Knowledge of the last reactivation age of faults in the western Basin and Range province (western North America) is key to understanding the recent evolution of the province in the context of the Pacific–North America plate boundary region. The Dixie Valley fault system (DVFS; Nevada, USA) is an important active normal fault in this region, but its slip history is not well known. Our goal is to determine the age of faulting initiation and the history of vertical displacement at the million-year time scale. We used well data from the Dixie Valley geothermal field and geologic maps to calculate offsets of geologic units. Tectonic events and their ages were determined from knickpoint analysis in the Stillwater Range. We then matched basin unconformities and knickpoints to obtain a vertical displacement history. The age of inception of the modern DVFS is likely between 13 and 7 Ma. The total vertical displacement of the geologic units is between 3 and 3.5 km, and most of it occurred in the last 2–3 Ma. Assuming an erosion rate between 0.03 and 0.07 m/ka, the maximum vertical displacement rate (∼1–2 mm/a) occurred between 3 and 0.5 Ma, and resulted in ∼2 km of vertical displacement. Finally, the age-versus-distance profile of the Stillwater Range crest is tapered at both ends, as expected for a propagating range-bounding normal fault, but it also has two maxima matching the deepest parts of the Dixie Valley basin. This indicates that the present-day DVFS may have started out as two separate strands that connected within 2–3 Ma of inception. Our work shows that the combination of stratigraphic and structural data from the basin with geomorphological data from the adjacent mountain range is suitable for placing constraints on the displacement history at the million-year time scale of an active range-bounding normal fault. The resulting displacement-time paths for normal faults may not be as uniform as commonly assumed.
Deformation ages, rates, and magnitudes for continental fault systems in the vicinity of plate boundaries are key pieces of information to test current hypotheses on the lithospheric response to plate tectonics. One of the most prominent examples is the evolution of the Pacific–North America plate boundary and the adjacent >800-km-wide Basin and Range province (western North America) over the past 30 Ma (Atwater, 1970; Atwater and Stock, 1998). The major plate-boundary fault of this system is the San Andreas transform (California, USA), which currently accounts for ∼75% of the dextral motion. The remaining 25% of the motion is widely distributed over numerous fault systems, which include (1) the many faults parallel to the San Andreas fault (e.g., Jennings, 1994; Plesch et al., 2007), (2) the Eastern California shear zone (Dokka and Travis, 1990), (3) the sinistral Garlock fault (California) (Davis and Burchfiel, 1973), and (4) the normal fault systems of the Basin and Range province (e.g., Wernicke, 1981; Eaton, 1982). Process-driven hypotheses at the lithospheric scale have been invoked regarding the formation of these fault systems (Jones et al., 1994; Sonder and Jones, 1999; Dickinson and Wernicke, 1997; Bourne et al., 1998; Flesch et al., 2000; Wernicke et al., 2000, 2008; Dickinson, 2002, and references therein; Unruh et al., 2003). Because these processes differ in both temporal and spatial scales of deformation, the underlying hypotheses are testable and falsifiable by additional knowledge on incremental fault-displacement paths and their interpretation in terms of the present.
Traditionally, fault displacement rates have been inferred, on the basis of a single data point in the past, as uniform either over many thousands of years (e.g., Hetzel et al., 2002, their figure 4) or over millions of years (i.e., since a fault’s inception; e.g., Thompson and Burke, 1973). This resulted in compilations of linear long-term average fault displacement rates, which were then compared to fault-slip rates inferred from instrumental seismological or geodetic data. Because agreement between the measurements was accepted a priori, the discrepancies needed to be discussed in terms of data complexity, inaccuracy, poor measurement quality, or limited reliability of the geologically derived “contemporary” fault displacement rates (cf. Bell et al., 2004; cf. comment and reply by Brown et al., 2005, and Chevalier et al., 2005; Chuang and Johnson, 2011). Authors who have taken an actualistic view have instead pointed to temporal variability in fault displacement rates of a fault or fault system over a range of scales (e.g., Wallace, 1987; Friedrich et al., 2003; Niemi et al., 2004; Chevalier et al., 2005; Kirby et al., 2008; Colgan et al., 2008; Gold et al., 2013). Such variability in fault displacement rate may be understood in process-based frameworks and by retrodiction (e.g., Bennett et al., 2004; Friedrich et al., 2004; Gourmelen and Amelung, 2005; Wesnousky, 2005; Davis et al., 2006). Therefore, in addition to information about deformation age, duration, rate, and magnitude, it is necessary to determine whether deformation over a broad region was coeval, was episodic, or occurred by propagation, which in the case of the Basin and Range province requires knowledge of the slip history of multiple range-bounding faults.
Detailed fault slip histories for many of the faults in the Basin and Range province are not available (for exceptions, see, e.g., Friedrich et al., 2003; McQuarrie and Wernicke, 2005; Koehler and Wesnousky, 2011; Pérouse and Wernicke, 2017). Therefore an obvious prerequisite to the analysis of temporal and spatial deformation patterns in this region is finding a simple way to quickly determine the slip history of several normal faults.
The Dixie Valley fault system (DVFS) is one of these faults. Dixie Valley (Nevada, USA) (Figs. 1, 2) also has a well-developed geothermal energy field. This has resulted in a wealth of geologic data collected over decades in the basin, including dozens of deep wells and seismic reflection profiles, making it an ideal case for developing a simple method to place first-order constraints on the fault slip history.
