Crustal block-controlled contrasts in deformation, uplift, and exhumation in the Santa Cruz Mountains, California, USA, imaged through apatite (U-Th)/He thermochronology and 3-D geological modeling

Abstract

Deformation along strike-slip plate margins often accumulates within structurally partitioned and rheologically heterogeneous crustal blocks within the plate boundary. In these cases, contrasts in the physical properties and state of juxtaposed crustal blocks may play an important role in accommodation of deformation. Near the San Francisco Bay Area, California, USA, the Pacific–North American plate-bounding San Andreas fault bisects the Santa Cruz Mountains (SCM), which host numerous distinct, fault-bounded lithotectonic blocks that surround the San Andreas fault zone. In the SCM, a restraining bend in the San Andreas fault (the SCM bend) caused recent uplift of the mountain range since ca. 4 Ma. To understand how rheologic heterogeneity within a complex fault zone might influence deformation, we quantified plausible bounds on deformation and uplift across two adjacent SCM lithotectonic blocks on the Pacific Plate whose stratigraphic and tectonic histories differ. This was accomplished by combining 31 new apatite (U-Th)/He ages with existing thermochronological datasets to constrain exhumation of these two blocks. Additionally, surface exposures of the latest Miocene to late Pliocene Purisima Formation interpreted in 18 structural cross sections spanning the SCM allowed construction and restoration of Pliocene deformation in a three-dimensional geologic model. We found that rock uplift and deformation concentrated within individual Pacific Plate lithotectonic blocks in the SCM. Since 4 Ma, maximum principal strain computed for the more deformed block adjacent to the fault exceeded that computed for the less deformed block by at least 375%, and cumulative uplift has been more spatially extensive and higher in magnitude. We attribute the difference in uplift and deformation between the two blocks primarily to contrasts in lithotectonic structure, which resulted from diverging geologic histories along the evolving plate boundary.Restraining bends in large, plate-bounding strike-slip faults produce deformation and create topography in the surrounding crust even in the absence of significant fault-normal plate motion. In these settings, local convergence of strike-slip plate motion across the restraining bend produces contraction and uplift (Segall and Pollard, 1980; Bilham and King, 1989; McClay and Bonora, 2001; Cunningham and Mann, 2007; Cooke and Dair, 2011). Over geologic time scales, strike-slip motion along the plate margin accumulates synchronously with bend-induced uplift, such that crust may advect into, through, and out of the zone of bend-induced deformation and uplift as it translates along the plate margin. This general kinematic model for long term uplift surrounding restraining bends in strike-slip faulting regimes has been used to describe the formation of numerous mountain ranges, and deformational patterns therein, from around the world (Anderson, 1990; Cowgill et al., 2004; Cunningham and Mann, 2007; Mann, 2007; Gudmundsdottir et al., 2013; Lease et al., 2021). While illustrative, this kinematic model omits the role that crustal rheology and the dynamic evolution of deformation within the restraining bend play in shaping the landforms and records of deformation that are observed today. To capture these aspects, Baden et al. (2022) coupled a mechanical tectonic model with a numerical geomorphic model, and linked signatures of rock uplift, exhumation, erosion, and relief to restraining bend deformation along the San Andreas fault (SAF) in the Santa Cruz Mountains (SCM; Fig. 1), near the San Francisco Bay Area in California, USA. In that investigation, across-SAF contrasts in elastoplastic strain-hardening crustal rheologies captured the dynamic evolution of bulk crustal deformation surrounding the restraining bend (the SCM bend) as tens of kilometers of strike-slip plate motion accrued over four million years. This rheological formulation was designed to capture the bulk behavior of the crust, however, and drastically simplifies the complex crustal structure in the SCM.While simple, idealized models of deformation around the SCM bend predict the first-order aspects of the topography and structural configuration (Baden et al., 2022), details of how crust deforms in the vicinity of the bend appear strongly influenced by the long-term template created by the protracted geologic history of the plate margin. The composite lithologic, stratigraphic, and structural relationships observed in rocks within the SCM today resulted from the post–18 Ma history of right-lateral shear along the Pacific–North American plate boundary (Stanley, 1984; Graham et al., 1989, 1997; Brabb et al., 1998; McLaughlin and Clark, 2004; Graymer et al., 2006; Sharman et al., 2013; Gooley et al., 2021). Crystallization and depositional ages of the geologic units preserved in the SCM span millions to hundreds of millions of years, and recent deformation and uplift of the SCM developed within this complex geologic framework. Importantly, large regional faults divide the crust into distinct structural sub-blocks and microterranes with unique geologic histories (Figs. 1B and 1C). The boundaries between and geologic attributes of these blocks produce large differences in the magnitude of recent (younger than 4 Ma) deformation across these structural domains (Bürgmann et al., 1994; Page et al., 1998; McLaughlin et al., 2001; McLaughlin and Clark, 2004). For example, southwest of the SAF, numerous studies have noted more extensive folding of the Purisima Formation, a latest Miocene–late Pliocene marine unit largely representative of shallow shelf and upper slope conditions (Powell et al., 2007), between the SAF and the Zayante fault than southwest of the Zayante fault (Bürgmann et al., 1994; Valensise, 1994; McLaughlin et al., 2001). However, given the poor constraints on regional fault activity, and the widespread removal of structural relationships largely associated with range uplift and