Evolution of Miocene normal and dextral faulting in the lower Colorado River region near Blythe, California, USA

Deformation related to the Pacific–North America transform plate margin is distributed along several NW-striking dextral fault systems in the interior of North America east of the San Andreas fault. Geodetic studies demonstrate that the eastern California shear zone (ECSZ) accommodates as much as ~29% of the modern relative plate motion across a zone up to ~100 km inboard (northeast) of the primary plate boundary (Miller et al., 2001). However, inception of a paleo-ECSZ (sensu Dorsey et al., 2021) began as early as ca. 10 Ma (Dokka and Travis, 1990; Schermer et al., 1996; Nuriel et al., 2019) and was likely distributed across an even wider zone of deformation that is not well defined and has limited fault slip constraints. A global plate circuit model (Atwater and Stock, 1998) and subsequent block reconstructions suggest that plate-margin–connecteddextral shear may have jumped inland from the offshore transform plate boundary during the Late Miocene (McQuarrie and Wernicke, 2005; Bennett et al., 2016). At the latitude of the lower Colorado River region, deformation related to the Pacific–North America transform plate margin is located well inboard of the plate boundary with Late Miocene and younger dextral slip documented on NW-striking faults along the California-Nevada border (Guest et al., 2007) and in western Arizona where NW-striking dextral faults offset shallowly dipping detachment faults in the Colorado River extensional corridor (Singleton, 2015). A better understanding of the spatial extent and temporal evolution of the paleo-ECSZ and fault interactions with older structures is essential to characterize the evolution of the Neogene Pacific–North America margin and could inform our understanding of the evolution of transitional diffuse plate boundaries across the globe, which account for ~15% of Earth‘s surface (Gordon, 1998).

The paleo-ECSZ, which was composed of Miocene-aged faults east of the Quaternary-active ECSZ, may have included right-separation faults near the California-Arizona border surrounding Blythe, California (e.g., Richard, 1993; Bennett et al., 2016; Dorsey et al., 2021; Fig. 1), but it is unclear if these faults are temporally and kinematically compatible with paleo-ECSZ deformation. A structural grain defined by NW-striking faults is predominant in regions along strike from Blythe to the northwest and southeast and to the southwest of Blythe closer to the San Andreas fault system. However, northeast of Blythe, the dominant structural grain is a result oflarge-magnitude extension accommodated by NE-directed detachment faulting and development of metamorphic core complexes in the Early to Middle Miocene (e.g., Foster and John, 1999). NW-striking brittle faults withkm-scale displacement of Mesozoic and Cenozoic rocks and structures are prevalent in the region surrounding Blythe and may have driven Neogene structural control of landscape evolution (e.g., Bennett et al., 2016; Dorsey et al., 2017, 2021; Thacker et al., 2020). However, the timing and kinematics of these NW-striking faults are poorly constrained, and they have been variably interpreted to record Early to Middle Miocene extension as dip-slip faults (Rotstein et al., 1976; Hamilton, 1982, 1984) or Late Miocene to Pliocene dextral strike-slip (Richard, 1993; Salem, 2009; Bennett et al., 2016). Compilation of the cumulative dextral shear accommodated on faults across southeastern California and western Arizona by Bennett et al. (2016) indicates that the measured magnitude of dextral shear in the lower Colorado River region is ~26 km, which is ~17–33 km less than the 43–59 km of cumulative dextral shear known in regions along strike to the northwest (eastern Mojave Desert) and southeast (southwestern Arizona). From this, Bennett et al. (2016) suggested that dextral shear may be significantly underestimated in the lower Colorado River region, particularly if the eastern Mojave, lower Colorado River, and southwestern Arizona regions are kinematically linked, as the similar deformation styles between them might suggest.

In this paper, we use structural analysis of NW-striking faults near Blythe to evaluate the kinematic evolution of fault slip and determine if these faults are more compatible with regional extensional, dextral, or dextral transtensional strain regimes. We use ⁴⁰Ar/³⁹Ar and U-Pb dating to assess when fault slip occurred and a compilation of shear magnitude estimates to determine if these faults could have accommodated sufficient dextral shear to balance the apparent shortfalls in along-strike regional kinematic budgets.

Geologic mapping of the mountain ranges northwest of Blythe has identified numerous NW-striking faults (Fig. 2), many with 100-m- to km-scale right-lateral separation of Mesozoic markers (Emerson, 1981; Hamilton, 1984; Stone and Pelka, 1989; Ballard, 1990; Salem, 2009; Stone et al., 2022). These NW-striking faults cut folds and thrust faults of the S- to SW-vergent Cretaceous Maria fold-and-thrust belt (of Reynolds et al., 1986), and, for most faults, brittle slip can only be bracketed to postdate these Late Cretaceous structures and predate the Quaternary fan surfaces that overlap almost all fault traces. Evidence for Quaternary fault slip is present only at the Blythe graben (BG, Fig. 2), and only a few faults have clear evidence of Cenozoic activity. For example, a fault in the Big Maria Mountains cuts a chain of intermediate to felsic intrusions dated to 21.7 ±2.8 Ma (hornblende K-Ar; Martin et al., 1982), indicating this fault is Early Miocene or younger. Cenozoic rocks are also faulted at Palen Pass, where the Packard Well fault zone (PW, Fig. 2) cuts a previously undated conglomerate that Stone and Kelly (1989) considered to be broadly Tertiary in age. For other faults northwest of Blythe, no such offsets of Cenozoic rocks are exposed, and the timing of fault slip is poorly constrained. Southeast of Blythe, Oligocene to Middle Miocene volcanic and clastic rocks are exposed in the hanging wall of the shallowly W-dipping ―south Dome Rock Mountains fault‖ on the west side of the Dome Rock Mountains (Spencer et al., 2016; Ferguson et al., 2018; Brickey, 2022). In the footwall, Mesozoic metasedimentary and metavolcanic rocks are cut and displaced by steeply dipping, NW-striking faults with both right- and left-separation of map units that dip SE (Stone, 1990; Richard et al., 1992; Johnson et al., 2021). The lack of constraints for fault timing, dip directions, and slip vectors obfuscate whether right- and left-separation faults in the Dome Rock Mountains are conjugate normal faults (e.g., Johnson et al., 2021), strike-slip faults with opposing slip senses formed during different deformation episodes, or some combination thereof.

An elliptical gravity low (herein called the Blythe gravity low; Fig. 2) extends from the Colorado River near the town of Blythe into the valley to the northwest. Inversions of gravity data indicate 4–5 km of sedimentary basin fill (Saltus and Jachens, 1995) overlying higher-density basement rocks, and the gravity low may mark the deepest concealed sedimentary basin in the region (Rotstein et al., 1976). Reported sharp linear gradients in both gravity and electromagnetic data (Rotstein et al., 1976; Hoover et al., 1981; Hamilton, 1984) on the southeast flank of the Big Maria Mountains bound the gravity low and suggest structural control of basin geometry and perhaps structurally controlled deposition patterns of Miocene and younger basin-fill strata.

Subsurface records indicate that the base of the 4.5–3.5 Ma Bullhead Alluvium locally lies at least 62 m to as much as >100 m below sea level near the town of Blythe, well below a river profile graded to Pliocene sea levels (Howard et al., 2015; Crow et al., 2019). Compaction of underlying basin fill is insufficient to explain this elevation discrepancy, suggesting that Pliocene or younger faulting and/or isostatic subsidence down-droppedColorado River deposits (Thacker et al., 2020). Low-displacement (decameter to centimeter scale) faults locally cut the Early Pliocene Bouse Formation; notable examples are exposed in outcrops near Parker, Arizona, and Cibola, Arizona (e.g., Buising, 1992; Gootee et al., 2016; Thacker et al., 2020). Pliocene growth strata in the hanging wall of W-dipping and SE-dipping normal faults near Cibola reflect syndepositional tectonism near the time of Colorado River integration, and Dorsey et al. (2021) suggested a model of a paleo-ECSZ releasing stepover basin between dextral NW-striking faults near Blythe and the Laguna fault system to the south. However, the role of NW-striking faults near Blythe in controlling basin subsidence and nascent Colorado River sedimentation is unclear, and it remains possible that these faults were kinematically or temporally incompatible with that model.

To determine the geometry, kinematics, and strain fields of faulting in the Blythe region, we collected fault kinematic data from mapped NW-striking faults that likely record Cenozoic slip. Sense of slip on faults was determined in outcrop using the fracture geometry and surface texture criteria of Petit (1987) and Doblas (1998). For some measurements where outcrop-scale kinematic indicators are not observed, we assume slip sense from offset of geologic features, either at the outcrop scale or along the trace of a mapped fault, or we assume slip sense from parallel faults of known slip sense. We used Stereonet 10.2.9 software (Allmendinger et al., 2012; Cardozo and Allmendinger; 2013) to plot planar and linear measurements and FaultKin 7.5 software (Marrett and Allmendinger, 1990; Allmendinger et al., 2012) to determine incremental shortening axes (P-axes) and incremental extension axes (T-axes) and evaluate kinematic compatibility. Averages were computed using maximum eigenvectors and reported for each stereoplot figure in Dataset S11 and Mavor et al. (2023). Stereoplot Kamb contours to linear features and poles to planes were calculated with interval = 2σ, significance level = 3σ, grid nodes = 20, and using an exponential smoothing algorithm, except where these parameters produced anomalous contouring patterns (in which case other settings are noted). Fault measurements were separated into principal slip surface measurements or measurements along subsidiary fault surfaces. Additional fault orientations of mapped faults were determined using the PlanarOrientationsTools Python scripts (Haugerud, 2020) extracted from 1-m-resolution lidar topographic data (U.S. Geological Survey, 2020) for faults northwest of Blythe and a photogrammetry-derived digital elevation model (DEM) from 2017 National Agriculture Imagery Program (NAIP) imagery (University of Arizona, 2019) for faults in Arizona.

We also collected samples for geochronologic analysis at five sites (Fig. 2). Detailed methodology for ⁴⁰Ar/³⁹Ar and calcite U-Pb geochronology is available in Item S1, and the resulting data from these methods are available in Datasets S2 and S3, respectively, and Mavor et al. (2023). Interpreted dates are reported with 2σ uncertainty and with a mean square of weighted deviation (MSWD) value to assess the data scatter.