Here, we use stratigraphic and structural data from the basin and combine them with geomorphologic analysis of the adjacent mountain range to constrain the vertical displacement history of the range-bounding fault system of Dixie Valley. Our approach is also applicable to other similar settings, in particular to other basin-range pairs in the Basin and Range province for which subsurface basin data exist (e.g., Surprise Valley–Warner Range, California; Crescent Valley–Cortez Mountains, Nevada; Railroad Valley–Grant Range, Nevada; Marys River Valley–Snake Mountains, Nevada) and to large rift basins worldwide (e.g., Rhône-Bresse-Rhine graben system, central Europe; range-basin pairs in the central Apennines, Italy). It can be used both when thermochronology data are unavailable, and in combination with these to refine results.
Geology and Tectonics of Dixie Valley
The Dixie Valley fault system is a complex normal-fault network with down-to-the-east motion, in which the main faults are the range-bounding fault and the piedmont faults (Dixie Valley Baseline Conceptual Model; U.S. Department of Energy, 2015). The fault network partly ruptured in the A.D. 1954 Dixie Valley earthquake.
These faults form an asymmetric graben, where most of the motion has taken place on subparallel faults that are located within a 2–3-km-wide zone immediately east of the Stillwater Range front. The resulting basin is filled with Tertiary and Quaternary volcanic and sedimentary deposits that lie on Jurassic basement rock. The basin reaches a maximum depth of 3000 m in southern Dixie Valley, whereas in the north, within the geothermal field, well, seismic line, and gravity data suggest that the basement contact is at ∼2500 m below land surface (Fig. 2; Schaefer, 1982). Its axis is consistently west of the valley centerline (Fig. 2), as expected for an active range-bounding fault system.
There are significant unresolved questions concerning the dip of these normal faults and their connectivity at depth (e.g., Okaya and Thompson, 1985; Benoit, 1999; Blackwell et al., 1999; Abbott et al., 2001, U.S. Department of Energy, 2015). The fault system is 120 km long but only a few kilometers wide, thus we treat it as a single fault and focus on vertical displacement between range and basin center alone. We therefore do not use individual cross-sections, but rather a very schematic lithological and topographic swath profile that includes the area of the Dixie Valley geothermal system and the Stillwater Range just west of it to the top of Table Mountain (Fig. 3). On this profile, along which data are plotted in a roughly NW-SE alignment and which only shows depths and elevations, we simply measure depths of lithologic contacts in wells and their corresponding elevations on the range slopes.
Previous Estimates of Vertical Displacement Rates for the DVFS
There is no coherent displacement history for the DVFS, because previous authors have in all cases calculated average rates from a specific time to present (Table 1). Whereas the most recent displacement rates (for the past 10–20 ka) are useful for earthquake hazard studies, long-term rates (i.e., rates calculated over time spans of 10 Ma or longer) are of limited usefulness (e.g., Friedrich et al., 2003), because a complete displacement history at the million-year scale is necessary in order to understand the evolution of deformation in the entire region over time.
Our goals are therefore to determine (1) the age of faulting initiation of the modern DVFS, and (2) the vertical displacement history of the fault since its inception. In order to accomplish this, we integrated data from the Dixie Valley sedimentary basin and from the Stillwater Range as described below.
DATA AND METHODS
Construction of time-displacement graphs followed Friedrich et al. (2003). By matching geologic unit contacts within the basin with those on the range slopes, we both determined the vertical displacement of each unit and placed first-order bounds on the timing of the displacement. We then identified unconformities in the basin and the equivalent knickpoints in channels on the eastern flank of the Stillwater Range. Finally, we determined the age of occurrence of changes in fault activity from knickpoint-migration analysis, and the complete vertical displacement history of the DVFS.
Geology and Stratigraphy: Vertical Displacement and Ages
Geologic units, with a brief description, age, and vertical displacement, are listed in Table 2. We determined the thickness of all geologic units in the vicinity of the Dixie Valley geothermal field by using published geologic maps (Willden and Speed, 1974; Speed, 1976; Johnson, 1977) in combination with publicly available well logs from geothermal wells (NBMG, 2015; U.S. Department of Energy, 2015). The ages of these units are from published works as well (Appendix 1). A more detailed description of units, which also explains how we estimated thickness, vertical displacement, and age constraints of each unit, is provided in Appendix 1.
We then compared the overall basin stratigraphy, reconstructed from geothermal wells, with the geology on the slopes of the Stillwater Range to determine the vertical displacement of contacts between the different geologic units (Fig. 3). For units present both in the basin and on the slopes of the range (units J-Jm-Jbr, Ts-Tv, Tmb), the contact depth in wells was added to the elevation on the range to get vertical displacement. For basin fill alone (units Tvs, Qal and Qpl), instead we calculated vertical displacements under both the assumption that sedimentation kept pace with displacement, and the assumption that displacement rates outpaced sedimentation rates in the late Tertiary (Fig. 4, Tvs case 1 and case 2, respectively; see Appendix 1 for detailed discussion).
Because we are not aiming for a 10 ka resolution, we can assume constant sedimentation rates during the Quaternary, which means that any significant unconformity in the basin should be tectonic in origin. Other basin features that indicate tectonic origin for the unconformities are the layer dips observed in the seismic profiles (Figs. 5, 6), which are too steep to be simply original depositional angles and are tilted as expected for motion along the DVFS, and the correlation between the westward shift of the basin depocenter and the position of unconformities (Fig. 7).