erosion, the variability in deformation, uplift, and exhumation throughout the range has yet to be quantified and linked to particular lithotectonic domains within the SCM.In this work, we investigated how lithotectonic domains within the SCM partition deformation in the presence of a restraining bend (the SCM bend) by coupling preserved exposures and structural attitudes in the Purisima Formation (regarded as an approximate 4 Ma sea-level datum) with 31 new apatite (U-Th)/He ages (AHe ages) throughout the SCM. These AHe ages serve as approximations for the time at which rock samples cooled to below effective closure temperatures of ~65–70 °C (Zeitler et al., 1987; Wolf et al., 1998; Farley, 2000; Flowers et al., 2009), and by inference reveal rates and patterns of exhumation in the SCM over time scales of millions of years (Ehlers and Farley, 2003). We used the newly collected AHe ages, in conjunction with pre-existing thermochronology (Naeser and Ross, 1976; Bürgmann et al., 1994; Baden et al., 2022), to constrain plausible reconstructions of the 4 Ma Purisima Formation datum in a series of 18 parallel cross sections spanning the SCM southwest of the SAF (Fig. 2; Figs. S2A–S2R1). We then retrodeformed a 3-D geologic model of this datum (based on our cross sections) to quantify plausible magnitudes of deformation and uplift throughout the SCM, and variations in deformation within lithotectonic domains. We found that strain, uplift, and exhumation that accumulated in the SCM since 4 Ma preferentially localized in the La Honda structural block (Fig. 1C). This localization of deformation may be driven by differences in stratigraphic architecture and lithologic composition in the upper crust of the juxtaposed La Honda and Ben Lomond blocks, which resulted from the accumulation of geologic deformation over millions of years. Our results demonstrate that the evolution of deformation and uplift of mountains surrounding complex plate boundaries are substantially impacted by the lithotectonic template produced by the long-term geologic history of the plate margin. As such, structures hosted within certain lithotectonic blocks may be more prone to damage during and in between earthquakes as plate motion accumulates, and so the rheological state and configuration may constitute an important aspect of understanding seismic hazard along these complicated plate boundaries.The SCM have been the focus of geologic investigation for over a century, with preliminary stratigraphic and structural studies spurred by the Mw 7.9 San Francisco Earthquake in 1906 (Lawson and Leuschner, 1906; Branner et al., 1909). Bisected by the plate-bounding SAF, the SCM host a myriad of geologic units and relict deformational features that indirectly record the evolution of the North American continental margin over the past ~150 m.y. (Graham et al., 1997; Brabb et al., 1998; Wentworth et al., 1999; Graymer et al., 2006). In the vicinity of the present-day San Francisco Bay Area, subduction that initiated at ca. 180 Ma (Mulcahy et al., 2018) led to the formation and accretion of Jurassic ophiolite sequences, serpentinite blocks, and mélange of the Franciscan Complex currently preserved northeast of the SAF (Graymer et al., 2006; Fig. 1B). The subsequent initiation of strike-slip motion at 30–25 Ma fragmented and shifted crust within the plate-bounding San Andreas fault zone, which led to the northwestward translation of Cretaceous arc-magmatism-derived Salinian granites and several-km-thick Cenozoic marine deposits along the Pacific–North American plate boundary (observed southwest of the SAF; Fig. 1B). Current contrasts in stratigraphy and geologic structure across the SAF distinguish the geologically complex and extensively thrust-faulted Bay Block northeast of the SAF (Fig. 1C; e.g., McLaughlin and Clark, 2004; Supplemental Text S1.1) from crustal blocks southwest of the SAF that host one of the most well-preserved Cenozoic stratigraphic sections in California.In the SCM, classic studies (Clark and Rietman, 1973; Graham, 1978; Stanley, 1984; Graham et al., 1989; McLaughlin and Clark, 2004) and industrial data compilations (California Division of Oil and Gas, 1982; Brabb et al., 2001) provide a geologic framework for understanding deformation in the SCM along the SAF and within the adjoining plates since ca. 30 Ma. We use these studies to elucidate patterns of recent crustal deformation that accumulated within the structurally complex and geologically heterogeneous crust along the evolving plate boundary. In the Bay Block, structural complexity, discontinuous stratigraphy, and the lack of preserved Pliocene deposits within faulted blocks of the Foothills thrust belt (Fig. 1; Aydin and Page, 1984; McLaughlin and Clark, 2004; Graymer et al., 2015; Langenheim et al., 2015; see Supplemental Text S1.1) obfuscate the distribution and magnitude of uplift and deformation northeast of the SAF since uplift of the SCM. Therefore, we focus our study on the SCM southwest of the SAF, where the moderately deformed yet well-preserved Cenozoic stratigraphic section provides an indirect record of both plate margin evolution and recent (4 Ma or younger) deformation associated with SCM uplift.Southwest of the SAF on the Pacific Plate, Cenozoic marine strata reside atop granitic, gabbroic, and Franciscan basement (Supplemental Texts S1.1 and S1.6) within distinct, faulted crustal blocks that vary in their geologic history. Within this Cenozoic sedimentary succession, the well-documented lithologies, stratigraphic relationships, unit thicknesses and ages, and depositional environments (and changes therein) inferred from these units (Supplemental Texts S1.3–S1.5) record geologic responses to tectonism along the evolving plate margin. Similarly, deformation observed within these sedimentary rocks records the effect of time-integrated crustal motion and strain accrued since their deposition. The deposition of the majority of these Cenozoic units substantially predates the restraining-bend-induced deformation beginning at 4 Ma. However, the largely shallow-marine Purisima Formation, thought to have been widely deposited within the