indicating that over the past 2.5 m.y., the valley floor has fluctuated between minor subsidence and aggradation relative to the mountain block. The large sediment supply from the Owens River may thus explain the additional decrease in the fan/drainage ratio from ∼0.2 north of Silver Canyon, where Owens River sediment is not available, to 0.10–0.05 south of Silver Canyon.

The Bishop tuff, erupted at 758.9 ± 1.8 ka (Sarna-Wojcicki et al., 2000), provides an excellent stratigraphic marker across the Owens Valley. As noted by previous investigators (Bateman, 1965; Pinter and Keller, 1995), the tuff forms a broad, asymmetric arch, highest toward the western side of the valley and downwarped toward the edges. On both the west (Round Valley) and east (Hammil Valley) side of the valley, the tuff sheet is not simply buried beneath alluvial fans prograding from the adjacent mountains, but rather it curves smoothly downward in the direction of the facing mountains. When well logs are used to contour the base of the tuff beneath the valley (Fig. 3), the structure is even more apparent. On the east side of the valley (the only one where the well log data were contoured), the tuff has been warped smoothly downward to the east with a vertical displacement of 100–150 m over ∼3 km horizontal distance (Figs. 4A and 4B). Wells were lacking in immediate proximity to the fault zone, and vertical displacement may be even more there. Across the same interval, the present topography slopes upward from the Owens River toward the fault at the base of the White Mountains, with a total elevation difference of 100–150 m. Assuming the topography at the time of the eruption of the Bishop tuff (760 ka) was similar to the present, total downwarping has been at least 200 m and possibly as much as 300 m in some locations. The Bishop tuff is found in outcrop on the footwall side of the fault (Bateman, 1965; Kirby et al., 2006). At the mouth of Silver Canyon, Kirby et al. (2006) documented 400–430 m of vertical separation between such outcrops and the buried Bishop tuff at the mountain front. At the southern end of the area, where the subcrop of the Bishop tuff has been mapped (Fig. 4B), the marker shows ∼260 m of eastward tilting across the valley that is not reflected in the northern profile. This tilting demonstrates that the Coyote Warp and the Owens Valley floor are being deformed uniformly, with consistent downwarping toward the White Mountains fault zone.

Along the eastern edge of the mapped portion of the Bishop tuff subcrop, the depth to the tuff increases smoothly from ∼120 m opposite Coldwater Canyon to 260 m opposite Black Canyon. Given the 760 ka age of the tuff, this implies an increase in the burial rate from 0.16 mm yr⁻¹ at Coldwater Canyon to 0.33 mm yr⁻¹ at Black Canyon. The fan/drainage area ratio at Coldwater Canyon is 0.20. Given the burial rate and the fan/drainage ratio at Coldwater Creek and the burial rate at Black Canyon, Equation 2 predicts that at Black Canyon, the fan/drainage ratio should be 0.10; the actual value is 0.07, corroborating the use of fan/drainage area ratio as a proxy for subsidence rate. The northward decrease in the subsidence rate may be due to a portion of the normal component of displacement stepping westward to the Fish Slough fault.

Subsurface data on the structure of the Bishop tuff are not available on the western side of the valley adjacent to the Round Valley fault, so there we measured profiles up the moraines that lie on top of the Pine Creek alluvial fan (Fig. 4C). The canyon bottom and the two lateral moraines show similar quasi-linear slopes upward from Round Valley except within 1.0–1.5 km of the Round Valley fault. At this distance, they begin to roll over. At the fault trace, the valley bottom is ∼45 m lower than the projected profile, and the moraine crests are ∼100 m lower.

Such anticlinal rollovers have traditionally been considered to result from simple volumetric compensation in the hanging wall during progressive flexure over a curved (i.e., listric) fault surface (Hamblin, 1965; Yamada and McClay, 2003; Twiss and Moores, 2007). There appears to be little doubt that this simple geometrical mechanism can produce pronounced anticlinal rollover in the hanging wall. However, recent numerical and analog modeling has shown that similar hanging-wall deformation can be produced by mechanical stresses during slip on planar faults (Grasemann et al., 2005; Resor, 2008).