Concerning the identification of unconformities, we based our work mainly on the structural interpretation from the Dixie Valley Baseline Conceptual Model (U.S. Department of Energy, 2015), especially on seismic reflection lines SRC-3, 102, 6, and 9/104 (Fig. 5; line SRC-3 is not shown separately because it is very close and parallel to line 102). The conceptual model, however, contains line drawings of reflectors, rather than the actual image, for three of the four lines (102, 6, and 9/104). Line 102 is parallel to line SRC-3 and <1 km from it, therefore we were able to verify reflectors between these two lines directly. For lines 6 and 9/104, we compared them with the uninterpreted reflection profiles of neighboring parallel lines from another work (William Lettis and Associates, 1998) to verify whether the traced reflectors are realistic. We also verified the stratigraphy interpreted in the conceptual model against well logs from the NBMG (2015) database, and adjusted the position of geologic unit boundaries on the four profiles where necessary.
In each profile (Fig. 5), we first marked reflector terminations associated with dip changes, then used these to trace all identifiable main unconformities, beginning with the top of unit Tmb (Fig. 5). We also used the depocenter path to help constrain the stratigraphic position of the unconformities, as most of the main unconformities appear to correspond to a faster westward shift of the depocenter (Fig. 7). We were able to identify five clear unconformities in all seismic lines and two additional ones in line 9/104 only. From lowermost to uppermost, the five unconformities are the top of unit Tmb, the top of Tvs, one unconformity in the middle of Qal, another one at the likely boundary between Qal and Qpl, and one in the middle of Qpl. The permissible age ranges of these unconformities can be inferred from Figure 4 and are listed in Table 3. The uncertainties in picking unconformities are given in Table 4, and vary from a few tens of meters to ∼200 m. Uncertainties are due either to regions where no coherent reflectors are present, or to closely spaced unconformable surfaces (Fig. 5). In the first case, we can see a dip change, but there is no information in between. In the second case, we can identify two unconformities close together, and either one could therefore correlate with a knickpoint.
Knickpoint-Migration Analysis and Calculation of Vertical Displacement Rates
In order to determine the age of unconformities at a better resolution and independently from their stratigraphic position, and to constrain displacement rates for different time intervals, we matched knickpoints in channels along the mountain range to basin unconformities (Table 3; Fig. 8). We also used the combination of crest lithology and distance between range crest and range front to constrain the timing of faulting initiation.
Dixie Valley is a basin the existence of which is entirely due to the presence of a range-bounding normal fault system. It is part of the internally drained Great Basin, has been periodically occupied by a shallow (maximum 70 m depth) lake, and has been isolated from the rest of the Great Basin lakes at least since the Middle Pleistocene (Reheis, 1999). The area has not been covered by glaciers, and the climatic conditions prior to ca. 2.5 Ma appear to have been more uniform and humid than in the Quaternary (Thompson, 1991), making repeated base-level drops related to lake fluctuations less likely. Therefore any significant (>>50 m) rapid drop in base level is bound to be local to the valley itself and tectonic in origin; to first order, all significant knickpoints in the channels should thus be tectonic.
Knickpoints form at the range front whenever slip on the range-bounding faults occurs. We cannot of course see every slip increment (i.e., every individual earthquake scarp), which would be well below the resolution limit of the terrain model (Fig. 1), but rather the accumulation of slip during relatively short periods of time (i.e., increases in fault vertical displacement rate). The combination of several dozens of individual events of size similar to that of the 1954 A.D. Dixie Valley earthquake would have to produce a visible step in the profile of channels that are cut by the fault.
The age of a knickpoint can be estimated based on its position along the channel and on the erosion rate of the catchment. Each time a significant knickpoint forms, there should also be a corresponding tectonically induced unconformity of the same age in the basin (Fig. 8), and our goal is to determine the age of that unconformity.
We picked knickpoints by careful inspection of channel profiles. Most channel profiles need to be viewed at two different aspect ratios to confirm all knickpoints (as shown in Supporting Information File S1, Fig. S11): one for the lower part of the channel, and a different one for the upper part. We considered only those knickpoints that either show the typical concave-upward channel segments above and below, or those that produce a deviation from the channel profile that affects a stretch of at least 50–60 m vertically and 100–150 m horizontally (i.e., well within the resolution limits of the digital elevation model [DEM]). All other small steps have been ignored, as many if not all of those are artifacts of the DEM. In addition to the channel shown in Figure S1 (footnote 1), which is located in the central part of the fault, the knickpoints for the three channels that drain into the geothermal area and for one channel in the southern half of the fault are shown in File S1, Figure S2 (footnote 1).
Calculation of Vertical Displacement Rates
We correlated knickpoints with unconformities, from youngest to oldest, on the basis that all knickpoints should have left a trace in the basin at multiple locations (Fig. 8). The knickpoint-unconformity pairs that we identified are listed in Table 3. The unconformities that match knickpoints Ak, Bk, Dk, and crestk are better identifiable than the unconformity matching knickpoint Ck, for which the depth uncertainty is as much as 200 m (Table 4).