Several NW-striking faults are exposed within or along the margins of the Big Maria Mountains (Fig. 2). The Eagle‘s Nest fault (Hamilton, 1984; informally named by Salem, 2009) is a NW-striking brittle fault that transects the Big Maria Mountains with an ~24-km-long, discontinuously exposed trace (EN, Fig. 2). In the central and northwest parts of the range, the fault is marked by a well-defined lineament of recessive topography and dips moderately to steeply NE. The Eagle‘s Nest fault cuts Proterozoic through Mesozoic rocks but does not intersect mapped Cenozoic units. East of the Eagle‘s Nest fault, existing geologic mapping of the Big Maria Mountains (Hamilton, 1984) depicts another fault with a sharp curve that connects a NW-striking, NE-dipping segment in the central Big Maria Mountains with a NE-striking, SW-dipping segment in the northeastern Big Maria Mountains (from MW to ST, Fig. 2). In contrast to a sharply curved fault, we mapped the NE-striking fault to be truncated against a throughgoing NW-striking fault (Fig. 3). We propose the name ―Slaughter Tree Wash fault‖ for the NE-striking fault that continues to the northeast of the Big Maria Mountains in the drainage of the same name. As mapped by Hamilton (1984), the Slaughter Tree Wash fault cuts a WNW-trending chain of rhyolite and dacite intrusions, one of which was dated by the hornblende K-Ar method to 21.7 ± 2.8 Ma (Martin et al., 1982). We propose the name ―Maria Wilderness fault‖ for the NW-striking fault that truncates the Slaughter Tree Wash fault and displaces the Big Maria syncline in the central Big Maria Mountains (Fig. 3). From the intersection with the Slaughter Tree Wash fault, the Maria Wilderness fault traces northwestward along the southwestern toe of the range (Fig. 3), for a total exposed trace length of ~20 km in the Big Maria Mountains. We postulate that the Maria Wilderness fault merges with the Eagle‘s Nest fault where both are concealed in the northwestern Big Maria Mountains. We use the 2.3 km of right-lateral separation of the Big Maria syncline (pink lines on Fig. 2) from Hamilton (1984) to infer that the offset equivalent of the Slaughter Tree Wash fault is concealed by alluvium on the southwest side of the Maria Wilderness fault (Fig. 3).

In the northeastern Big Maria Mountains, west of Quién Sabe Point (QP, Fig. 2), NW- and NNW-striking faults cut Cretaceous to Proterozoic crystalline rocks and project SE toward the NE-striking Slaughter Tree Wash fault. The NW-striking Quién Sabe fault (locality QS on Fig. 2; Richard, 1993; Dembosky and Anderson, 2005) cuts an undated landslide breccia with volcanic and metamorphic clasts considered to be questionably Miocene by Hamilton (1984). The intersections between these structures are concealed by alluvium, and the NE-striking Slaughter Tree Wash fault is shown either cutting (Hamilton, 1984) or displaced by (Hamilton, 1964; Carr, 1991) the NW-striking Quién Sabe fault. Additional concealed NW-striking right-separation faults are postulated between the Big Maria and Riverside Mountains (the Big Wash fault; BW in Fig. 2) and between the Riverside and Whipple Mountains; these faults are postulated to cut Early to Middle Miocene detachment systems (Carr, 1991; depicted as geophysically defined faults in Fig. 2).

The southeastern Big Maria Mountains host an even more complicated fault architecture with several unnamed faults that branch between the Eagle‘s Nest fault and the Maria Wilderness fault (Fig. 4). The southwesternmost of these faults (Fig. 4G) strikes north-northwest and likely splays from the Eagle‘s Nest fault, though the intersection is concealed by alluvium. As mapped by Hamilton (1984), the Eagle‘s Nest fault does not continue to the southeasternmost parts of the Big Maria Mountains, but branching fault connections between the Eagle‘s Nest fault and the Maria Wilderness fault (e.g., Fig. 4F) could suggest that strain is partitioned to the latter in this part of the range.

The principal slip surface of the Slaughter Tree Wash fault is exposed in the eastern Big Maria Mountains (locality ST on Fig. 2) and is sub-vertical to steeply SSE-dipping with NE-raking slickenlines (Fig. 5A). Map relations of supradetachment rocks in the Slaughter Tree Wash fault hanging wall at Quién Sabe Point juxtaposed against subdetachment rocks in the Slaughter Tree Wash fault footwall suggest southeast-downdisplacement, and this relationship combined with the oblique slickenlines indicate normal-sinistral oblique slip. Measured subsidiary fault zone surfaces are likewise ENE-WSW striking, subvertical, and have NE-rakingslickenlines on average. P-axes (incremental shortening axes) plunge moderately north, and T-axes (incremental extension axes) plunge southeast (Fig. 5A).

We subdivide the fault kinematic data along the Maria Wilderness fault into three sections: the southeastern Big Maria Mountains (Figs. 4A and 4B), the central section (Fig. 5B), and the northwest section (Fig. 5C), and likewise divide kinematic data of the Eagle‘s Nest fault into data collected in the southeast (Fig. 4E), northwest (Fig. 5D), and central (Fig. 5E) parts of the Big Maria Mountains. Several outcrops of the principal slip surface of the Maria Wilderness fault both southeast and northwest of the intersection with the Slaughter Tree Wash fault expose similar fault surface orientations that dip ~54°–59° to the northeast with steeply SE-rakingslickenlines (Figs. 5B and 5C). The core of the fault hosts a 3–5-m-wide zone of foliated fault gouge and brecciated wall rock. In one location, gouge foliation is systematically oblique to the principal slip surface exposed in a nearby outcrop and serves as a dextral-normal oblique slip kinematic indicator (green symbols on Fig. 5B); a calculated slip vector from the gouge and fault orientation using the graphical method of Moore (1978) is 127°/07°, which is more shallowly SE-plunging than the average slickenline at that locality (093°/41°; Dataset S1 [^{footnote 1}]; Mavor et al., 2023).

In the northwestern part of the Big Maria Mountains, the principal slip surface of the Eagle‘s Nest fault dips steeply northeast with highly variable slickenlines that on average rake steeply southeast (E-plunging; Fig. 5D). Subsidiary fault surfaces are variably oriented but are largely parallel to the steep northeast dips of principal slip surface exposures, with more shallowly raking slickenlines on average. At fault exposures in the central Big Maria Mountains, the Eagle‘s Nest fault principal slip surface measurements are NE-dipping with nearly down-dip NE-plunging slickenlines (Fig. 5E).

Faults in the southeastern Big Maria Mountains have slickenlines that are steeply SE-raking, with lesser down-dip or subhorizontal lineations and rare NW-raking oblique lineations. Some faults dip steeply southwest (e.g.,Figs. 4A and 4B) and have left-lateral separation, opposite that of the main trend of NE-dipping fault surfaces. Similar steep SE-raking oblique slickenline rakes on these faults suggest that they may have originated as conjugate normal faults to NE-dipping fault surfaces. Where visible, the faults that splay from or connect between the Eagle‘s Nest and Maria Wilderness faults have right-lateral separation of map features, and we use this relation to assume slip sense for faults without textural slip-sense indicators.

Some faults have multiple orientations of slickenlines. In most cases, overprinting relations are not readily apparent between slickenline sets, but an exposure of the principal slip surface of a splay to the Maria Wilderness fault (Fig. 4C) has shallowly raking lineations that appear to truncate steeply raking lineations, suggesting that dextral strike-slip followed normal and oblique dextral-normal slip. Subsidiary fault surfaces to the Maria Wilderness fault in the southeastern Big Maria Mountains have largely down-dip slickenlines, while exposures of the principal slip surface have subhorizontal slickenlines with dextral slip-sense indicators (Fig. 4B). T-axesvary from NE-SW-trending with gentle plunges for normal faults to subhorizontal and E-W-trending for faults with dominantly dextral slickenlines. P-axes are more variable and range from subhorizontal N-S to steeply S- or SW-plunging. Together these axes fit an overall dextral transtensional kinematic regime with NE-SW to E-W T-axes and P-axes varying from N-S subhorizontal to steeply plunging. If normal and slightly oblique normal-dextral kinematics predate dextral strike-slip kinematics, as suggested by slickenline overprinting relations, a kinematic shift would be required on these faults in the Big Maria Mountains, which progressed from NE-SWextension to NW-directed dextral transtension, the latter characterized by E-W extension and vertical to N-S subhorizontal shortening.

Estimates of fault-slip offset in the Big Maria Mountains are confounded by the lack of suitable piercing points and previously unknown fault kinematics. Cenozoic rocks are sparse in the Big Maria Mountains and do not offer suitable linear offset markers. Several lines of evidence from geologic mapping suggest a significant component of NE-down vertical displacement on the Maria Wilderness fault. For example, Jurassic metavolcanic rocks are in the core of the Big Maria syncline northeast of the fault trace, but where this syncline occurs on the southwest side of the fault, lower Paleozoic rocks are exposed in the syncline core (Fig. 4). This suggests that the exposure level is deeper on the southwest side, likely due to a component of NE-down motion across the fault. No such relations requiring dip-slip displacement are apparent for the Eagle‘s Nest fault, where Mesozoic structural features have comparable exposure levels on either side of this fault.

Although the exact magnitudes of the strike-slip and dip-slip components of faulting are unknown, they can be constrained within a range of possible values using the orientations of planar offset markers. Salem (2009) mapped several generations of folds (F1–F3) and the associated axial cleavages (S1–S3) and used these to estimate separation. He describes an estimated 1.5 km of right-lateral separation of the Eagle‘s Nest fault by aligning traces of Late Cretaceous F3 folds and 104 m of NE-down separation of the Eagle‘s Nest fault on a cross section that reconstructs the largest of the F2 folds, the Big Maria syncline. By these measurements, the calculated overall slip vector rake in the fault plane would be 4° SE, significantly shallower than many of the steeply raking slickenline measurements from the principal slip surface collected during our study (Figs. 4E, 5D, and 5E).