northern Salinian Block in the present-day SCM between 6.9 Ma and 2.8 Ma (Powell et al., 2007), is sufficiently young to record deformation that exclusively coincided with the uplift of the SCM. In the following sections, we first outline the relationship between the crustal blocks and associated bounding faults that compose the SCM southwest of the SAF, and the stratigraphic variation therein. We then discuss the deposition of the Purisima Formation within these blocks, and subsequent deformation that occurred during uplift of the SCM.Southwest of the SAF, large strike-slip faults subdivide the crust into unique geologic and structural blocks (Fig. 1C). The Pilarcitos fault, a former trace of the plate-bounding SAF that was abandoned between 6 Ma and 4 Ma following plate motion reconfiguration (Pollitz, 1986; Harbert and Cox, 1989) bounds the Pilarcitos Block to the southwest. The San Gregorio–Hosgri fault (SGHF), which satellite geodesy suggests currently accommodates ~2.4 mm/yr of slip (d’Alessio et al., 2005) and may influence the long-term uplift rate along the Santa Cruz and San Mateo coasts (Anderson and Menking, 1994), bounds the Pigeon Point Block to the northeast (Supplemental Text S1.6). McLaughlin and Clark (2004) further subdivide the southwestern SCM along the Zayante fault, partitioning the region into the Ben Lomond Block, largely composed of minimally buried, warped crystalline basement with a thin sedimentary cover, and the La Honda Block, which hosts a several km thick, folded, and faulted sedimentary succession. In the southwestern SCM, the strike-slip activity along subsidiary regional faults, like the Zayante fault, is poorly constrained, and the geologic and deformation history associated with individual structural blocks appears to be distinct. However, the Ben Lomond and La Honda blocks comprise the large majority of the land area southwest of the SAF, and the stratigraphic sections and units within these blocks are corelative, despite tectonically imparted differences. In contrast, the geologic histories and units contained within the Pilarcitos and Pigeon Point blocks vary dramatically from those observed in the Ben Lomond and La Honda blocks (Supplemental Text S1.6). Additionally, unconstrained Pliocene–Pleistocene strike-slip motion along the Pilarcitos fault and the SGHF complicate correlations of Purisima outcrops in these blocks with those found in the Ben Lomond and La Honda blocks. As a result, this investigation focuses exclusively on deformation that developed in the Ben Lomond and La Honda blocks.The Ben Lomond Block. Within the Ben Lomond Block, sedimentary marine rocks ranging from Eocene to Pliocene in age locally overlie the Salinian granitic crystalline basement to form a thin, irregularly preserved sedimentary cover that rarely exceeds 1 km in thickness. Paleocene to early Eocene (Locatelli Formation) and Eocene (Butano Sandstone) marine rocks that pre-date the initiation of the San Andreas fault at ca. 28 Ma nonconformably overlie the Salinian granitoids (Supplemental Text S1.3). Oligocene rocks are largely absent, and geologic relationships, detrital zircon U-Pb signatures, and tectonic inferences based on lithologies and stratigraphic relationships in the Oligocene–Miocene section within adjacent sections in the neighboring La Honda Block suggest that the Ben Lomond Block may have been uplifted and exhumed during this time (Clark and Rietman, 1973; Stanley, 1984; Gooley et al., 2021). Miocene (Lompico Sandstone, Monterey Formation, Santa Margarita Sandstone, Santa Cruz Mudstone; Supplemental Text S1.5) and Miocene–Pliocene (Purisima Formation) marine rocks, their inferred depositional environments, and multiple unconformities in the stratigraphic record (Fig. 3; Table A1) indicate that numerous tectonically driven marine transgressive-regressive cycles swept across the Ben Lomond Block since the early Miocene. Where present, Pliocene strata are flat-lying or very gently dipping. Lithified marine rocks whose deposition post-dates the Pliocene deposition of the Purisima Formation are not observed in the SCM. However, subaerial fluvial and eolian Pleistocene deposits (Aromas Sand) outcrop to the south of the SCM bend, and multiple marine terraces ranging in age from 610 ka to 81 ka flank the western, coastal slopes of the Ben Lomond Block (Lajoie et al., 1979; Anderson, 1990; Gudmundsdottir et al., 2013).The La Honda Block. In contrast to the crystalline-basement-dominated Ben Lomond Block with minimal sedimentary cover, the La Honda Block hosts thick sedimentary successions with a composite thickness of up to 14 km (Stanley, 1984; Gooley et al., 2021), reaches an inferred preserved thickness of up to 6 km (Stanley, 1984; McLaughlin and Clark, 2004), and exhibits one of the most extensively preserved, continuous sections of Cenozoic marine rocks in California (Clark, 1981). The stratigraphy of the La Honda Block includes formations and assemblages ranging from Paleocene to Pleistocene in age, and stratigraphic relationships record numerous shifts in tectonic activity and associated transgressive-regressive marine cycles (Fig. 4). Regional structural interpretations suggest that the deepest Paleocene sedimentary units lie in faulted contact with the inferred gabbroic basement (McLaughlin and Clark, 2004). Though the Paleocene units do not crop out at the surface and have not been encountered in the subsurface, observations do not preclude their presence at depth (Stanley, 1984; McLaughlin and Clark, 2004). Eocene to Oligocene deposits characteristic of deep-water environments (Butano Sandstone, San Lorenzo Formation) amount to >3 km of the sedimentary section in some locations, and may correspond to rapid basin subsidence associated with wrench-tectonism caused by oblique convergence along the North American continental margin prior to ca. 42 Ma (Fig. 4B; Stanley, 1984). During the Oligocene, shifts in tectonic activity along the plate margin, potentially associated with the approach and collision of the Pacific-Farallon spreading ridge with the North American continental margin at ca. 30 Ma, restructured the La Honda Block (Clark