The crest of the Stillwater Range, to first approximation, marks the time when the modern DVFS first started moving, because the crest is the oldest knickpoint. Since inception of the DVFS, the crest has migrated away from the range front at a speed that depends on the erosion rate of the area. The crest position is therefore a proxy for fault age, which we calculated following the same procedure described above for calculating the age of other knickpoints. In turn, a plot of age versus distance along the fault (Fig. 10) is a proxy for the shape of the total vertical-displacement function along the fault, and can be used to infer fault evolution.
Age of the DVFS, and Along-Strike Fault Propagation
The possible maximum age of inception of the DVFS is between ca. 13–14 and 7–8 Ma (Fig. 10), if we apply an average long-term erosion rate for the Stillwater Range of between 0.03 and 0.07 m/ka. Determination of erosion rates requires significant discussion and is therefore described separately in the “Discussion” section below. The specific age profile for the crest (Fig. 10) has been derived from knickpoint migration analysis.
The along-strike age profile of the DVFS (Fig. 10) shows the typical tapered ends of an active, propagating normal fault, with older ages (i.e., larger displacements) in the middle (e.g., Scholz, 1990). There is however a stretch that corresponds to a marked change in fault strike (“The Bend” of Wallace and Whitney ; Fig. 2), where the age of faulting initiation is ∼2 Ma younger than expected.
Vertical Displacement History of the DVFS
The vertical displacement history of the DVFS since ca. 13 Ma is provided in Figure 11. This path does not account for all of the observed vertical displacement of the geologic units older than 13–14 Ma, which is an additional 200–700 m. Overall, the path shows slip accelerating gradually until 1.5–3 Ma, then marked acceleration until 0.5–1 Ma, and finally slowing down again toward the present.
The highest vertical displacement rate (between ∼1 and 2 mm/a) occurred between ca. 3 and 0.5 Ma, and produced ∼2 km of vertical displacement.
Age of the DVFS and Erosion Rates
In our calculations of vertical displacement ages and rates, we used a value between 0.03 and 0.07 m/ka for the average erosion rate of the Stillwater Range. These values are based on two observations. The first is that both the northernmost and southernmost parts of the range crest are capped by Tertiary basalts (unit Tmb) with a radiometric age of 13–14 Ma (Fig. 2 and references therein). The crest is the oldest knickpoint, therefore any erosion rate used in calculations cannot return ages of faulting initiation older than 13–14 Ma. The second observation is that the Tertiary volcaniclastic sediments (Tvs) must have at least in part deposited in a developing basin bound by the DVFS, because this unit tends to be thicker near the Stillwater Range front, whereas it pinches out toward the Clan Alpine Mountains to the southeast (Fig. 5). Therefore, faulting must have started prior to or concurrently with deposition of unit Tvs. The age of unit Tvs has not been determined directly in Dixie Valley, but in the Carson Sink there are basalts with an age of 7.7 ± 0.2 Ma (Fig. 2) intercalated within unit Tvs, ∼200 m above its basal contact. We therefore postulate that the age of faulting initiation should not be any younger than ca. 7–8 Ma. In practice, the age of inception of the DVFS may be narrowed down further, by considering that unit Tmb is widespread across the region, has similar thickness in wells and on the ranges, and was therefore likely emplaced on relatively flat topography (Hastings, 1979). Thus unit Tvs may be indeed the very first indication of basin formation in Dixie Valley, making the likely inception age of the modern DVFS younger than 13–14 Ma.
Values of average erosion rates <∼0.03 m/ka produce ages of faulting initiation that are too old, and values of erosion rates >∼0.07 m/ka produce ages that are too young (Fig. 12). Our preferred erosion rates are comparable to the 4 Ma cosmogenic nuclide rates of Jungers and Heimsath (2015) for the Pinaleño Mountains in southeastern Arizona (USA) (0.03–0.06 m/ka) and to the 16 ka cosmogenic nuclide rates of Granger et al. (1996) for the Fort Sage Mountains in California (0.03–0.06 m/ka), and in general are compatible with the low-end rates for large drainages in the United States (∼0.04 m/ka) calculated by Judson and Ritter (1964). The Fort Sage Mountains site in particular is only ∼150 km directly west of the Stillwater Range, is located at a similar elevation, and has a similar annual rainfall.
Along-Strike Propagation of the DVFS
The younger fault initiation age found in “The Bend” (Figs. 2, 10) is likely due to the modern DVFS starting out as two separate segments that only later on merged into the present-day configuration. The age difference of The Bend segment cannot be explained by systematic lithological differences along the strike of the fault, because there is no recognizable pattern (all channel lithologies are listed in File S2 [footnote 1] and the lithology of five of the channels is detailed in Figs. S1 and S2 [footnote 1]; crest lithology is shown in Fig. 10). Also, The Bend is located where the DVFS changes strike from NE-SW to N-S. Finally, the oldest ages of the DVFS correspond to the deepest portions of the Dixie Valley basin (significantly deeper than 2000 m; Fig. 2), whereas The Bend flanks a high in the basin floor (depth <2000 m). It seems thus reasonable to infer that for the first ∼2 Ma of the fault’s existence, The Bend was a relay ramp in the system (Fig. 13). Another possibility for the age minimum in The Bend could be the reactivation of a previous low-angle fault (as reported by Abbott et al.  for The Bend and south to Coyote Canyon, which is about 10 km from The Bend [Fig. 9]). If the central part of the system is exclusively low angle, whereas to the north and south the fault has been reactivated as a higher-angle one, then we would expect smaller vertical displacements in the center, as ages and rates have been calibrated in the geothermal area on the assumption that the fault system has behaved similarly along strike for the past 15 Ma. However, the stretch of fault reported as low angle by Abbott et al. (2001) does not fully correlate with the younger ages we found, as shown in Figure 10. Both the Coyote Canyon channel and the channel immediately north of it (Fig. 9), where Abbott et al. (2001) infer a low-angle modern fault, return older fault ages instead. This points toward the absence of a low-angle reactivated fault at the latitude of Coyote Canyon: either there is no modern low-angle fault at the latitude of Coyote Canyon, and the low-angle fault is instead limited to the stretch between approximately Rough Creek Canyon and Wood Canyon (Fig. 9C), or the low-angle fault hypothesis as an explanation for the anomalous, younger crest ages is not valid.