Using our average Eagle‘s Nest fault principal slip-plane measurement, we calculated that the 1.5 km of horizontal right-lateral separation of S3 folds described by Salem (2009) can be explained by as little as 400 m ofNE-down pure dip-slip displacement, and the right-lateral separation of the Big Maria syncline axial surface can be explained by 900 m of NE-down oblique dextral-normal displacement in the direction of our 83° SE-raking average slickenline measurement. Likewise, along the Maria Wilderness fault, the 2.3 km of right-lateral separation of the Big Maria syncline (Fig. 4) can be explained by as little as 500 m of NE-down normal-sensedisplacement or 1.2 km of NE-down oblique dextral-normal displacement along the 50° SE-raking average slickenline measurement for that fault.

The range of measured slickenline rakes from NW-striking faults in the Big Maria Mountains suggests a multistage slip history and that the true magnitude of displacement is a composite of dip-slip, oblique, and strike-slipdeformation. Given that the Big Maria syncline is significantly refolded around NW-tracing folds (F3 of Salem, 2009), the assumption in the above calculations that the syncline axial surface is a suitable planar offset marker is demonstrably not met, but we suggest that between the Eagle‘s Nest and Maria Wilderness faults there could be as much as ~3.8 km of cumulative dextral shear if separation of the Big Maria syncline is the product of dextral strike-slip (Fig. 4). Alternatively, the mapped right-lateral separations could be primarily an artifact of NE-down dip-slip displacement of dipping markers. If the average measured slickenline is taken to represent the true slip vectors, total dextral slip across the Eagle‘s Nest and Maria Wilderness faults in the Big Maria Mountains is ~1.0 km, and from our observations of strike-slip and oblique slickenlines with dextral slip-senseindicators, we suggest this value as a practical minimum estimate of dextral shear. Thus, we suggest that right-lateral separation is not a product of only dip-slip faulting and estimate the cumulative dextral slip shared by these faults to be 1.0–3.8 km (Dataset S4 [^{footnote 1}]; Mavor et al., 2023).

Calcite along the Eagle‘s Nest fault in the northwestern part of the Big Maria Mountains (Fig. 2) is mineralized in a 3–5-cm-thick banded vein on the principal slip surface (Fig. 6A; sample 1912-SM215 in Dataset S3 and Mavor et al., 2023), and slickenlines on the vein margins indicate that veining occurred before fault slip had terminated. Two sets of slickenlines are present on opposing vein margins, raking 77° SE and 62° NW. In a thin section perpendicular to the vein, calcite has syntaxial growth textures with coarse-zoned dogtooth calcite terminating inwards from an outer margin of fine-grained calcite, and we targeted both textural domains for laserablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) analysis (Fig. 6B). The overall regression age for sample 191-SM215 is 7.9 ± 1.0 Ma (MSWD = 1.6, Fig. 6C). Calcite from the finer-grainedtextural domain has lower U concentrations (<6 ppm) and a high uncertainty regressed age of 1.5 ± 7.4 Ma (MSWD = 1.4). The regressed age for analyses from only the coarse-zoned textural domain (n = 125) is the same age and uncertainty as a whole-sample age (with MSWD = 1.7, Dataset S3), indicating that the coarse calcite is Late Miocene and the dominant signal in the whole-sample age.

Richard (1993) used the name ―Maria fault‖ (MF, Fig. 2) to refer to a postulated system of largely concealed NW-striking faults that offset the eastern Little Maria Mountains from the northwestern Big Maria Mountains (MF, Fig. 2). Some studies have suggested that a syncline of metamorphosed Paleozoic rocks in the southern Little Maria Mountains is a continuation of the Big Maria syncline and is displaced by a concealed NW-strikingfault between the ranges (Ellis, 1981; Richard, 1993). However, no faults are depicted in this saddle on geologic maps by Hamilton (1984), Ballard (1990), or Stone et al. (2022), and regional structural depictions by Stone (2006) and Salem (2009) show the concealed trace of the Big Maria syncline as an unfaulted feature between the Big Maria and Little Maria mountains. Citing telluric traverses (Hoover et al., 1981) and unpublished gravity traverses by R.W. Simpson Jr., Hamilton (1984) infers a concealed NW-striking, SW-down fault on the west side of the Big Maria Mountains that traces toward but does not reach the saddle between the Big Maria and Little Maria Mountains near the abandoned town of Midland, California (locality MF on Fig. 2). A NW-trending alignment of low-lying hills composed of Miocene–Oligocene boulder breccia (Stone et al., 2022) lies southwest and subparallel to this geophysically defined NW-striking fault. Gravity maps of southeastern California and western Arizona (Rotstein et al., 1976; Mariano et al., 1986; Carr, 1991) depict a strong NW-to WNW-striking gravity gradient southwest of these hills, which forms the northeast margin of the Blythe gravity low and continues westward south of the Little Maria Mountains (Fig. 2). Stone (2006) suggested that the topographic lineament on the northeast side of aligned breccia hills may be controlled by a structure that splays from the geophysically defined basin-bounding fault.

The eastern contact of the breccia with crystalline bedrock is marked by a subtle NNW-tracing lineament visible in lidar and aerial imagery (MF, Fig. 2); fault surfaces and cataclastic zones in the pre-Cenozoic bedrock along this lineament dip moderately west, slightly oblique to the lineament trace. Slickenlines are largely down-dip, though one observed set rakes shallowly from the south-southeast (Fig. 7A). Riedel shears on fault surfaces with dip-slip slickenlines suggest W-down normal-sense slip. Incremental strain axes for these fault measurements suggest E-W subhorizontal extension and vertical shortening. Where exposed, the Cenozoic breccia is cut by NW- and NNW-striking, steeply dipping faults marked by several centimeters of gouge, indicating that at least some deformation postdates deposition of the breccia.

Faults that displace Mesozoic structures are mapped in the bedrock in the eastern Little Maria Mountains (locality MD on Fig. 2), with both right and left map separation of Paleozoic units in the overturned northern limb of the Big Maria syncline (Emerson, 1981; Ballard, 1990). We measured fault surfaces along these mapped faults and found a range of orientations between steeply W-dipping, N-striking and subvertical, or steeply NE-dipping surfaces, all with steeply NNW-plunging slickenlines (Fig. 7B).

From our observations of the topographic lineament, faults that cut the pre-Cenozoic bedrock, and faults that cut the Miocene–Oligocene breccia, we infer that the low pass between the Big Maria and Little Maria Mountains is a structural depression that may be controlled by a NW- or N-striking fault, similar to the structure envisioned by Richard (1993). We propose that this fault continues to the southeast following the geophysically defined trace depicted by Hoover et al. (1981) and Hamilton (1984) and the topographic lineament on the northeast side of the Miocene–Oligocene breccia outcrops. The available bedrock exposure allows but does not require significant right-lateral separation along this fault. Regardless, the 4.5 km (Richard, 1993) or 5 km (Richard and Dokka, 1992) estimates of right-lateral separation across the Maria fault are likely near the upper end of permissible dextral slip magnitudes that align map units in the north limb of the Big Maria syncline between the two mountain ranges. Fault slip may be entirely W-down normal sense, as suggested by our brittle fault kinematic data (Fig. 7A). If this is the case, the magnitude of normal slip is probably small as this results in apparent left-lateral separation of the N-dipping Paleozoic map units in the overturned northern limb of the Big Maria syncline. In summary, we concur with Stone (2006) that the Maria fault is likely of lesser importance than faults with stronger horizontal gravity gradients to the southwest and consider the range of total dextral slip on the Maria fault to be 0–5.0 km (Dataset S4 [^{footnote 1}]; Mavor et al., 2023).

The Blythe graben (locality BG on Fig. 2) is the only evidence in the study area for surface fault rupture in Quaternary time. Faults that bound the up to 230-m-wide graben are generally NW-striking. A paleoseismic trench study of the Blythe graben (Fugro Inc., 1975) reported fault and fracture surfaces that cut and displace older alluvial deposits but are overlapped by the youngest alluvium. The largest displacement on any individual fault is 1.5 m of SW-down offset measured at the base of the youngest alluvial fan unit along a fault oriented 116°, 77° SW on the northeast side of the graben (Fig. 7C). Fault and fracture surfaces observed in these trenches strike northwest, northeast, or east. We suggest that the smaller-displacement surfaces that are oriented different than the ~280–300° surface trace of the graben likely do not reflect the trace of the larger structure. No slickenlines are reported from the trench study, but oblique dextral-normal slip would be compatible with regional Pliocene–Pleistocene E-W extension and the modern geodetic strain field (Thacker et al., 2020).

The location of the Blythe graben corresponds with the trace of the gravity lineament depicted by Rotstein et al. (1976) that bounds the northeast edge of the Blythe gravity low. From this spatial correspondence, and the paleoseismic observation that the largest displacement fault is along the northeastern margin of the graben, we interpret that the sense of vertical displacement on the structure responsible for the Blythe graben is dominantlySW-down and that the NE-down faults along the southwestern margin of the graben are likely either antithetic to the SW-down faults along the northeastern graben margin or are related to complex surface rupture of an oblique dextral-normal fault (e.g., negative flower structure). The Blythe graben surface lineament appears to bend toward the west at its northern end, and we infer that the underlying structure does not project toward the Maria fault but instead follows the gravity gradient mapped by Rotstein et al. (1976) southwest of the Little Maria Mountains.

Two NW-striking right-separation faults transect the northern Little Maria Mountains, each tracing for 5–8 km within linear valleys. We follow the nomenclature of Emerson (1981) and refer to the western fault as the ―Valley fault‖ and the eastern as the ―Border fault‖ (VF and BF, respectively, Fig. 2). Mapping by Emerson (1981) shows the southeast termination of the Valley fault as a horsetail splay of NNW- and N-striking faults. Both faults displace Paleozoic and Mesozoic rocks with Late Cretaceous metamorphic fabrics and do not deform Quaternary alluvial fan deposits. Using the attenuated Paleozoic stratigraphic section in the northern Little Maria Mountains as a marker, mapped horizontal right-separation is ~150 m and ~200 m for the Valley fault and Border fault, respectively (Emerson, 1981), though Stone (2006) points out that right-lateral separation could be a result of either dextral strike-slip or E-down normal faulting.