and Rietman, 1973; Stanley, 1984). At this time, deep-water submarine fan and shallow marine coarse deltaic deposits (i.e., Vaqueros Sandstone and Zayante Sandstone) record a major marine regression, which appears synchronous with the accumulation of motion across the Zayante-Vergeles fault, and with relative uplift of the basin margins along the Ben Lomond massif (Fig. 4C; Stanley, 1984; Supplemental Text S1.3). Between 25 Ma and 20 Ma, interfingering mudstones and sandstones (Lambert Shale) and submarine volcanics (Mindego Basalt) suggest that a period of crustal extension and thinning associated with regional transtension along the SAF system caused rapid subsidence and development of a pull-apart basin and associated volcanism within the La Honda Block (Stanley, 1984; Supplemental Text S1.4). Following this period of 25–20 Ma extensional tectonism, later deformation, uplift, and erosion within the La Honda Block at ca. 18 Ma eroded the upper surfaces of the Oligocene–Miocene units, generating the unconformity observed at the base of lower–middle Miocene deposits (Lompico Sandstone, Monterey Formation). After this time, Miocene and Pliocene units preserved within the La Honda Block mimic those found in the Ben Lomond Block (Figs. 3, 4D, and 4E). However, within the La Honda Block, Miocene–Pliocene deposits (Purisima Formation) exhibit more steeply dipping beds (≤75°) than those observed in the Ben Lomond Block. While lithified marine strata post-dating the Purisima Formation are absent, Pleistocene subaerial deposits crop out south of the SCM bend (Aromas Sand), and along the peaks near the northern terminus of the SCM bend (Santa Clara Formation). The transition from marine to nonmarine deposition broadly records the initiation of uplift in the SCM.Stratigraphy of the Purisima Formation. The Miocene–Pliocene Purisima Formation (6.9–2.8 Ma) consists of sandstones, mudstones, and conglomerates, and is geographically more widely distributed in the SCM than any other Tertiary formation (Stanley, 1984; Powell et al., 2007). The majority of the Purisima Formation rests unconformably or nonconformably on underlying rocks, which range in age from Cretaceous (Ben Lomond granitoids) to Miocene (Santa Cruz Mudstone). However, in select localities adjacent to the La Honda Basin, the contact between the Purisima and the Santa Cruz Mudstone is interpreted as conformable (Cummings et al., 1962). While the majority of the top surface of the Purisima has been eroded or is unconformably overlain by Quaternary marine terrace deposits, Purisima deposits in the southern SCM (near Aptos, California) are conformably overlain by the Aromas Sand (late Pliocene to Pleistocene; Stanley, 1984). Though widely distributed, truncation of the Purisima Formation across regional faults complicates stratigraphic correlation and reconstruction. Powell et al. (2007) divides the Purisima Formation into two distinct, informal sections: the Santa Cruz section (preserved to the southeast of cross section F; Fig. 2) and the San Mateo coast section (preserved to the northwest of cross section I). These two sections vary in their stratigraphic continuity and inferred depositional environment through time. The Santa Cruz section is ~325–550 m thick (Shaw et al., 1994; Powell et al., 2007), and is inferred to have been deposited in inner to mid-shelf water depths of <100 m, with the upper portion of the section representing foreshore, shoreface, and potentially non-marine deposition. Age-distinctive diatoms and magneto-stratigraphic constraints together suggest that the Santa Cruz section was deposited between 6.77 Ma and 2.58 Ma, with a distinct hiatus from 4.9 Ma to 3.65 Ma (Dumont and Madrid, 1987; Powell et al., 2007). The San Mateo coast section is ~1750 m thick, and represents the Purisima Formation type section as described by Touring (1959). This section, which was deposited between ca. 6.8 Ma and ca. 2.8 Ma, is separated into the informal Tahana, Pomponio, San Gregorio, Lobitos, and Tunitas members (Powell et al., 2007). The Tahana member (ca. 6.8–5.3 Ma) and lower portion of the Pomponio member (ca. 5.3–3.7 Ma) were likely deposited in shallow shelf water depths (<100 m). Over the course of Pomponio deposition, plausible depositional water depths increased substantially, ranging from 150 m to 750 m in the upper Pomponio section. Water depths once again shallowed to <100 m during the deposition of the San Gregorio member (ca. 3.7–3.4 Ma) and the Lobitos member (ca. 3.4–3.1 Ma), and further shallowed to <50 m during the deposition of the Tunitas member (ca. 3.1–2.8 Ma).While the long-lived geologic structures that detail this protracted history (i.e., regional faults, folds, and stratigraphic relationships) are persistent in the geologic record, uplift that formed the Santa Cruz Mountains themselves is fairly recent (younger than 4 Ma), and likely initiated when plate motion reorganization from 6 Ma to 4 Ma formed a steep bend in the SAF within the incipient SCM. This restraining bend in the SAF (the SCM bend) caused regional contraction and uplift within the bend while strike-slip motion along the plate margin accumulated (Anderson, 1990, 1994; Baden et al., 2022). Researchers attribute uplift observed over a suite of disparate time scales to SCM-bend-induced uplift, including uplifted and preserved Pliocene marine strata (Valensise, 1994), uplift and exhumation inferred from low temperature thermochronology (Baden et al., 2022), uplift recorded in marine terraces along the Santa Cruz coast (Anderson et al., 1990; Anderson and Menking, 1994), increased erosion rates surrounding the inferred locus of uplift in the SCM (Gudmundsdottir et al., 2013), and coseismic uplift during the Mw 6.9 Loma Prieta earthquake in 1989 (Valensise and Ward, 1991). These geologic, thermochronologic, topographic, and geomorphic data broadly constrain the recent evolution of the SCM, and numerical model predictions that bridge the disparate time scales over which these observations apply are consistent with SCM-bend-induced uplift (Baden et al., 2022). In this conception, minimal translation of Bay Block crust relative to