Vertical Displacement History of the DVFS
Most of the westward shift of the depocenter trajectory is in the lower 1 km of basin fill (unit Qal) (Fig. 7); the logical explanation is that fault slip rate was faster then, allowing the basin to stay close to the range front. When the rate of slip decreased, the basin migrated more slowly too, if we assume that sediment supply rate did not vary much. This is consistent with the larger displacement rates (Fig. 11; Table 3) observed between 0.5 and 3 Ma.
The most likely displacement history (Fig. 11) is one for which the range crest age is younger than the oldest permissible age (13–14 Ma is the oldest permissible age for the crest). The maximum age path results in at least one unconformity, Cu, having an age incompatible with its stratigraphic position (Table 3). Another unconformity, Au, also has a questionable age on the maximum age path, though in this case it is possible that it is actually the boundary between units Qpl and Qal that is misplaced, due to the very limited information about Qpl in well logs. Fault initiation or reactivation ages determined from thermochronology for several other range-basin pairs in the northwestern Basin and Range province (e.g., Colgan et al., 2006a, 2006b; Fosdick and Colgan, 2008; Scarberry et al., 2010) are between 8 and 12 Ma, which also suggests that the younger age path in Figure 11 is more likely.
The lowest possible value for the Pleistocene vertical displacement rate (0.61 mm/a) is three times the Pleistocene rate of dePolo and Anderson (2000) (0.2 mm/a; Fig. 14), but it is comparable to the upper end (0.5 mm/a) of the 12 ka rates found by Bell and Katzer (1990), Caskey et al. (1996), and Bell et al. (2004) (Table 1). In any case, our method cannot resolve displacement at the scale of the earthquake cycle, therefore the youngest segment of the lines plotted in Figure 11 must be considered only indicative. The relatively high value of displacement rates for the Pleistocene (0.61–1.15 mm/a) may for example indicate that there is at least one additional, younger knickpoint-unconformity pair, which is not visible at the resolution of the DEM and seismic lines we had available. Whereas it may be possible in the future to reliably identify additional knickpoints by using a higher-resolution terrain model (e.g., Tandem-X mission data, https://tandemx-science.dlr.de), there are no better published seismic lines for the basin.
There is an extra 200–700 m of vertical offset of units Ts-Tv and the Jurassic units that is not accounted for by slip along the modern fault system. This basement offset is likely related to an older extensional episode. There are several possible candidates for this extension (Fig. 11), because Fosdick and Colgan (2008) identified an unroofing episode at 15–17 Ma in the neighboring East Range, and Hudson et al. (2000) reported an extensional phase for the southern Stillwater Range at 22–24 Ma and possibly a second one at 19–21 Ma. All three of these extensional episodes may have affected the Jurassic units and most of Ts-Tv (with the possible exception of youngest Ts, which may have been affected only by the 15–17 Ma extension).
The amount of unroofing in the southern Stillwater Range between 20 and 24 Ma suggests low-angle normal faulting at this time (John, 1993, 1995). The present-day DVFS may have at least in part reactivated some older low-angle faults south of The Bend, where a low-angle (30°) listric fault has been imaged to 2 km depth by Abbott et al. (2001). In the geothermal area in northern Dixie Valley, however, given the amount of confirmed vertical displacement on the DVFS (∼3.6 km) compared to the width of the Dixie Valley half-graben there (at most ∼4 km), the present-day fault system is likely to be mostly high angle.
Additional Sources of Uncertainties
A source of uncertainty that we have not discussed so far is the estimate of eroded cross-sectional area: we have tested x = 2 and x = 3 in Equation 1, with x = 3 being a more realistic option, but the value may in principle also be lower than x = 2. The age of unit Tmb however still places good constraints on erosion rates. The limit case of x = 1 requires erosion rates of at least 0.08 m/ka to produce an age for the Stillwater Range crest that is not significantly older than 13–14 Ma, and a rate no higher than 0.15 m/ka to produce an age no younger than 7 Ma (Fig. 12). Such erosion rates are somewhat too high when compared to erosion rates for similar settings in the region, but they cannot be excluded a priori. Average erosion rates >0.15 m/ka can however be excluded, as x < 1 is not possible. The plot in Figure 11 therefore represents the maximum solution space, and it is not possible to reliably narrow it down further without additional or better data for the basin, or independently determined erosion rates for the range.