The steeply NE-dipping principal slip surface of the Valley fault is exposed in several places where eroded gullies expose the fault core (Fig. 8B). Slickenlines are subhorizontal to shallowly NW-raking, indicating dextral and oblique dextral-reverse deformation with an average slickenline orientation of 324°/12° (Fig. 8A). Unlike faults in the Big Maria Mountains, no dip-slip slickenlines are recognized on these NW-striking faults in the Little Maria Mountains, which supports the inference that right-lateral separation of Paleozoic units in the Little Maria Mountains is the result of strike-slip rather than dip-slip faulting (Emerson, 1981). Subsidiary fault surfaces that merge with the principal slip surface are consistent with a dextral Riedel shear orientation (Fig. 8C).

Faults along the mapped central splay fault in the southern horsetail of the Valley fault are mineralized with calcite. Slickenfiber steps and calcite veins filling T-fractures oriented ~30° to the main fault surface indicate syn-kinematic calcite precipitation during dextral slip (Figs. 6D–6G). Two samples of calcite were collected ~30 m apart from exposures of this fault for U-Pb analysis (Figs. 6D–6I). Sample 1912-SM183 (Dataset S3 [^{footnote 1}]; Mavor et al., 2023) is cut perpendicular to the fault surface and parallel to slickenlines on dextral oblique slickenfiber steps. An ~5-mm-thick zone of fine- to medium-grained calcite coats the fault surface. Discrete shear fractures within this sample trace parallel to the exposed surface and are associated with cataclasis. The presence of calcite slickenfibers, calcite along shear- and T-fractures, and involvement of calcite in brecciation indicates that calcite mineralization was synchronous with faulting. Laser ablation spots were targeted along both the fine- to medium-grained calcite as well as the mineralized shear fractures. A regressed age for the whole sample of 6.8 ± 2.9 Ma (MSWD = 1.5) indicates faulting in the Late Miocene or Early Pliocene.

Sample 1912-SM184 (Figs. 6G–6I), collected from the same fault, has calcite slickenfiber steps on an exposed fault surface. In thin section, calcite is present in three textural domains: slickenfiber steps, en echelon veins, and fine-grained calcite in the sample matrix. Patchy fine-grained calcite in the matrix incorporates some crystalline wall-rock material, likely where calcite cemented and replaced fault gouge and breccia. Sub-mm-thick en echelon T-fracture veins mineralized with colorless calcite cut the fine-grained calcite (Figs. 6G and 6H) and, along with Riedel shears observed in both outcrop and thin section, support interpretations of dextral kinematics.LA-ICP-MS spots targeting the slickenfiber textural domain have low U (<0.5 ppm, n = 58) and do not yield a negatively sloped regression age. Calcite from the crosscutting T-fracture textural domain (n = 72) yielded awell-constrained intercept age of 9.8 ± 0.8 Ma (MSWD = 1.4; Fig. 6I). Analyses from the fine-grained calcite matrix (n = 35) yielded a low-precision age of 9.7 ± 7.9 Ma (MSWD = 0.9), which is compatible with the Late Miocene age from T-fracture calcite.

The Packard Well fault zone is a system of WNW-striking, NNE-dipping faults that crosses Palen Pass, which separates the Palen Mountains to the south from the Granite Mountains to the north (Fig. 2). Richard and Dokka (1992) suggested that the Packard Well fault zone kinematically connects northwestward to concealed faults in Cadiz Valley, and southeastward to faults of the Cibola Pass fault zone. At Palen Pass, the Packard Well fault trends 294°–277°, while the trend of a fault required to pass to the southeast between the Little Maria and McCoy Mountains is ~304°–321°. Richard (1993) recognized this geometry and suggested obliquedextral-reverse kinematics for the Packard Well fault zone with an estimated minimum 2.3 km of right-lateral separation and 0.9 km of reverse slip on this fault system required to restore the southern contact of the Cretaceous Cadiz Valley batholith.

At Palen Pass (Figs. 2 and 9), two NE-dipping faults bound an alluvial fan conglomerate (fanglomerate unit Tf of Stone and Kelly, 1989) that consists primarily of gravelly sand or interbedded sand and gravel with subangular clasts. Crudely defined stratification is visible in most outcrops, though massive intervals and well-sorted sandstone beds with ripple marks are locally present. The northernmost fault is WNW-striking and moderately NNE-dipping with NW-raking slickenlines (Figs. 9 and 10A), indicating that oblique dextral-reverse slip juxtaposed plutonic rocks of the Granite Mountains over the younger conglomerate, corroborating the kinematics that Richard (1993) interpreted for the Packard Well fault zone. The fault is overlain by Quaternary alluvium (unit Qa₂ of Stone and Kelly, 1989; Figs. 9 and 10F). P-axes record N-S shortening (Fig. 10A). Bedding within the conglomerate is folded asymmetrically against the bounding faults and is locally slightly overturned with a SSW-vergent tight anticline between synclines (cross section A–A‘, Fig. 9). Farther east, the conglomerate bedding defines a single syncline between the two primary faults with the steepest bedding dipping SSW (cross section B–B‘, Fig. 9). Poles to bedding (Fig. 10B) define a scattered girdle distribution with a NNE-striking cylindrical best-fit plane and a shallowly NW-plunging fold axis. Based on cross sections (Fig. 9), folding accommodated ~110 m of horizontal shortening, or ~23% shortening between the bounding faults. The axial plane strike is 9° counterclockwise of the measured fault, which is compatible with fold-fault obliquity expected within a dextral restraining bend (e.g., Sylvester, 1988). The 325° slip vector azimuth reported by Richard (1993), measured from the average strike of faults used in regional palinspastic reconstructions, matches the 326° average trend of slickenlines on the principal slip surface of the Packard Well fault, supporting suggestions that the Packard Well fault is linked to regional structures with a localized restraining bend at Palen Pass.

Numerous small faults cut and displace strata within the conglomerate map unit (e.g., Figs. 10C and 10D) with gouge zones up to 3 cm thick. Where visible, many of these faults have separation on the order of a few meters or less, though one fault juxtaposes conglomerate of different bedding orientations and likely has displacement of tens of meters or greater. Orientations of these minor faults are scattered but are dominantly NW-striking and steeply dipping with an average orientation of 133°, 88° SW (Fig. 10C). Most faults within the conglomerate map unit do not have visible slickenlines or a determined slip sense, but of those measured, the average slickenline plunges shallowly northwest with predominantly N-plunging P-axes and shallowly ENE-WSW–plunging T-axes (Fig. 10C).

Sparse outcrops of relatively flat-lying olivine basalt flows are located in the Palen Pass area ~4 km SSE of the Packard Well fault zone (PB, Fig. 2). Stone and Kelly (1989) suggested that these basalts postdate faulting at Palen Pass and associated folding of the conglomerate. However, paleomagnetic data from these outcrops indicate 31.4° ± 11.5° of clockwise vertical-axis rotation (Carter et al., 1987), which Richard (1993) interpreted to reflect dextral deformation of the nearby Packard Well fault zone rather than rotation of the Palen Mountains as a whole. In outcrop, the basalt flows are largely undeformed; however, in the northeast part of the exposure, an incipient shear fracture fabric cuts and displaces basalt flow horizons (Fig. 11A). Shear fractures are steeply dipping and strike either south-southeast or west-southwest as apparent conjugates (Fig. 11B). Where visible, slickenlines have shallow rakes and are consistent with a conjugate strike-slip relation. The overall shortening direction recorded by kinematic data (199°/05°) corroborates the shortening direction indicated by the acute conjugate bisector (203°/01°). These shear fractures indicate that at least some tectonic deformation occurred after emplacement of the basalt flows, most likely recording the later stages of dextral slip on the Packard Well fault zone.

A slightly reworked ~75-cm-thick ash bed is interbedded in steeply S-dipping Palen Pass conglomerate (Fig. 10D). Sanidine separates dated by the ⁴⁰Ar/³⁹Ar method yielded ages from 260 Ma to 10.9 Ma (35 grains; Dataset S2; Mavor et al., 2023), with a distinct young population of Middle to Late Miocene grains. A weighted-mean age of the youngest 14 grains is 11.71 ± 0.13 Ma (2σ), and the MSWD of 1.4 is below the upper 95% confidence value of 1.78 (Fig. 10E). The abundance of young grains and sample texture inconsistent with extensive reworking suggests that there was not a prolonged hiatus between eruption and final deposition of the ash bed, and that the 11.7 Ma weighted-mean age is a good approximation for the depositional age of the reworked ash bed.

Conflicting ages have been reported for the basalt southwest of Palen Pass. Stone and Kelly (1989) presented whole-rock K-Ar ages of three samples—two from basalt flows and one from a dike of similar lithology; all three samples have overlapping ages between 6.4 Ma and 6.2 Ma. However, a 13.7 ± 1.6 Ma K-Ar date (Carter et al., 1987) suggests that the flow could be significantly older. Given this uncertainty in the possible age, we dated a basalt groundmass sample with ⁴⁰Ar/³⁹Ar methods (Dataset S2; Mavor et al., 2023). The resulting data display a complicated argon release pattern with an age spectrum characterized by decreasing apparent age with increasing temperature. These data do not conform to a plateau by the criteria of Schaen et al. (2021) but are characteristic of argon recoil (e.g., Fleck et al., 2014) commonly observed in fine-grained volcanic rocks. Calculating a recoil model age yields a date of 7.03 ± 0.23 Ma (2σ); a high MSWD value of 63.02 reflects the sloping distribution of the data. Fleck et al. (2014) demonstrated that even in the absence of a plateau, recoil model ages that overlapped within 2σ uncertainty with isochron ages determined in their study were consistent with external constraints. The 6.7 ± 0.59 Ma (MSWD = 52.86) isochron age for our basalt sample overlaps with the recoil model age and lends confidence that the recoil model age reflects the true eruption age of the sample. If the minor faults and paleomagnetic rotation reflect Packard Well fault zone deformation imparted on the basalt, this date indicates that at least some deformation on the Packard Well fault zone continued after ca. 7 Ma.