the SCM Bend caused uplift and exhumation to localize near Mount Loma Prieta early in SCM formation, which ultimately led to bulk crustal strain-hardening northeast of the SAF. In contrast, the advection of relatively weaker Pacific Plate crust into and through the SCM Bend caused uplift to localize in the southern SCM southwest of the SAF (Baden et al., 2022). Today, the SCM bend accommodates a gradual 12.5° change in strike of the SAF over ~50 km (Fig. 1A), and bend-induced deformation continues to pose a significant seismic hazard to the San Francisco Bay Area while plate motion across the SAF accumulates.In the SCM, high- and low-temperature thermochronology helps to constrain time-temperature cooling histories and inferred exhumation and uplift during the evolution of the plate margin. While high temperature systems (40Ar/39Ar in K-feldspar; Z. Shulaker, 2021, personal commun.); sphene fission tracks, Naeser and Ross, 1976) quantify the long-term unroofing histories of Salinian granitic rocks, lower temperature systems and the cooling histories that they record can resolve recent uplift and exhumation on the scale of 2–3 km. Investigators have utilized low temperature apatite fission track and (U-Th)/He thermochronometers to quantify cooling and inferred uplift and exhumation in the range. Bürgmann et al. (1994) showed that units in the vicinity of Mount Loma Prieta have been more deeply exhumed than the crystalline basement of the Ben Lomond Block within the Pacific Plate. Apatite fission track data collected from granitic bedrock on Ben Lomond Mountain and Montara Mountain (Naeser and Ross, 1976) also suggest that recent exhumation within the Ben Lomond Block has been minimal. However, Baden et al. (2022) showed that, within the La Honda Block, apatite (U-Th)/He ages have been recently reset as a result of uplift within the SCM bend, and these ages increase in the northwest direction.We collected 31 bedrock samples from throughout the SCM, and extracted apatite crystals at Stanford University (California). Twenty-three of these samples were analyzed at the Berkeley Geochronology Center from 2015 to 2018 following procedures described in Tremblay et al. (2015). The remaining six samples were analyzed at the University of California, Santa Cruz, in 2015 following the bulk sample gas extraction and measurement procedures described in Johnstone et al. (2013). We included an additional six samples that were measured and reported in Baden et al. (2022). For all crystals, we applied an α-ejection correction (Ft) after Ketcham et al. (2011) to all measured single-crystal apatite (U-Th)/He ages (hereafter referred to as AHe ages) assuming no U-Th zonation. After data acquisition, we used the Robust Z-score Method (Iglewicz and Hoaglin, 1993) to identify individual Ft-corrected crystal cooling age outliers. The Robust Z-Score Method calculates univariate dispersion of individual values from the population median. Modified Z-Scores are calculated as follows:where MAD = median , and is the sample median. In this approach, samples whose modified Z-Scores exceed a set threshold, D, are labeled as outliers and excluded from analysis (~3% of grains). Modified Z-Scores attained using this method are more resistant to the influence of sample outliers than standard Z-Scores that are calculated using the sample mean. For small sample sizes (n = 10), Iglewicz and Hoaglin (1993) show that 5% of simulated, pseudo-normally distributed populations are excluded when D = 4.19, and 2.5% of these populations are excluded when D = 5.08. We assume that AHe crystal ages in sedimentary rocks may exhibit complex distributions due to variable apatite provenance and complicated thermal histories related to sedimentary deposition, burial, and subsequent exhumation. In this context, we sought to exclude crystal ages that strongly deviated from other observations. Therefore, we set D equal to 5, such that individual crystal age measurements with | Mi | > 5 were considered outliers and excluded from further analysis. We then used crystal cooling age uncertainties to calculate weighted mean ages for each sample. We report the significance of crystal rejection and weighted mean sample calculations in the Supplemental Material. We provide laboratory measurements in Table S1 and sample locations in Table S4.To quantify plausible time-temperature histories for each AHe sample, we used the QTQt kinetics model (Gallagher, 2012), which uses a Markov Chain Monte-Carlo (MCMC) sampler to calculate the time-temperature history of the sample that fits He concentrations and/or fission track densities to within their uncertainties. QTQt also accounts for the role of radiation damage in apatite crystals that develops as a by-product of radioactive decay and He production and impacts the temperature-dependent diffusion of He in apatite (Shuster et al., 2006; Flowers et al., 2009; Shuster and Farley, 2009; Willett et al., 2017). In the SCM, the complex geologic histories of the Ben Lomond and La Honda blocks likely caused multiple episodes of exhumation-driven cooling, burial-driven reheating, and potentially prolonged sample residence in the He partial retention zone (HePRZ; Shuster et al., 2006). Heating, cooling, and extended HePRZ residence times serve to obfuscate the role of radiation-damage-induced He retention in the rock sample, complicating the calculated, thermally driven diffusion history of He. To account for this effect, we modeled the development of radiation damage in all samples using the Alpha Damage and Annealing Model (Willett et al., 2017) within the QTQt program, which tracks the accumulation and annealing of radiation damage that develops from α-decay over the duration of the modeled thermal history.To model thermal histories for our samples, we first compiled and implemented a set of thermal model constraints applicable to each individual sample. These constraints included time and temperature estimates derived from the crystallization of the sampled crystals/rock, depositional constraints based on sedimentary relationships and maximum depositional ages as reported in Gooley et al. (2021), and