Within the uncertainties due to data quality and availability, we have been able to determine a likely vertical displacement history of the Dixie Valley fault system (Figs. 11, 13), including the age of faulting inception (between 13–14 and 7–8 Ma) and the acceleration of displacement to a maximum of 1–2 mm/a (between ca. 3 and 0.5 Ma). On the basis of its age-distance displacement profile, the DVFS may be the result of the merging of two separate strands north and south of The Bend, which occurred within a couple of million years from the time displacement started (Figs. 10, 13). Whereas the southern DVFS may have some low-angle faulting (i.e., 30° or less), this seems unlikely in the area of the geothermal field, based on the constraints on the amount of total vertical displacement compared to the width of the half-graben there. Finally, we constrained the average erosion rate in the Stillwater Range for the past 13–14 Ma to between 0.03 and 0.07 m/ka, and most likely no higher than 0.05 m/ka (Fig. 12).
Our work shows that combining stratigraphic and structural data from the basin with geomorphological data from the adjacent mountain range is an effective way of using data that for the most part already exist in literature to place reliable constraints on the displacement history at the million-year time scale of active range-bounding fault systems. The next step would be the application of this method to several other basin-range pairs in the region, possibly also testing it against thermochronology data at suitable locations.
All data for this paper are properly cited in either the main text, in Appendix 1, or in the Supporting Information (footnote 1), and referred to in the reference list. We thank two anonymous reviewers and the Associate Editor, C. Amos, for their insightful comments and suggestions. AMF would like to thank John Caskey for having introduced her to the paleoseismological setting of the Dixie Valley fault system.
APPENDIX 1. DETAILS AND DISCUSSION OF LITHOLOGY AND AGE OF STRATIGRAPHIC UNITS, AND OF THEIR VERTICAL DISPLACEMENT
Quaternary Playa Sediments (Qpl)
Characteristics of Unit
Playa sediments (unit Qpl) are composed mainly of clays, with some silt, and tend to be present mostly in the lowest part of Dixie Valley (Huntington et al., 2014). The measured thickness close to the range front varies from 480 to 900 m (Fig. 2), and playa sediments interfinger and grade into alluvial fans close to the foot of the range (Huntington et al., 2014). Alluvial sediments are present below Qpl too. In well 62A-23 (Fig. 3), the thickness of alluvium below Qpl is comparable to the thickness of Qpl itself. We did not find any estimates of Qpl thickness for other wells. An indication of Qpl thickness in the basin is given by seismic profiles (Okaya and Thompson, 1985; U.S. Department of Energy, 2015). These show a package of reflective, continuous, thinly bedded layers reaching maximum depth of ∼1.2 km below the basin’s surface. This package lies unconformably above a package of less-continuous layers. The boundary between the two correlates with the base of Qpl in well 62A-23.
The oldest possible age for the base of this unit should be the oldest age for Pleistocene lakes reported for this area. Caskey et al. (2000) hypothesized 0.2–0.5 Ma for the unit they name “Qbfo”, which they consider equivalent to Lake Lahontan Late Pleistocene deposits. Reheis (1999) and Reheis et al. (2002) assigned the oldest Lake Dixie deposits to the Middle Pleistocene highstand of Lake Lahontan, therefore the base of Qpl could be as old as 0.65 Ma. Playa sediments are still forming today in the Humboldt salt marsh (Fig. 2).
Minimum displacement. Minimum displacement would be zero if the entire thickness of the unit had been deposited in preexisting space. Of course this is not realistic, because the DVFS is still active. The reasonable minimum displacement is therefore probably closer to 100–200 m.
Maximum displacement. Maximum displacement would be the maximum measured Qpl thickness from Huntington et al. (2014) (900 m) if sedimentation of Qpl kept pace with vertical displacement rates.
Quaternary Alluvium (Qal)
Characteristics of Unit
Quaternary alluvium (unit Qal) consists of unconsolidated alluvial fan sequences with silt, sand, gravel, cobbles, boulders, and locally minor interbedded clay (Huntington et al., 2014). In the Dixie Valley geothermal field, the Quaternary deposits as a whole are between 1300 and 2500 m thick. In well 62A-23 (Fig. 3), for which a measurement of Qpl thickness exists, Qal alone is 900 m thick. Its maximum thickness is ∼1200 m. The oldest age is constrained to be the base of the Quaternary (2.5 Ma), for lack of any further information about the age of this unit. Alluvial fans are still forming today, therefore the youngest age is 0 Ma.
Minimum displacement. In principle, the minimum depth of Qal base in wells within 2 km of the Stillwater Range front (1400 m) should be taken as the minimum displacement. However, in the limit case, the displacement could be zero if all of the Quaternary sediments were deposited into a preexisting empty basin. The latter option seems unrealistic, though, given that the DVFS is still active and was likely active during the rest of the Quaternary as well. Because in Dixie Valley the ratio between basin depth and range crest elevation is closer to 2/3 than to 1/2 (i.e., the basin is “overfilled”), we then take 2/3 of 1400 m as a reasonable minimum displacement (i.e., ∼900 m).
Maximum displacement. Maximum displacement would be the maximum depth of Qal base in the center of the basin (2500 m), assuming that sedimentation kept pace with displacement along the fault.