At this point, the significance of the pronounced rollovers observed in the Bishop tuff is uncertain. One possible distinction between the stress-induced rollovers on planar faults and the geometrically forced ones on listric faults may be the scale of deformation. The stress-induced rollovers depend on mechanical properties of the rock that are limited in terms of the magnitude and spatial extent of deformation that can be induced. Geometrically caused deformation, however, depends only on the scale of curvature of the underlying fault, which can be over distances of many kilometers. The observed rollover anticlines in the Bishop tuff extend up to 4 km from the range-bounding faults toward which they plunge (for example, at Hammil Valley). This scale of deformation may support an interpretation that these features result from geometrically induced deformation over listric range-bounding faults. At a minimum, the deformation of the tuff is consistent with deformation over listric faults.

Due in part to concerns regarding volcanic activity in the Long Valley caldera, a relatively dense network of seismographs has been installed in the study area since the early 1980s. The operation of this network has coincided with a high level of seismic activity in the northern portion of the study area. Combined with recent advances in hypocenter relocation (Waldhauser and Schaff, 2008), this database has enabled high-resolution imaging of seismic structures beneath the Volcanic Tableland. However, relatively low seismic activity north of the Volcanic Tableland has not allowed similarly refined imaging in that area.

Hypocenters related to the 1984–1986 episode of unusual activity in the region form a distinctive pattern. Figure 5 is a cross section illustrating relocated hypocenters of events between 1980 and 2008 in the latitude interval 37.46°N–37.50°N (under the center of the Volcanic Tableland). Hypocenters of events belonging to three brief swarms in 1984 and 1986 are highlighted. Many of the hypocenters are associated with the Chalfant Valley fault to the west of the White Mountains, but most of those not associated with the Chalfant Valley fault form a band of seismicity dipping westward from a depth of ∼1 km under Chalfant Valley to ∼8 km under the Sierra Crest (10° to 15° dip). This band of seismicity is suggestive of activity on a low-angle fault.

To further test this hypothesis, we examined the dips of events in the latitude interval (37.40°N–37.50°N) that had fault-plane solutions. Figure 6 illustrates the dips of events with reported azimuths between 200° and 300° (i.e., west-directed slip that would be consistent with motion on a west-dipping fault plane). A large proportion of these events have calculated dips <15°, and an even larger proportion have dips <30°. The combination of a plane of subhorizontal seismicity with a large number of low-angle fault-plane solutions on that plane suggests that the seismicity may be produced by a low-angle west-dipping detachment at depths of 4–7 km beneath the Owens Valley.

In this section, we present several types of evidence that bear on the geometry of the faults bounding the northern Owens Valley. These types of evidence include geometrical projections from fault-trace mapping, the spatial distribution of alluvial fan/drainage basin area ratios, distributions of earthquake hypocenters, deformation of the Bishop tuff, and land-surface profiles adjacent to faults. All of these types of evidence indicate that many, but not all, of the bounding faults exhibit moderate- to low-angle dips. In general, the high-angle faults appear to be characterized by lateral or oblique slip and the low-angle ones by dip slip.

To what extent do these moderate- to low-angle fault planes accommodate the active tectonics of the Owens Valley? The mechanical feasibility of low-angle normal faulting is still controversial (Collettini and Sibson, 2001). It is well known that in extensional settings, faults that were originally active as high-angle features can be rotated to much shallower dips (Morton and Black, 1975). Two mechanisms for such rotation have been commonly invoked: domino-style tilting (Proffett, 1977) and rolling-hinge footwall deformation (Wernicke and Axen, 1988). These mechanisms could transform formerly active high-angle faults bounding the Owens Valley into inactive low-angle ones that are not characteristic of the current tectonic regime.

We argue that contemporary extension of the Owens Valley is indeed being accommodated on the low- to moderate-angle structures we have described herein. The fundamental basis for our argument is that the evidence we cite in support of low- to moderate-angle fault geometries is almost entirely derived from active tectonic features. Most importantly, the fault-trace maps from which the fault-plane angles are calculated are based on visible fault scarps and other geomorphic indications of active faulting. Similarly, the alluvial fans used in our analysis are all active fans. The hypocenter locations were measured within the past 40 yr.