observable geologic relationships in the SCM (i.e., timing and estimated proximity to the surface at the time of subsequent unit deposition atop the sampled unit). These constraints are reported in Table S2. All samples were modeled with an uninformative prior distribution spanning 80–0 Ma and 15–500 °C, and a specified present-day surface temperature of 15 °C. For each sample, we input individual crystal ages, dimensions, and the measured concentrations of U, Th, Sm (where available), and He. Since the reported analytical measurement error associated with individual crystal ages may not fully capture the uncertainty in interpreting the reported cooling ages, we allowed QTQt to scale the reported errors for individual crystal concentration measurements to bound the permissible set of time-temperature paths more conservatively.To begin, each inverse model was run until the corresponding likelihood chain for the model run converged. This “burn-in” phase typically consisted of ≥100,000 iterations, after which we ran 100,000 “post-burn-in” iterations, and used the time-temperature paths associated with these iterations to calculate 95% credible intervals of the joint posterior time-temperature probability density function (pdf) and the posterior-probability-weighted average time-temperature path, or the “expected” time-temperature path, for the resultant pdf following the procedures and calculations outlined by Gallagher (2012). We then converted the 100,000 “post-burn-in” time-temperature paths and credible intervals to time-depth paths assuming constant 25 °C/km and 35 °C/km geothermal gradients, as is consistent with the range of geothermal gradients previously assumed for the SCM (Lachenbruch and Sass, 1980; Bürgmann et al., 1994; Baden et al., 2022), while requiring surface temperatures to be 15 °C for all paths. We used the converted 95% credible time-depth intervals to define the maximum and minimum permissible depths at 4 Ma (the time at which we assume that the range began to uplift), which, in turn, constrained the exhumation experienced by each thermochronologic sample since that time. We used the expected model depth values for the 25 °C/km and 35 °C/km geothermal gradients to bound the range of expected model depths for each sample. We report credible intervals and expected model values for time-temperature and time-depth model predictions in Table S3. We include MCMC time-temperature path density plots, model constraints, and model prediction results for all samples in Figures S1i–S1xxxvii. Model input and run files are included in the Supplemental Material.To create a 3-D geologic and structural model of the Purisima Formation, we first constructed 18 geologic structural cross sections spanning the SCM southwest of the SAF (Figs. S2A–S2R). These cross sections are spaced ~5 km apart, are oriented roughly parallel with one another, and are perpendicular to the trace of the SAF (Fig. 2). To produce these cross sections, we first compiled regional geologic maps (Graham et al., 1997; Brabb et al., 1998; Wentworth et al., 1999; Wagner et al., 2002), subsurface well data (California Division of Oil and Gas, 1982; Stanley, 1984; Brabb et al., 2001), and pre-existing regional geologic and geophysical cross sections and seismic velocity profiles for reference (Cummings et al., 1962; Clark, 1981; Shaw et al., 1994; Page et al., 1998; McLaughlin et al., 2001, 2002; Jachens et al., 2004; Catchings et al., 2004; Zhang et al., 2018; Fuis et al., 2022). We then chose our cross-section locations, and, where applicable, interpolated geologic structural attitudes from strike and dip measurements on or adjacent to the cross-section trace. Similarly, we projected subsurface contact information derived from compiled regional well data into relevant cross sections, and generated subsurface interpretations of the existing geologic configuration in the SCM that were consistent with the available data. We do not construct our cross sections using typical balancing techniques tied directly to a thermokinematic model (e.g., McQuarrie and Ehlers, 2015) given the geologic uncertainties associated with out-of-section-plane (i.e., strike-slip) faulting, lateral variations in depositional extents and thicknesses of units in the SCM, and multiple unconformities in the SCM stratigraphy (Fig. 3; Table A1).After constructing the structural cross sections, we used current elevations of the preserved Purisima Formation, in conjunction with the predicted exhumational magnitudes derived from the inverse thermal models (described above), to bound plausible uplift and exhumation extents throughout the SCM (Fig. 5). We first assumed that the upper depositional extent of the variably preserved Purisima Formation provides an approximate sea level datum at ca. 4 Ma. We then constructed maximum, minimum, and preferred models for the reconstructed position of this sea level datum in each structural cross section to bound the extent of deformation and associated uplift that has developed since 4 Ma. These reconstructions allow the thickness of the Purisima Formation to vary regionally between ~325 m and 550 m (Shaw et al., 1994; Powell et al., 2007) in the southern SCM in the vicinity of cross section A–A′ to >1700 m thick in the La Honda Basin (California Division of Oil and Gas, 1982; Horn, 1983; Hector, 1986; Wright, 1990; Brabb et al., 2001) near section M–M′. We assume that the thick Purisima section in the La Honda Basin is not representative of its typical thickness, since subsurface wells and cross sections suggest that it has been locally thickened as a result of syndepositional extension in the basin north of the SCM Bend (e.g., cross section M). Plausible Purisima depositional extents and thicknesses are geologically unconstrained in the central SCM near section H–H′, where the Purisima Formation is absent. Elevated portions of the landscape (i.e., Ben Lomond Mountain, Butano Ridge, Montara Mountain, and along the SAF) also lack preserved Purisima remnants (Graham et al., 1997; Brabb et al., 1998; Wentworth et al., 1999; Graymer et al., 2006; Powell et al., 2007). In these areas, we used the modeled