Tertiary Volcaniclastic Sediments (Tvs)
Characteristics of Unit
Unit Tvs consists of partially consolidated volcaniclastic sediments (Huntington et al. 2014). In the Dixie Valley geothermal field, this unit has a maximum thickness of ∼500 m (<300 m in most wells) and no clear age constraints, but a similar volcaniclastic unit in the northern Carson Sink (Fig. 1) reaches 600–700 m thickness and contains two dated basalt flows (ca. 8 Ma in the lower third of the unit, and ca. 4 Ma near the top, though the latter is likely a sill; Stewart et al., 1994). It does not crop out in the Stillwater Range west of the geothermal field, where any volcaniclastic sedimentary rocks are in all cases located below Tmb. In wells in Dixie Valley, Tvs is in nearly all cases above the Tertiary basalts (only a couple of wells have Tvs intercalated in the basalt flows), therefore we constrain the age of this unit to be equivalent to that of the Tertiary volcaniclastics found in the Carson Sink above the main basalt flow package (ca. 8 Ma). Its youngest possible age is 2.5 Ma (base of Quaternary), for lack of better constraints. Tvs is clearly differentiable from Qal because of the type of sediments (more abundant tuffaceous material in Tvs) and the higher degree of consolidation (Maurer et al., 1995; Huntington et al., 2014), but this unit has not been dated independently in wells.
There are two possible end members for the timing of Tvs deposition:
• Case 1: Tvs originally deposited on top of Tmb, then was fully eroded on Table Mountain but preserved in the valley, where it can now be found in wells. In this scenario, after the deposition most of Tvs, the DVFS must have accumulated 3000 m of vertical displacement (value controlled by the well-constrained Tmb basalts base at ca. 13 Ma, less so by the Tmb top, which may be ca. 8 Ma). Also, displacement could not have started until Tvs deposition was completed, because otherwise Tvs would only have been deposited in the basin.
• Case 2: Tvs never deposited on top of Tmb on Table Mountain, and only deposited within the basin and was not eroded before being covered by Qal. All of Tvs would have deposited after most of the Tertiary displacement had ended.
Minimum displacement. Minimum displacement would be the minimum depth of Tvs base in wells (1600 m) if case 2 applies.
Maximum displacement. Maximum displacement would be the vertical distance between top of Table Mountain and the deepest position of Tvs base in wells (a displacement of 3200 m), if case 1 applies.
Tertiary Basalts (Tmb)
Characteristics of Unit
Unit Tmb consists of flows of porphyritic olivine basalts and basaltic andesites (Nosker, 1981; Morton et al., 1977; Evans and Brown, 1981; Stewart et al., 1994; Alm and Walker, 2016) with thickness in the Dixie Valley geothermal wells of ∼300–400 m, with a maximum of 500 m. The basalts on Table Mountain have a maximum thickness of ∼300 m, which can be inferred from the geologic map of Willden and Speed (1974). The radiometric ages of Tmb in the area of interest vary from ca. 4 Ma (Stewart et al., 1994) to 17 ± 0.8 Ma (Evans and Brown, 1981). The only sample from the northern end of Table Mountain has been dated at 13.4 ± 0.5 Ma (Nosker, 1981) and comes from near the contact between Tmb and Tv (Fig. 2). A sample in a similar stratigraphic position at the southern end of Table Mountain has an age of 13.3 ± 0.4 Ma (Stewart et al., 1994). Hastings (1979) reported also younger ages of ca. 8 Ma for basalts at ∼2000 m below the northern Carson Sink (7.7 ± 0.2 Ma; Stewart et al., 1994) (Fig. 2), and ages as young as ca. 5 Ma are reported by Stewart et al. (1994) for a drill core beneath Soda Lake in the southern Carson Sink.
For vertical displacement estimates, we consider mainly the basalts of northern Table Mountain because, given their location, these are most likely lateral equivalents of the basalts found in the geothermal wells at the foot of the range itself (Fig. 2). Hastings (1979) pointed out that the similar thickness of basalts in wells and in the surrounding mountains indicates that little topography should have been present at the time of emplacement of the basalts, excluding therefore a major extensional episode. These basalts are present at about the same elevation on top of the Clan Alpine Mountains, supporting the hypothesis of a fairly flat terrain during most of the emplacement of the basaltic flows. Tmb also caps the Stillwater Range crest in several places, making it very likely that this unit is indeed the last one emplaced before the start of Basin and Range extension. Alm and Walker (2016) inferred from a perched Neogene lacustrine basin (likely the same Ts unit found in the geothermal field) in La Plata Canyon (southern Stillwater Range; Fig. 2) that early Basin and Range extension here is younger than the age of the basalts that cap the lacustrine sequence (14.34 ± 0.04 Ma).
Minimum displacement. Minimum displacement would be the vertical distance between the position of Tmb base in well 28-33 (1200 m elevation b.s.l. [below sea level]; Fig. 3) and the base of Tmb on Table Mountain (1800 m elevation a.s.l. [above sea level]). The minimum vertical displacement is therefore 3000 m. Well 28-33 was chosen because the base of Tmb is not faulted, and the thickness of the Tmb section here is similar to that on Table Mountain.
Maximum displacement. Maximum displacement would be the vertical distance between the deepest position of Tmb base in the basin (see Dixie Valley conceptual model cross sections [(U.S. Department of Energy, 2015]) (1800 m elevation b.s.l.) and the base of Tmb on Table Mountain (1800 m elevation a.s.l.). The maximum vertical displacement is therefore 3600 m. This value is more realistic than the minimum value because it accounts for the depth of all the basalts in the basin.