With regard to the mechanisms of fault rotation we cite herein, domino-style tilting is quite unlikely to have influenced the faults we have described. The rationale is that we have mapped faults that define the boundary between the Owens Valley terrain and the mountain blocks on either side. Domino-style faulting can affect only subsidiary faults within the hanging wall. On the other hand, given the magnitude of tectonic denudation at the margins of the Owens Valley, we consider it likely that footwall rebound is to some extent rotating the bounding faults we have mapped toward shallower dips. These faults are nevertheless clearly still active.

In Figure 7, we illustrate three alternative interpretations for the tectonic structures responsible for the formation of the Owens Valley. The first is traditional high-angle normal faulting. This interpretation is supported by the measurement of steep dips along some of the margins of the valley (particularly, the northern part of Wheeler Crest and the White Mountains fault zone south of Straight Canyon). However, it is not supported by the measurement of low fault dips (>35°) in outcrop over a considerable portion of the margins of the valley. As described previously, this interpretation also overestimates the total vertical displacement of the valley floor if the modern strain regime is representative of the long-term average.

The second interpretation is one in which the east-west extension across the valley has been accommodated on a shallow, east-dipping detachment that originates along the Sierra Nevada frontal fault (here represented by the Round Valley fault). The primary motivating factors for this interpretation are that it permits ∼5 km of east-west extension without requiring excessive vertical displacement and that the east-dipping Sierra Nevada frontal fault forms the western boundary of extension along the entire Owens Valley, with the exception of the Coyote Warp. Monastero et al. (2002) interpreted reflection seismology data from the Indian Wells Valley, south of the Owens Valley, to show a listric, east-dipping low-angle fault originating at the Sierra Nevada frontal fault. This interpretation is consistent with that structure. Along the east side of the southern Owens Valley, the range front of the Inyo Mountains is relatively inactive, which argues against tectonic structures originating from that side of the valley as the primary mechanism for extension.

The third interpretation accommodates extension on a west-dipping detachment. Like the second interpretation, it allows significant extension without undue vertical displacement. It is consistent with the inferences, based on seismic data, that indicate a west-dipping plane of seismic activity beneath the valley. Finally, it is similar in orientation to the well-evidenced westward-dipping detachments along which Death and Panamint Valleys opened. As previously suggested by Wesnousky and Jones (1994), the low-angle faults in Death, Panamint, and Owens Valley may all sole into a common detachment.

Both of the low-angle faulting interpretations address the vertical displacement discrepancy associated with the high-angle model. Horizontal motion at 4.1 mm yr⁻¹ for 3.4 m.y. would produce 14 km of horizontal displacement. If the motion were on a detachment dipping 15° at depth, ∼4 km of vertical displacement would be produced, which is in agreement with the observed vertical separation.

The east-west component of regional transtension across the Owens Valley has traditionally been thought to be accommodated on high-angle valley-bounding faults. The topography of the valley and at least some observations on exposed fault planes seem to support this model. However, this fault geometry appears to be inconsistent with that of nearby, tectonically associated basins and is also inconsistent with outcrop observations of fault dips as low as 26° bounding the valley. These factors encourage reconsideration of the traditional model.

Calculation of fan/drainage area ratios for canyons on the west side of the White Mountains shows a very large range of variation, by a factor of ∼20. Paradoxically, the highest ratios are at the north end of the range, where the mountain front appears to be highly active, and the lowest ratios are at the south, where the range front appears much less active. However, these variations in fan/drainage ratio are correlated with variations in the dip angle and style of normal/oblique faulting along the mountain front. Fault-plane dip appears to decrease northward, and as it does, the fan/drainage ratio increases. The association between these parameters supports the idea that low-angle faulting is important along at least parts of the valley margin.

The Bishop tuff is an important stratigraphic marker in both the surface and subsurface environments. In localities near the faulted valley margin, where the distal portions of alluvial fans begin to slope upward toward the mountains (and probably also did so at the time the tuff was erupted), the exposed surface of the tuff now exhibits notable anticlinal rollovers. The cross-strike width of the rollovers is in the range of 1–4 km. Although the mechanism for production of rollovers of this scale is somewhat ambiguous, it is plausible that in this case it is produced by geometric deformation over listric valley-bounding faults.