Other WNW- and W-striking faults cut Paleozoic and Mesozoic bedrock at Palen Pass and may have been involved in Packard Well fault zone deformation. The N-dipping, S-vergent frontal thrust of the Maria fold-and-thrust belt traces between the McCoy and Little Maria Mountains and through Palen Pass (Salem, 2009). N-dipping overturned strata and reverse faults with Late Cretaceous slip juxtapose Paleozoic and Mesozoic map units. Some of these faults are overprinted by Cenozoic brittle faults with breccia zones (Salem, 2009), while other N-dipping structures lack brittle deformation and instead are marked by thin zones of ductile shear recording Mesozoic strain related to the Maria fold-and-thrust belt. We interpret Neogene faulting on steeply N-dipping structures at Palen Pass to have preferentially localized onto the WNW-striking Mesozoic discontinuity, which introduced a rightward bend in the otherwise NW-striking fault system and imparted the localized transpressional deformation.

The Palen and McCoy Mountains expose a thick package of mostly SSE-dipping Mesozoic volcanic and clastic rocks. In the center of the Palen Mountains, these rocks are cut by several NNW-striking, right-separationfaults with pronounced topographic lineaments along fault traces (PA, Fig. 2; Stone and Pelka, 1989; Fackler-Adams et al., 1997). Extracted planar feature orientations from digitized fault traces on a 1 m digital elevation model are NNW-striking and steeply NE-dipping to vertical. An outcrop of the principal slip surface of a NNW-striking mapped fault has subhorizontal slickenlines (black symbols in Fig. 12A) and oblique gouge foliation indicating dextral strike-slip kinematics.

Slickenline measurements from all fault measurements are near horizontal on average, and faults with shallow slickenlines have dextral kinematic indicators. Textural slip-sense indicators are rare for fault surfaces with dip-slipslickenlines. Because bedding in the central Palen Mountains is SSE-dipping (Stone and Pelka, 1989) and faults dip east-northeast, normal-sense dip-slip would impart apparent left-lateral separation. Thus, dextral strike-slip must be dominant to explain any observed right-lateral separation unless reverse-sense slip is invoked for down-dip lineations, which would be seemingly inconsistent with the known Neogene regional kinematics of NE-SW to E-W extension (e.g., Thacker et al., 2020) rather than NE-SW shortening.

To evaluate the amount of dextral shear accommodated by faults in the Palen Mountains, we measured separation of displaced marker horizons in the McCoy Mountains Formation as mapped by Stone and Pelka (1989) and Fackler-Adams et al. (1997) and found 2.1–3.5 km of cumulative dextral displacement. The marker horizons in the central Palen Mountains are not recognized in the western Palen Mountains (Stone and Pelka, 1989), and thus no displacement magnitudes are estimated for mapped faults in the southwestern Palen Mountains. As such, the true amount of distributed slip may be greater than estimates presented here.

In the McCoy Mountains, NW-striking faults are mapped by Stone and Pelka (1989) as NE- and SW-down normal faults. We observed sparse outcrop exposures of the central fault in the range (MC, Fig. 2), with predominantly dip-slip slickenlines, though shallowly plunging and subhorizontal lineations are also observed on subsidiary fault surfaces (Fig. 12B). In contrast to the relationship mapped by Stone and Pelka (1989), we interpret that the Cenozoic breccia is not offset by the fault in the center of the McCoy Mountains. If true, this unit may postdate SW-down normal slip, but the depositional age of the breccia is not presently known.

The base of the McCoy Mountains Formation does not align along strike between the Palen and McCoy Mountains and is apparently offset with ~4 km of right-lateral separation (Stone and Pelka, 1989). A NNW-strikinggravity gradient on the west side of the McCoy Mountains (Fig. 2) likely marks the location of a concealed structure and has been interpreted as a WSW-down normal fault (Rotstein et al., 1976; Stone, 2006). A concealed fault with dip-slip kinematics would be compatible with the steeply raking lineations on NW-striking normal faults that we observe in the McCoy Mountains. However, to generate 4 km of apparent right-lateralseparation of the 35° SSE-dipping McCoy Mountains Formation, a dip-slip fault perpendicular to bedding and dipping 60° WSW would need 2.8 km of vertical throw. This amount is unlikely for a Miocene fault given the ~0.5–0.7 km depths to pre-Cenozoic bedrock between the ranges estimated by a gravity inversion (Saltus and Jachens, 1995), and thus, we suggest that this fault likely has a significant dextral slip component. A fault with 0.7 km of WSW-down throw would generate only 0.9 km of map right-lateral separation of the McCoy Mountains Formation; therefore, we suggest that a rough minimum dextral slip component is ~3 km.

The Dome Rock Mountains expose SE-dipping Mesozoic clastic and volcanic rocks that are cut on the west flank of the range by the ―south Dome Rock Mountains fault‖ (SDRMF), a shallowly W-dipping normal fault with E-tilted Early to Middle Miocene conglomerates and volcanic rocks in its hanging wall (Fig. 2; Spencer et al., 2016; Brickey, 2022). A series of NW-striking faults cut the Dome Rock Mountains and are marked by sharply recessive topographic lineaments. At Ehrenberg Wash (locality EH in Fig. 2), Mesozoic metasedimentary rocks are faulted with ~4 km of right-lateral separation across a system of interconnected faults, with ~3 km of this offset concentrated on the main Ehrenberg Wash fault (Johnson et al., 2021). Northeast of Ehrenberg Wash, an ~10-km-wide zone of NW- to NNW-striking faults has a cumulative ~5 km of left-lateral separation of these same markers, the largest of which is at Copper Bottom Pass (CB, Fig. 2) with 1–2 km left-lateral separation. Southwest of Ehrenberg Wash, in the Trigo Mountains, an ~3-km-wide zone of NW-striking faults of the Cibola Pass fault zone (CP, Fig. 2) displaces the SDRMF with ~7 km cumulative dextral displacement (Richard et al., 1992).

Although no principal slip surface of the Ehrenberg Wash fault has been found during our study or previous geologic mapping, an exposure of predominantly steeply SW-dipping subsidiary fault surfaces and fault gouge along the fault trace may approximate its orientation (Fig. 13A). Slickenlines rake shallowly, with an average orientation of 120°/01°. Along the southernmost splay fault to the Ehrenberg Wash fault system (ES, Fig. 2), brittle fault surfaces strike northwest with subvertical dips, with shallowly NW-raking slickenlines and dextral kinematic indicators (Fig. 13B). Together, the slickenlines of the Ehrenberg Wash fault zone are subhorizontal to shallowly NW-raking, indicating it is a dominantly dextral strike-slip fault.

Fault surfaces of left-separation faults near Copper Bottom Pass are less well exposed; one fault south of Copper Bottom Pass has a small exposure of the principal slip surface oriented 306°, 67° NE, and other mapped faults have measurable subsidiary fault surfaces that trace northwest or north-northeast with steep dips (Fig. 13C). Measured slickenlines on these fault surfaces are variable and are most commonly obliquely SE- or NW-raking. Some strike-slip slickenlines are accompanied by very sparse dip-slip lineations, suggesting that these faults may have a polyphase slip history. For faults without surface exposures, planar best-fit orientations calculated by extracting mapped fault trace profiles from a digital elevation model yielded steep planes dipping >58° to both the northeast and southwest, which could be consistent with the hypothesis that these faults originated as steeply dipping normal faults or as near-vertical strike-slip faults.

Faults of the Cibola Pass fault zone generally do not crop out in the Trigo Mountains. However, reported shallowly plunging slickenlines in fault exposures along strike ~25 km to the southeast in the Middle Mountains (Richard et al., 1992) suggest that the ~7 km of right map separation is a product of dextral strike-slip rather than dip-slip faulting.

Richard et al. (1992) interpreted the Ehrenberg Wash fault to predate slip on the gently W-dipping SDRMF based on deeper exposure levels of Jurassic metavolcanics to the north. However, Brickey (2022) interpreted the opposite and suggested that dextral strike-slip on the Ehrenberg Wash fault postdated normal faulting of the SDRMF and deposition of the tilted conglomerate; by this model, slip on both the SDRMF and Ehrenberg Wash faults would be constrained to postdate the 14.6 ± 0.1 Ma maximum depositional age of the stratigraphically highest tilted tuffaceous sandstone dated in that study. Neither the SDRMF nor the tilted conglomerate in its hanging wall are recognized north of the Ehrenberg Wash fault, as would be expected if the ~4 km of right-lateral separation of the Mesozoic bedrock was due to dextral strike-slip displacement. Right-lateral separation ofSE-dipping Mesozoic markers can be explained by normal-sense displacement on the Ehrenberg Wash fault if the fault dips steeply SW, as some fault surface measurements indicate, and the Copper Bottom fault and Ehrenberg Wash fault zones could be kinematically compatible as oppositely dipping conjugate normal faults during NE-SW extension as suggested by Johnson et al. (2021). Nonetheless, shallowly raking slickenlines with dextral slip sense indicate that the Ehrenberg Wash fault has at least some dextral slip history. Even a small amount of SW-down dip-slip displacement along the Ehrenberg Wash fault would produce a large amount of apparent left-lateral separation of the shallowly dipping SDRMF, such that later dextral strike-slip movement of the Ehrenberg Wash fault would be unlikely to restore the apparent left-lateral separation. This example shows that the apparent deeper exposure levels and missing SDRMF trace north of Ehrenberg Wash do not preclude Cenozoic slip on the Ehrenberg Wash fault zone, but our observations do not confidently prove late Cenozoicstrike-slip kinematics of NW-striking faults in the Dome Rock Mountains. As mapped by Spencer et al. (2016) and Johnson et al. (2021), faults of the Ehrenberg Wash system do not cut overlying Late Miocene to Pliocene alluvial fan conglomerate, indicating that Cenozoic deformation on this structure had ceased by that time. Shallowly raking slickenlines for left-separation faults north of Ehrenberg Wash could also record a minor amount of dextral strike-slip with displacement insufficient to exceed previous left-lateral separation. This demonstrates that the slip history of these faults cannot be explained exclusively by normal-sense slip.