Tertiary Lacustrine Sediments (Ts) and Tertiary Volcanics (Tv)
Characteristics of Unit
Rocks of units Ts and Tv overlap in space and time and cannot be easily separated, therefore we treat them as a single unit in order to measure vertical displacements, with the understanding that in general most of Ts is younger than most of Tv and in higher stratigraphic position.
Unit Ts is mainly lacustrine sediments (volcaniclastic sediments and carbonaceous siltstones) (Waibel, 1987). It is found in most of the Dixie Valley geothermal wells, and on the Stillwater Range slopes west of the geothermal field, in all cases below Tmb when the two units are found in stratigraphic contact. It has not been dated directly (Waibel, 1987).
Willden and Speed (1974) called some of these sedimentary rocks “younger sedimentary rocks from the Pliocene”, coeval with “younger basalts” (also assumed to be Pliocene in age). From the comparison of published maps with the location of samples, however, it is clear that the “younger basalts” are in fact our Tmb unit, because the dated samples of basalts in the area all have coordinates that match the locations of the basalt outcrops shown in the geologic maps of Willden and Speed (1974) and Johnson (1977). Therefore, given that the age for the base contact of Tmb on Table Mountain is older than Pliocene, and that these sedimentary rocks are in all cases mapped stratigraphically below Tmb, all of the “younger sedimentary rocks” mapped on Table Mountain immediately west of the geothermal field must in fact be Miocene in age, and we have therefore shown them as Ts-Tv on our map in Figure 2. The youngest possible age for the top of Ts near Table Mountain corresponds to the oldest basalts above it, i.e., ca. 13 Ma, but it could be older.
Unit Tv is composed of silicic tuffs and flows that show up in significant thickness in only some of the geothermal field wells (likely faulted out of section at the location of other wells; Waibel, 1987) but are quite thick in the Clan Alpine Mountains to the east and also crop out on the Stillwater Range below Tmb. Waibel (1987) considered them to be Oligocene in age, and radiometric data on these rocks compiled by Hudson et al. (2000) confirm that the age range is ca. 23–27 Ma. At the latitude of the geothermal area, there are no exposed Tertiary intrusive rocks, unlike in the southern part of the Stillwater Range.
Any displacement in these units is pre–Basin and Range and it should have started no earlier than ca. 24 Ma, based on the age of steeply tilted Stillwater caldera tuffs within Tv in the southern Stillwater Range (Hudson et al., 2000). For this early episode, Hudson et al. (2000) also constrained the end of displacement, given by the likely 22 Ma tuffs that lie unconformably over the tilted caldera tuffs. A second possible extensional episode has been identified by Hudson et al. (2000) based on the presence of intruded dikes with an age range of 19–21 Ma. Given the age range of this unit, however, its youngest parts (Ts) might have undergone their earliest displacement only during the unroofing episode recorded by Fosdick and Colgan (2008) at 15–17 Ma.
Minimum displacement. Minimum displacement would be the vertical distance between the shallowest, non-faulted base of Tv (well 28-33) and the minimum elevation of Tv base within the range just west of the geothermal field (1600 m b.s.l. + 1600 m a.s.l. = 3200 m) (Fig. 3).
Maximum displacement. In principle, maximum displacement should be the vertical distance between the deepest, non-faulted base of Tv (well 28-33) and the maximum elevation of Tv base within the range just west of the geothermal field (1600 m b.s.l. + 1800 m a.s.l. = 3400 m). We would argue however that the good constraint on displacement value is only apparent due to the fact that Tv is missing in most wells, and commonly in its place there is a fault. In fact, if we assume that the thickness of Tv in well 74-7 was originally similar to that in well 28-33, the base of Tv would be located at 2000 m b.s.l., and the maximum displacement would therefore be 3800 m.
Jurassic Rocks (J)
Characteristics of Units
Two distinct units have been mapped near the geothermal field by Willden and Speed (1974) and Johnson (1977). Most of these rocks (unit Jm) are mafic rocks that belong to the Humboldt igneous complex and represent old oceanic crust. The other unit is the Boyer Ranch Formation (unit Jbr), which consists of quartzites that occur below and within the lower part of Jm and can be seen in well 28-33 (Fig. 3). All of these rocks are much older than the rest of the sequence above and form the crystalline basement of the Dixie Valley basin.
The maximum displacement along the DVFS shown by the Jurassic units is not larger than the likely displacement of the Tertiary IXL pluton (Willden and Speed, 1974), which is located between Rough Canyon and Coyote Canyon (Fig. S2 [footnote 1]), and the displacement of which is post–24 Ma (Hudson et al., 2000). Therefore, the displacement of the Jurassic units must also be post–24 Ma.
Minimum displacement. Because Jurassic units form a crystalline basement much older than any of the units above, the minimum displacement of the Jurassic rocks cannot be any less than the minimum displacement possible for Tv above it, therefore the minimum reasonable displacement for the Jurassic units is 3200 m.
Maximum displacement. Maximum displacement might be considered to be the vertical distance between the base of Jbr in well 28-33 (2000 m b.s.l.) and the highest contact between Jm and underlying Triassic rocks on the mountainside (1700 m a.s.l.), i.e., 3700 m total displacement. However, in over half of the wells, the base of Jurassic rocks is not reached. If we assume that in these wells the thickness of Jurassic rocks is similar to that found in wells 62A-23 and 62-21, then the maximum displacement could be up to 4300 m.