Earthquake hypocenters under the Volcanic Tableland during the 1984–1986 episode of heightened seismicity are arrayed in a subhorizontal planar fashion. Fault-plane solutions for these events include many that are consistent with westward dips <15°. This distribution of hypocenters can be interpreted as events along a westward-dipping detachment fault.

We note that an interpretation of active tectonics that accommodates east-west extension mostly along low-angle faults helps to resolve the discrepancy between geodetically based and geologically based estimates of displacement rates (McCaffrey, 2005). Most geological fault studies in the area have assumed fault-plane dips of ∼60°. If the actual dips are <30°, horizontal displacement rates are increased by a factor of ∼4. This brings them into much better agreement with the geodetic rates.

Direct field observations (triangulation of fault planes, alluvial fans with high fan/drainage area ratios under highly active mountain fronts) argue strongly against the first of the alternative interpretations presented in the discussion section: simple high-angle normal faulting. In contrast, these features are expected from low-angle faulting. The low-angle interpretations are further supported by indirect evidence such as large-scale rollover features and interpretation of seismic events. Finally, the low-angle interpretations allow extrapolation of current strain rates over the known period of transtension to explain the observed vertical displacement of the base of the Owens Valley, while the high-angle interpretation does not. We tend to favor the third interpretation (westward-dipping detachment) over the second because it is supported by the regional pattern and by seismic data, but a clear choice between interpretations is not possible without further evidence.

With regard to the three hypotheses presented under “objectives,” our data clearly do not support the first one: that Owens Valley and the basins to the east all initially opened on high-angle faults, but that the older eastern ones have subsequently rotated to lower dips, inasmuch as the active normal tectonics of the Owens Valley appear to be on moderate- to low-angle faults. The same observations argue against the second hypothesis that high-angle faulting characterized only the Owens Valley. These observations, however, do support the third hypothesis, that the mechanisms of extension in all the basins are similar but that the Owens Valley is at an earlier stage of extension.

The tectonic framework of the Owens Valley is clearly complex and not amenable to any simple classification. Strain is accommodated by strike-slip faulting, normal faulting, oblique faulting, extension distributed across small intravalley faults, and warping and tilting. In many locations, a combination of these mechanisms appears to be active within a few kilometers distance, or even less. The transition along the Round Valley fault from apparently a simple low-angle normal fault (near Pine Creek) to a high-angle dextral-oblique fault at the north end of Wheeler Crest is particularly interesting. It echoes similar variations along the strike of the Death Valley and Panamint Valley fault systems (Cichanski, 2000; Walker et al., 2005) and may imply similar underlying tectonics. In any case, our observations indicate that low-angle normal faulting probably plays an important role along at least portions of the valley margins. Our data cannot unequivocally address two corollary questions: (1) are the low-angle faults planar or listric, and (2) are the low-angle faults simply components of a complex but essentially local tectonic structure or are they manifestations of an underlying detachment that is a controlling structure in relation to regional transtension?

With regard to the first question, we note that a somewhat modified version of the first interpretation (high-angle normal faulting) involving moderate-angle planar faults might still be consistent with our observations. Active slip along planar moderate- to low-angle normal faults has been demonstrated elsewhere in the Great Basin (Abbott et al., 2001; Louie and Pullammanappalli, 2007). If this interpretation is correct, it could allow local tectonics to be explained without recourse to regional detachments. Subsurface fault geometry is difficult to demonstrate without seismic-reflection data, and this may be the only definitive way of addressing this question. Such data are very expensive to obtain; the only present indication of the possible results is from one of the few seismic surveys across the Sierra front, in Indian Wells Valley, which did identify a listric east-dipping detachment fault (Monastero et al., 2002).

The relocated earthquake hypocenter data set presents the most convincing case for answering the second question in the affirmative. If the planar alignment of hypocenters does image a detachment, it is attractive in a regional tectonic context to speculate that it extends north and south under the Owens Valley and that it keys into a regional detachment at 5–10 km depth that integrates displacement between Death Valley and the Sierra Nevada. In summary, our data indicate that any tectonic interpretation explaining extensional opening of the Owens Valley should include some component of low-angle faulting. However, the regional significance of this observation cannot be adequately evaluated without further investigation.