In summary, field relationships, fault geometry, and fault surface kinematic indicators along NW-striking faults in the Dome Rock Mountains suggest at least some history of dextral strike-slip for the Ehrenberg Wash fault and that such dextral strike-slip is perhaps most kinematically compatible with the strain field documented for the Late Miocene in the Lower Colorado River region (e.g., Thacker et al., 2020). The possibility of a polyphase Neogene slip history that combines a phase of NE-SW extensional conjugate normal faulting with a phase of dextral strike-slip is not ruled out by left-lateral separation of offset markers, though the amount of dextral offset is unlikely to exceed several kilometers.

To evaluate patterns of brittle fault kinematics in the region that surrounds Blythe, California, we separated faults by sense of slip using our field observations of slip sense and slickenline rake (Fig. 14). Faults with acute rake (≤45°) and dextral slip-sense indicators (Figs. 14A1 and 14A2) or assumed dextral slip sense (Figs. 14A3 and 14A4) have N-S shallowly plunging to subhorizontal P-axes and E-W shallowly plunging T-axes. Faults with acute slickenline rake >45° and normal slip sense have near-vertical P-axes and NE-SW shallowly plunging T-axes (Figs. 14B1 and 14B4). Faults with either dextral (40%) or normal slip sense (37%) are 77% of all measured faults with slickenlines and slip-sense indicators (43%, 39%, and 82%, respectively, if faults with assumed slip sense are included). Slickenlines of normal faults are, on average, slightly oblique and SE-raking, related to a minor dextral component of oblique slip (Figs. 14B2 and 14B4).

Faults with left-lateral slip sense (Figs. 14C1–14C4) are uncommon and make up only 5% of all measured faults with observed slip sense (and 7% of faults when including those with assumed slip sense). Some of these faults are NW-striking and steeply NE-dipping, and from the similar geometry to the population of dextral faults, it is possible that these faults are either pre-Neogene (most sinistral faults in this orientation are within the Mesozoic McCoy Mountains Formation), or have slip sense misinterpreted, in which case the WNW-ESE P-axes and NNE-SSW T-axes may not accurately depict Neogene strain regime patterns. Conflicting incremental strain axis orientations suggest the latter for at least some faults with assumed sinistral sense (Fig. 14C4), and apparent left-lateral separation along an oblique dextral-normal fault is possible where a planar marker dips more shallowly than the slip vector obliquity. Other sinistral-sense faults are NE-SW striking, near vertical (Fig. 14C3), and have similar orientations to the Slaughter Tree Wash fault and to late Neogene left-lateral faults in the Mesquite and northern Plomosa Mountains (Thacker et al., 2020).

Faults with observed reverse-slip sense are scattered through the map area and are 18% of measured faults with slip-sense indicators (11% when including faults with assumed slip sense). T-axes are steeply plunging and on average near vertical (Figs. 14D2 and 14D4). The shallowly NNE-SSW–plunging P-axes of reverse-sense faults are broadly compatible with N-S shallowly plunging to subhorizontal P-axes for dextral faults (Figs. 14A2 and 14A4) but not with the near-vertical P-axes for normal-sense faults (Figs. 14B2 and 14B4).

Across the entire data set, faults are most commonly NW-striking and steeply dipping: 82% of fault surfaces strike in the northwest or southeast quadrants, 96% dip 45° or greater, and the average fault surface is oriented 318°, 77° NE (Fig. 15A), indicating a predominance of NE-dipping planes. Few NE-striking faults are measured in this data set, and shallowly dipping fault surfaces are near absent. Only six of 642 (0.9%) faults dip <30°, and 491 (76.5%) are 60° or steeper. Slickenlines broadly fit a scattered girdle distribution around the average fault surface; shallowly NE-plunging and horizontal to steeply SW-plunging lineations are rare. Within the range of slickenline orientations, steeply E-plunging lineations are the most common orientation as depicted by stereoplot contours and the overall average lineation of 094°/68° (Fig. 15B).

Incremental strain axes of the full fault data set can help to approximate the overall strain regime during fault slip, with the caveat that many faults in the data set displace Mesozoic and older rocks and may record only pre-Neogene strain. Incremental strain axes in our data set have an overall pattern of NE-SW to E-W shallowly plunging T-axes and N-S shallowly to steeply plunging P-axes, with an oblique normal-dextral transtensional fault-plane solution (Figs. 15D–15F). A P-axis compatibility determination using FaultKin found that 97 of 126 (77%) faults with slip-sense indicators and 315 of 385 (82%) faults including those with slip sense assumed are compatible in the same strain regime, with the caveat that such solutions are non-unique. Incremental strain axes that plot in fields incompatible with the fault-plane solution could be from older episodes of faulting, have misinterpreted slip sense, or could reflect strain partitioning or localized strain fields influenced by bends or interactions between faults.

The bimodal distributions of steeply SE-raking and subhorizontal principal slip-surface slickenlines (Fig. 15C) and of N-S subhorizontal and steeply S-plunging P-axes (Fig. 15F) could suggest two distinct episodes of brittle deformation, one with oblique normal-dextral slip and another with dextral strike-slip, rather than a smooth transition between the styles. However, a histogram of acute slickenline rakes has a near even distribution across the range between dip-slip and strike-slip, with a slight prevalence of steeply oblique and strike-slip lineations (Fig. 15H), and the overall data better support a continuous range of fault styles from normal, oblique dextral-normal transtensional, dextral strike-slip, and oblique dextral-reverse on NW-striking, NE-dipping faults. Fault surfaces with more than one set of slickenlines are common, but crosscutting relations needed to determine the relative timing between slickenline sets are rare. The best examples, encountered in the southeastern Big Maria Mountains (at the location of Fig. 4C), suggest that shallowly SE-raking slickenlines (dextral strike-slip) overprint down-dip slickenlines (normal slip).

Timing constraints presented in this study indicate that NW-striking faults in the Blythe region were active in the Late Miocene and possibly into the Early Pliocene. The 21.7 ± 2.8 Ma dacite that is cut by the Slaughter Tree Wash fault, which is, in turn, cut by the Maria Wilderness fault indicates that both faults are Early Miocene or younger. Two U-Pb dates on synkinematic calcite from the Valley fault in the Little Maria Mountains with shallowly NW-raking slickenlines (Figs. 6F and 6I) can be interpreted in two ways. The uncertainties of the two ages nearly overlap at ca. 9 Ma and could indicate that strike-slip deformation, dextral or oblique dextral-reverse, was active at that time. Alternatively, these two samples could reflect multiple periods of fault activity, one 10.6–9.0 Ma and another 9.7–3.9 Ma.

Kinematic context for the calcite sample from the Eagle‘s Nest fault is less clear, but the U-Pb date indicates that this fault was active between 8.9 Ma and 6.9 Ma (Fig. 6C). The youngest age of the three calcite dates (6.8± 2.9 Ma) has the largest 2σ uncertainty, and deformation in this sample could be as young as ca. 4 Ma.

The Packard Well fault, potentially the largest offset fault in the Blythe region, was active after deposition of the Palen Pass conglomerate at 11.7 Ma, with deformation persisting long enough into the Late Miocene to deform that unit into overturned folds. If the paleomagnetic rotation and weak brittle deformation in the sequence of ca. 7.0 Ma basalt flows near Palen Pass (Fig. 11) do indeed reflect activity of the nearby Packard Well fault, then Neogene brittle deformation at Palen Pass must have continued into the latest Miocene or later.

Together, these timing constraints suggest that NW-striking faults in the region surrounding Blythe, California, were active, at least some with dextral strike-slip or dextral oblique-slip kinematics, by 9.8 ± 0.8 Ma (the oldest date of synkinematic calcite) and at least some brittle deformation continued until after the ca. 7.0 Ma eruption of the Palen Pass basalt flows. Most faults in the Blythe region have no indications of deformation in the Late Pliocene or Quaternary, with exception of the Blythe graben that displaces Pleistocene to Holocene alluvium (Fugro Inc., 1975). Post-Miocene activity has been proposed for faults in this region, with studies suggesting displacement of the 4.5–3.5 Ma Bullhead Alluvium near Blythe (e.g., Howard et al., 2015; Crow et al., 2019; Thacker et al., 2020). Geophysically defined faults at the margins of the Blythe gravity low (Fig. 2) may bound a NW-striking, ~6–9-km-wide graben of displaced Bullhead Alluvium and connect along strike to the Packard Well fault to the northwest. Epicenters of recorded seismicity are sparse in the region surrounding Blythe, but a vague northwest trend aligned near the trace of the Blythe graben may suggest historical activity (Thacker et al., 2017). We speculate that the Blythe graben could be the southeasternmost fault of the ECSZ as defined by Miller (2017) that meets the criterion of recent activity.

Northwest-striking, steeply northeast-dipping faults near Blythe, California, with a range of normal, oblique dextral-normal, dextral, and oblique dextral-reverse kinematics were persistent structural weaknesses that record a transition between extensional and oblique dextral faulting. The predominance of steeply NE-dipping faults, with few E-W– or N-S–striking surfaces, and a near complete absence of shallowly dipping fault surfaces in this data set suggests that NE-dips of dextral faults may be inherited from Early to Middle Miocene normal faults that accommodated NE-SW extension. Multiple orientations of slickenlines on similarly oriented faults indicate that these faults were active across changing kinematic regimes and are consistent with the documented shift in the regional strain regime from Middle Miocene NE-SW extension to Late Miocene NW-directed dextral transtension accompanied by E-W extension and N-S shortening (Singleton, 2015; Singleton et al., 2019; Thacker et al., 2020). Few constraints exist to estimate the onset of faulting in the present study area, but NE-downnormal faulting is kinematically compatible with Early to Middle Miocene NE-SW extension, and shallowly raking slickenlines and E-W–trending T-axes are best compatible with the known Late Miocene and younger transtensional strain regime.