exhumation since 4 Ma implied by each AHe sample location to bracket plausible depositional extents and thicknesses for the Purisima Formation at 4 Ma. We projected these bounds above interpolated sample positions in applicable cross sections to predict the projected location of the permissible maximum, minimum, and preferred upper boundaries of the Purisima Formation. Since young (younger than 4 Ma) AHe ages may poorly constrain maximum temperature bounds at 4 Ma given the broad range in the permissible prior temperature (15–500 °C) at the start of the simulation, we limited the position of the maximum Purisima Formation reconstruction to 5 km in elevation. We then used structural attitudes within the Purisima Formation (where present) to project the horizon from preserved outcrops to these interpolated exhumational bounds to form maximum, minimum, and preferred top-of-the-Purisima horizon reconstructions in each cross section (Fig. 5). Where applicable, we also used geologic structural constraints, neighboring AHe sample results, and neighboring cross-section interpretations to inform these reconstructions. Given that the structure of folding and/or faulting within the now eroded Purisima Formation adjacent to the SAF is largely unconstrained in cross sections C–R, we project reconstructions horizontally northeastward from thermal sample locations closest to the SAF in each cross section to minimize construction bias.We combined geologic cross sections (A–R; Fig. S2) to produce 3-D models for each of the deformed Purisima Formation reconstructions spanning the length of the SCM southwest of the SAF. These 3-D models were created using PetEx’s Move3D geologic modeling software, in which stratigraphic horizons and fault lines were digitized for each cross section. We then used a linear interpolant to define the shape and extent of the maximum, minimum, and preferred reconstructions for the top of the Purisima Formation between cross sections. We adjusted the 3-D surfaces to conform with the surface location of mapped faults and structural discontinuities between section lines, where applicable. We then smoothed the linearly interpolated 3-D surfaces using inverse distance weighting (exponent p = 2) within an elliptical moving window with a major axis length equal to three-quarters of the cross-section spacing (3750 m), a minor axis length equal to one-quarter of the cross-section spacing (1250 m), and with the major axis of the ellipse oriented perpendicular to cross-section traces (azimuth = 143°). In parameterizing the smoothing ellipse dimensions, we selected a relatively long major axis length (less than the cross-section spacing width of 5 km, but three times that of the shorter axis length) to smooth interpolative artifacts between parallel cross sections while simultaneously prioritizing local cross-section interpretations.In a second step, we used the constructed geometry of the top of the Purisima Formation to derive uplift fields for the maximum, minimum, and preferred scenarios described above. We then retrodeformed these 3-D surfaces to recover the magnitude and extent of deformation recorded by the Purisima Formation since 4 Ma. For each reconstruction, we first reversed offsets along faults that cut the top of the Purisima Formation using the fault parallel flow algorithm (Ziesch et al., 2014). We only restored apparent dip-slip motion across regional faults, since strike-slip motion along the major regional faults in the area is largely unquantified (McLaughlin and Clark, 2004). Next, we unfolded the restored 3-D top-of-the-Purisima horizons using PetEx’s Move3D flexural slip algorithm, which preserves surface area within modeled horizons by allowing for sliding between bedding planes during deformation. We chose the modeled SAF as the pin surface, and unfolded the top-of-the-Purisima reconstructions to a horizontal, planar surface at 0 m (sea level) along a vertical plane trending roughly perpendicular to the SAF, and parallel to the 18 cross-section traces we created (azimuth = 53°). For each scenario, we unfolded top-of-the-Purisima reconstructions for the La Honda and Ben Lomond blocks separately, such that deformation resolved in each of the blocks during unfolding was not dependent on the distribution of deformation in the neighboring block. After unfolding each of the scenarios, we extracted maximum principal strain magnitudes calculated during the retrodeformation process for the maximum, minimum, and preferred top-of-the-Purisima Formation models to quantify deformation that has developed in the southwestern SCM since 4 Ma. We then compared maximum principal strain and uplift between the Ben Lomond and La Honda blocks for each reconstruction by calculating changes in uplift and strain within each block as a function of distance within the SCM bend. We subsampled uplift and strain fields using a 100 m grid, and divided vertices defining the uplift and strain fields into bins spaced every 1 km along the SCM-bend-relative coordinate system. We then calculated mean and quartile values within each bin, and plotted these metrics for each scenario as a function of distance along the SCM bend.AHe sample ages in the SCM show that cooling ages were, on average, younger in the La Honda Block than in the Ben Lomond Block (Fig. 6; Tables S1 and S4). The average of the weighted-mean sample ages within the La Honda Block was 8.9 Ma, as compared to 19.4 Ma for the Ben Lomond Block. Compiled individual crystal AHe age distributions for samples contained within the Ben Lomond and La Honda blocks show that crystal AHe ages within the La Honda Block cluster between 10 Ma and 0 Ma (median = 6.1 Ma), while crystal ages in the Ben Lomond Block are more evenly distributed from 50 Ma to 0 Ma (median = 20.0 Ma; Fig. 7). Many of the youngest AHe sample ages, previously reported in Baden et al. (2022), were located in the La Honda Block adjacent to the SAF in the southern SCM, where the geologic structure of the Ben Lomond and La Honda blocks sharply contrasts across the Zayante fault. The timing and magnitude of exhumation predicted for these samples appear to be the most