Northeast-directed extensional detachment faulting in the Colorado River extensional corridor initiated at ca. 22–20 Ma, and slip on most detachment faults ceased by ca. 14–12 Ma, as constrained by thermochronometry ~40–120 km to the northeast of the Big Maria Mountains (Foster and John, 1999; Stockli et al., 2006; Singleton et al., 2014; Prior et al., 2016, 2018). Thus, the Early to Middle Miocene NE-SW extensional strain regime predates our Late Miocene calcite U-Pb dates that are synkinematic to oblique dextral faulting. Our data infer a model where normal faulting in an Early to Middle Miocene NE-SW extensional strain regime is overprinted by dextral and oblique dextral slip in a Late Miocene to Pliocene E-W extensional strain regime. Oblique normal-dextral slip may record transitional kinematics between these regimes. Where observed outside of localized fault restraining bends, NW-raking slickenlines with oblique dextral-reverse kinematic indicators could reflect a youngest evolution to dextral transpression, as interpreted by Singleton et al. (2019) for NE-up oblique dextral slip for faults in the Buckskin Mountains that cut the youngest faulted deposits there.

The ENE-striking Slaughter Tree Wash fault, with steeply ENE-raking oblique slickenlines, could be compatible with the Middle Miocene strain regime as a transfer fault between steeply NE-dipping oblique normal-dextralfaults, such as the Maria Wilderness and Quién Sabe or Big Wash faults (Fig. 2). Elsewhere in the Colorado River extensional corridor, ENE-striking faults accommodate discrepancies in slip magnitude or polarity between extensional domains as transfer faults (e.g., Spencer and Reynolds, 1989; Beratan, 1991). Postdetachment sinistral-normal oblique faults with this orientation are present in the northern Plomosa Mountains (Strickland et al., 2017). In the Mesquite Mountains (Fig. 2), a NNE-striking sinistral strike-slip fault was likely coeval with deposition of basal carbonates of the Bouse Formation (Thacker et al., 2020). ENE- to NNE-striking left-separation faults with late Neogene slip are also recognized ~60 km to the south of Blythe, where they are conjugate to NW-striking dextral faults and displace the ca. 9.45 Ma basalts of Black Mountain (Ricketts et al., 2011), and ~180 km to the northwest of Blythe at the oblique sinistral-normal Nipton Mountains fault (Mahan et al., 2009; Miller et al., 2019).

Several lines of evidence indicate that fault reactivation may have played a significant role in localizing deformation across the mountain ranges surrounding Blythe, California. Thickness changes within Jurassic strata acrossNW-striking left-separation faults in the northern Dome Rock Mountains provide evidence for syndepositional slip on these faults. Cretaceous deposits that are displaced but do not change thickness across these faults indicate subsequent reactivation (Crowl, 1979). Restoration of the 069°, 30° SW average bedding of 140 measurements within Jurassic units in the Cunningham Mountain quadrangle (Johnson et al., 2021) to horizontal, would restore the 306°, 67° NE measured principal slip surface of a left-separation fault south of Cunningham Mountain to 300°, 85° NE, an approximation of fault orientation at the time of deposition, and this steep dip suggests that the fault may have originated as a strike-slip fault. While the timing of fault slip for NW-striking faults in the Dome Rock Mountains is not well constrained, these relations demonstrate that these faults may belong-lived structural features.

Neogene deformation on the Packard Well fault zone at Palen Pass most likely reactivated strands of a Mesozoic fault system. NNE-dipping faults that bound the Miocene conglomerate are also aligned with significantNNE-dipping discontinuities in the Mesozoic bedrock units that mark the boundary between the Maria fold-and-thrust belt and the McCoy basin (Stone and Kelly, 1989; Salem, 2009). Our transpressional kinematic data (Fig. 10A) support the suggestion that the E-W curve in the trace of the Packard Well fault zone through Palen Pass imparts a restraining bend in an otherwise dextral system (Richard and Dokka, 1992), and we argue that an inherited Mesozoic zone of weakness guided fault localization to form this bend. Slip magnitude estimates for the Packard Well fault zone of 12–16 km, or as much as 24 km northwest of Palen Pass (Powell, 1981; Richard and Dokka, 1992), likely include pre-Miocene deformation that also displaced Mesozoic markers (Richard, 1993). Salem (2009) hypothesized that the concealed frontal fault between the Little Maria and McCoy Mountains could have been a Jurassic–Early Cretaceous normal fault bounding the McCoy rift basin that was reactivated in the Late Cretaceous as a Maria fold-and-thrust belt accommodation zone for S-directed thrusting. If true, this suggests that faults with Late Miocene slip could in places be reactivating long-lived polyphase structures with previous episodes of Early Miocene, Cretaceous, and Jurassic slip. Examples of such influence of inherited structures are present elsewhere in the (paleo-) ECSZ: the Mule Mountains thrust system is reportedly reactivated as a brittle fault with unspecified kinematics in the southeastern Dome Rock Mountains, southern Palen Mountains, and the southern Coxcomb Mountains (Tosdal, 1990), and dextral restraining bends in the central ECSZ may have localized on Early Miocene normal faults (Singleton and Gans, 2008).

Palinspastic reconstructions and existing estimates of the cumulative magnitude of fault shear in the Blythe region do not include several right-separation faults within the Big Maria, Little Maria, and Palen Mountains. To estimate the possible range of cumulative dextral shear across a 50 km SW-NE transect from the Palen to Riverside Mountains (Fig. 2), we compile published descriptions of fault-slip magnitudes and add magnitudes ofright-lateral separation from this study. Our tabulated estimates indicate a lower bound of 11 km and an upper bound of 38 km of dextral shear between the Palen and Riverside Mountains (Fig. 16; Dataset S4 [^{footnote 1}]; Mavor et al., 2023). These estimates are complicated by the lack of suitable Cenozoic piercing points and largely rely on sub-planar Mesozoic markers. Estimating true Cenozoic displacement is not possible with these markers, and shear estimates are dependent on the true slip vector that we argue changed through the displacement history. Right-lateral separation on some faults in the Big Maria and McCoy Mountains could be purely a product of dip-slip motion that generates apparent map right-lateral separation. However, most faults have slickenlines indicating at least some oblique or strike-slip deformation, and we consider the minimum estimates a likely underestimate. We use a minimum of 1 km of right slip for the Eagle‘s Nest/Maria Wilderness fault system based on oblique slickenlines. Faults within the Palen and Little Maria Mountains have slickenlines that suggest mostly strike-slip deformation, and right separation of SE-dipping markers cannot be explained by normal-sense displacement on NE-dipping faults; therefore, for such faults, we approximate the magnitude of dextral shear using the map strike separation.

The greatest uncertainties with the cumulative dextral shear estimates are from the Packard Well fault and the Maria fault. We use the minimum dextral slip estimate of 2.3 km for the Packard Well fault from Richard (1993) as a lower bound of shear between the northern Palen Mountains and Little Maria Mountains. Upper estimates of Packard Well fault slip of ~12 km between the McCoy and Little Maria Mountains (Richard and Dokka, 1992) or ~16 km to align rocks in the Little Maria Mountains to Palen Pass (Powell, 1981; Richard and Dokka, 1992) may include Mesozoic displacement; the latter estimates are greater in part due to combined displacement through the Ford Lake North fault (3–4 km of right-lateral separation) and the Iron Mountains fault (~5.5 km dextral or 1 km normal slip; Miller and Howard, 1985; not included in Dataset S4) that likely merge with the Packard Well fault east of Palen Pass. Despite uncertainty in how the slip is partitioned between these faults, they likely share a kinematic linkage along strike with the southeastern Bristol–Granite Mountains fault where offset geophysical markers suggest 15–17 km of dextral offset (Langenheim and Miller, 2017); it is unlikely that the cumulative shear values of the Packard Well–Iron Mountains system greatly exceed those values. Even if dextral slip of the Packard Well fault is held to the minimum value of 2.3 km, up to 25 km of cumulative dextral shear is allowable on the recognized structures along our 50 km transect.

These estimates demonstrate that the cumulative dextral shear is possibly greater than the maximum value of 16 km proposed northeast of the Sheep Hole fault (Richard, 1993), and our upper bounding estimate affords another 20 km of dextral shear to that estimated by Bennett et al. (2016) between the Palen and Riverside Mountains (excluding faulting in the Buckskin Mountains). Additional unaccounted dextral shear in our transect may be accommodated by sub–map-scale distributed faulting, or by faults that lie entirely concealed in valleys between mountain ranges. The potential amount of underestimated diffuse shear is difficult to estimate, but kinematic models that reconcile geologic and geodetic slip rates across the modern ECSZ suggest that ~40% of total strain may occur away from mapped faults (Herbert et al., 2014). Future palinspastic reconstructions must be careful to allow for this cumulatively significant shear as intra-block deformation.

Pervasive, but difficult-to-constrain, dextral shear is likely present across much of the lower Colorado River and western Mojave regions. Carr‘s (1991) depiction of offset of the Whipple Mountains detachment fault by geophysically defined faults between the Riverside Mountains and eastern Buckskin Mountains (Fig. 2) could accommodate as much as ~7 km of unaccounted dextral shear, and Richard and Dokka (1992) report an additional ~10 km of dextral shear on minor NW-striking faults in the Coxcomb Mountains. However, it is unclear to what extent these apparent separations could be the result of dip-slip or oblique-slip faulting. The prominent gradients bounding the Blythe gravity low could indicate even more unrecognized dextral shear; Stone et al. (2022) estimated that as much as 50 km of dextral shear is needed to align the base of the McCoy Mountains Formation between the McCoy and Dome Rock Mountains (Fig. 2). Umhoefer et al. (2020) expanded on their then in-progress mapping to suggest that 30–35 km of dextral shear would restore a pull-apartbasin model of the gravity low. However, our kinematic data suggest that the Blythe gravity low formed in part due to the dip-slip component of oblique faulting, and shear estimates made by restoring a geophysical pull-apart basin with only strike-slip shear on NW-striking faults may overestimate the true magnitude of dextral shear. If the potential ~7 km of dextral shear along concealed faults near Parker, Arizona (Carr, 1991), is accounted for, only 23–28 km of dextral shear in the region surrounding Blythe is required to resolve the 30–35 km shear-magnitude discrepancy identified by Bennett et al. (2016) between the Lower Colorado and surrounding regions. This amount is permissible by our estimates of 11–38 km of dextral shear across our 50 km transect.

Although some Miocene dextral shear is possible along NW-striking faults in the Dome Rock Mountains, the permissible amounts are relatively small and cannot match the amounts of dextral shear in the transect to the northwest. In contrast to the hypothesized connections of fault lineaments between the Big Maria and Dome Rock Mountains (Hamilton and Myers, 1966; Powell, 1981), most of the dextral shear along NW-striking faults in the Big Maria Mountains and the faults bounding the Blythe gravity low apparently does not continue along strike into the NW-striking faults in the Dome Rock Mountains, and may instead step south (right step) onto the Cibola fault system, where ~7 km of dextral shear is permissible (Richard et al., 1992); this releasing stepover geometry could allow down-dropping across a NW-SE–oriented pull-apart model for formation of the gravity low.

Sharp NW-SE gravity gradients between the McCoy and Big Maria Mountains are suggestive of concealed faults that bound the Blythe gravity low, and the magnitude of these gradients suggests that dip-slip faults juxtapose low-density basin fill against more dense crystalline bedrock (Rotstein et al., 1976). The kinematic model presented here proposes that Early to Middle Miocene normal movement of these concealed faults (synchronous with detachment faulting of the Colorado River extensional corridor) formed the apparent vertical displacements evident in geophysical data, but also that normal faulting was later overprinted by dextraloblique-slip and strike-slip by the end of the Miocene, which likely involved further tectonic subsidence. Continued basin down-dropping in releasing stepovers between dextral faults into the Pliocene could explain the presence of Bullhead Alluvium deposits below sea level near Blythe described by Howard et al. (2015) and Crow et al. (2019).

The data presented here are consistent with hypotheses of strike-slip pull-apart basin deformation near the time of Colorado River integration (Bennett et al., 2016; Cassidy et al., 2018; Dorsey et al., 2021) but also suggest that the earlier basin development (Middle to Late Miocene) may have been controlled by normal and oblique dextral-normal deformation on the same faults. Our model is consistent with the small-scale N-striking normal faults recognized in the Bouse Formation that record Late Miocene to Pliocene E-W extension (Thacker et al., 2020; Dorsey et al., 2021). Our fault kinematic data set from bedrock mountain ranges does not identify a significant population of faults in this orientation; however, our data do not preclude the existence of such faults in Late Miocene and Pliocene deposits concealed by younger alluvial fans. N- and S-striking normal faults may not be recognized in the bedrock owing to the preexisting anisotropy of NW-striking surfaces that localized strain. NNE-striking normal faults (the Ferguson Lake and Lighthouse Rock faults) appear to control the path of the Colorado River between Cibola and Yuma (Beard et al., 2016), and similar faults are suggested to underlie La Posa Plain west of the northern Plomosa Mountains (Strickland et al., 2018; Thacker et al., 2020); the Castle Dome and Indian Wash basins west of the Castle Dome and Middle Mountains, respectively (Richard et al., 1992; Richard, 1993); and the valley between the Moon and Big Maria Mountains (Bennett et al., 2016). Farther north, N-S elongate extensional basins in Ivanpah and Mohave valleys are suggested to accommodate Miocene strain in the gap between the Stateline and Buckskin dextral systems (Mahan et al., 2009; Bennett et al., 2016; Thacker et al., 2020). If the NW-striking geophysical anomalies that bound the northeastern side of the Blythe gravity low do not connect to Copper Bottom or Ehrenberg Wash faults and instead step south (right step) to NW-striking faults of the Cibola fault system in the northern Trigo Mountains as we propose, a concealed W-dipping normal fault would be expected on the east margin of the basin, west of the SDRMF and similar to hypothesized W-dipping normal faults west of the Dome Rock Mountains as depicted on figure 14 of Dorsey et al. (2021). With timing constraints presented here that NW-striking faults were active by 10–9 Ma and may have remained active until Pliocene time, it seems likely that an interconnected fault system across the California-Arizona border was active both before and during the Early Pliocene integration of the Colorado River system. Future work could better constrain the subsurface architecture of the Blythe area and influence of faulting on the evolution of the Colorado River system.

The kinematic shift to E-W extension and dextral to oblique dextral faulting may have shortly followed the cessation of large-magnitude NE-SW extension in the Colorado River extensional corridor at ca. 14–12 Ma. The timing of faulting and the continuum of kinematic data from normal to oblique normal-dextral to dextral that we have documented in the Blythe region are similar to a progressive transition from NE-SW to E-W extension at the end of detachment faulting documented in the Buckskin–Rawhide Mountains (Singleton, 2015). Our interpretation is of a continuous progression between these strain regimes rather than a punctuated transition. Our observations that the same faults were active in both kinematic regimes and locally reactivated Mesozoic structures suggest that fault localization onto preexisting weaknesses may be an underappreciated factor for initiation and localization of paleo-ECSZ deformation.

Our timing constraints from the paleo-ECSZ are consistent with hypotheses of a distributed zone of dextral shear initiating in Late Miocene time, spanning from the Gulf of California to southeastern Nevada. Dextral shear in the Blythe region recorded by our oldest synkinematic calcite date of 10–9 Ma supports hypotheses of ECSZ initiation by ca. 10 Ma (Dokka and Travis, 1990; Schermer et al., 1996; Nuriel et al., 2019) and demonstrates that such dextral shear strain was more widespread than the extent of the Quaternary-active ECSZ. Transitional fault kinematics in the Blythe region may coincide with widespread observations of a kinematic shift across a broad zone of the Pacific–North America plate margin associated deformation. For example, the onset of deformation in the Blythe region is seemingly concurrent with: (1) the initiation of the southern San Andreas fault system in the northern Salton Trough at ca. 8 Ma (Mason et al., 2017); (2) a shift to rapid transtensional faulting in the northern Gulf of California (e.g., Seiler et al., 2010; Bennett et al., 2017) during development of the transtensional Gulf of California shear zone (Bennett and Oskin, 2014); (3) clockwise rotation of the eastern Transverse ranges (Carter et al., 1987); (4) initiation of NW-striking dextral faults in Death Valley where the southern Death Valley fault zone, with ~40 km dextral slip, active during regional extension between 12 Ma and 6 Ma (Pavlis and Trullenque, 2021); and (5) dextral slip on NW-striking faults in western Nevada that initiated at ca. 10 Ma (Reheis and Sawyer, 1997). These events could all relate to increasing relative plate velocities and a steady clockwise rotation of decreasingly divergent oblique relative vectors between the Pacific and North America plates through the Middle Miocene (DeMets and Merkouriev, 2016) or to eastward strain partitioning due to development of the Big Bend in the San Andreas fault between 12 Ma and 5 Ma (Liu et al., 2010). Our finding that faults were active in the Blythe area with dextral kinematics between 10 Ma and 7 Ma is supported by geodynamic modeling that depicts a transition from extension to transtension between 13 Ma and 9 Ma in this region, driven by plate boundary interactions more so than gravitational potential energy (Bahadori and Holt, 2019).

With exception of the Blythe graben, little evidence suggests that faulting continued into the Late Pliocene surrounding Blythe, and the wide footprint of Late Miocene dextral shear waned in the Pliocene to Quaternary. Strain transfer out of the Blythe region in the Pliocene was likely localized into a narrower San Andreas–ECSZ system and is consistent with a rapid increase of dextral shear by 4.7 Ma on the San Andreas fault (Oskin et al., 2001). Decreasing fault activity in the Blythe region by the Quaternary is also consistent with a westward shift in dextral shear identified in southeastern California (e.g., McQuarrie and Wernicke, 2005), within the ECSZ (Dokka and Travis, 1990), and documented by marine geophysical studies in the northern Gulf of California (Aragón-Arreola and Martín-Barajas, 2007).

Faults surrounding the Blythe region in southeastern California dominantly dip steeply northeast and have multiple orientations of slickenlines that indicate polyphase kinematics ranging from normal-slip in a NE-SWextensional regime to oblique dextral and dextral slip during transtension accompanied by N-S shortening and E-W extension. Calcite U-Pb dates indicate that faults were active with dextral-oblique kinematics between 10Ma and 7 Ma and perhaps as late as ca. 4 Ma, which is further supported by the down-dropping of similarly aged (4.5–3.5 Ma; Howard et al., 2015) Bullhead Alluvium near Blythe. Together with the timing constraints, transitional kinematics are best explained as recording a shift from Early to Middle Miocene NE-SW extension of the Colorado River extensional corridor to Late Miocene dextral transtension related to the incipient development of the ECSZ. The kinematic shift from normal to oblique dextral faulting in the Blythe region followed the end of extensional detachment faulting and appears to coincide with a widespread Late Miocene shift in the style and kinematics of Pacific–North America plate boundary-connected deformation. Faulting in the Blythe region was precursor to, and perhaps contemporaneous with, integration of the Colorado River to the Gulf of California, and this deformation likely influenced the river course and geometry of depocenters. Oblique dextral-reverse slip and overturned folds in a ca. 11.7 Ma conglomerate at Palen Pass record localized transpression of the Packard Well fault in a restraining bend. Collocation of a restraining bend in the Packard Well fault with a major Mesozoic structural boundary suggests that the geometries of Neogene structures are at least locally influenced by inherited anisotropy. Limited faulting affects Quaternary deposits in the Blythe region because most strain waned in the Late Miocene or Pliocene. Subsequent dextral shear localized farther west into a narrower zone of plate boundary deformation that includes the San Andreas fault and Quaternary ECSZ. Our new constraints for the evolution of diffuse transitional strain along Pacific–North America plate boundary may be useful to inform models of diffuse plate boundary strain localization at other plate margins.

Funding for this project was provided by supplementary funding to National Science Foundation EAR award #1822064 through the INTERN program and by U.S. Geological Survey National Cooperative Geologic Mapping Program. We thank the United States Army for access to outcrops on the Yuma Proving Grounds. We thank Megan Flansburg for fruitful discussions and thank Sean Long, Jacob Thacker, the Geosphere editor, the Geosphere themed issue guest editor, and an anonymous reviewer for helpful feedback that improved the manuscript. Data presented in this manuscript are available in Mavor et al. (2023). Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.