closely associated with and/or diagnostic of SCM Bend uplift and exhumation (Baden et al., 2022). Other young AHe ages in the La Honda Block also appear to cluster around the La Honda and Pilarcitos faults. Samples with the oldest AHe ages were located within the Ben Lomond Block.Inverse thermal models calculated with QTQt for all AHe samples (and associated fission-track data, where applicable) reveal probability density functions that describe the range of time-temperature paths that can reproduce sample ages for each location (Fig. S1). Notable contrasts in plausible time-temperature paths between sample SC4 (in the Ben Lomond Block) and CB_SCM_BUT4 (in the La Honda Block) in cross section D*–D′ (Fig. 8) over the past 10 million years highlight differences in what we infer to be exhumation-driven cooling histories between the two structural blocks in the southern SCM. Time-temperature histories predicted for samples sourced from the La Honda Block in the central and northern portions of the SCM (cross sections J*–J′, M*–M′*, Figs. 9 and 10), generally predict higher maximum permissible temperatures and plausible cooling between 10 Ma and 0 Ma than samples within the Ben Lomond Block.Sample time-depth histories, which were calculated from time-temperature histories using constant geothermal gradients, predict plausible ranges for a sample’s depth beneath Earth’s surface through time since 50 Ma with tighter constraints since 10 Ma, as resolved by the AHe dataset alongside additional geological constraints (Fig. S1; Tables S2 and S3). This conversion assumes that increases in temperature correspond to burial, while decreases in temperature correspond to exhumation. For most samples in the Ben Lomond Block (e.g., SC4, Fig. 8), time-depth models require that minimal exhumation amounting to <1 km to have occurred since the formation of the SCM at 4 Ma. In contrast, bounds on exhumation predicted for many samples within the La Honda Block (e.g., CB_SCM_BUT4, Fig. 8) generally require larger magnitudes of recent exhumation (>2.5 km) and associated cooling to produce the AHe ages observed. Both the Ben Lomond and La Honda blocks contain samples that deviate from these trends, however. For example, samples CB_SCM_BUT7 and JG_SCM1 (Figs. S1xv and S1xx) reside within the La Honda Block, and both predict minimal exhumation since 4 Ma. In contrast, samples SG36 and SG45 (Figs. S1vii and S1viii) reside within the Ben Lomond Block, and both predict substantial post-Purisima exhumation.The 3-D geologic model of the SCM and associated structural cross sections highlight the variations in deformation that have developed within the Ben Lomond and La Honda blocks over geologic time scales spanning tens of millions of years (Figs. 8–11; Fig. S2). In the southern SCM (e.g., cross section D*–D′; Fig. 8), the pervasive cylindrical folding of the thick Cenozoic stratigraphic section evident in the La Honda Block contrasts sharply with the broad regional warping of the Ben Lomond Block, which hosts only a thin veneer of post-Oligocene sedimentary rock atop Salinian granitic basement. In the central SCM (e.g., cross section J*–J′; Fig. 9), deformation and uplift of Butano Ridge appears to be facilitated by reverse motion along the Butano fault, while distributed folding of Cenozoic strata persists along the high topographic crest adjacent to the SAF. In the northern SCM (e.g., cross section M*–M′*; Fig. 10), normal fault slip along the La Honda fault and other buried faults to the west developed after the late Miocene, and accommodated the deposition of the thick section of the Purisima Formation observed within the La Honda Basin.Uplift and strain fields calculated from the Purisima Formation reconstructions and retrodeformation (Fig. 12) constrain the plausible extent of deformation and uplift that has developed in the SCM since 4 Ma. The minimum reconstruction predicts that the Purisima Formation was not deposited throughout the SCM, but rather onlapped onto emergent, topographic highs at 4 Ma, which we delineate with white hatched lines (Fig. 12A). These highs include present-day Butano Ridge, Ben Lomond Mountain, and the high topographic crest adjacent to the SAF. In the southern SCM, inclined structural attitudes and geologic contiguity of the Purisima Formation within the La Honda Block in the vicinity of cross sections A–A′, B–B′, and C–C′ require uplift in excess of ~2 km. The minimum top-of-the-Purisima reconstruction predicts, at maximum, 2280 m of uplift in the La Honda Block (Fig. 12A), and a block-wide mean of 517 m. In the Ben Lomond Block, the minimum Purisima Formation reconstruction predicts a maximum uplift of 625 m, and a block-wide mean uplift of 375 m (Fig. S4). Predicted strain is highest in the La Honda Block, in the southern SCM (cross sections B–G; Fig. 12D). Within the SCM bend, binned mean uplift in the La Honda Block exceeds that in the Ben Lomond Block by an average of 344%, and binned mean maximum principal strain in the La Honda Block exceeds that in the Ben Lomond Block by >2500% (Fig. 13).The preferred reconstruction assumes that the Purisima Formation was deposited throughout the SCM southwest of the SAF, albeit in variable thicknesses. This is supported by the fact that there are no outcrops that expose strata containing coarse clastic material that would suggest the presence of emergent topography in the vicinity of deposition. In this scenario, uplift concentrates along the SAF in cross sections B–J, which coincides with the SCM restraining bend. Here, peak uplift values exceed 3.7 km near section G–G′. The preferred model also predicts moderate uplift (<2 km) along Butano Ridge, along the northern slopes of Ben Lomond Mountain adjacent to the Zayante fault, along the Pilarcitos fault in cross sections L–P, and along the northwestern slopes of Montara Mountain. The preferred Purisima Formation reconstruction predicts, at maximum, 3744 m of uplift in the La Honda Block (Fig. 12B), and a block-wide mean of 1144 m. In the Ben Lomond Block, the preferred Purisima Formation reconstruction predicts a maximum uplift of 1790 m, a block-wide mean uplift of 785 m (Fig. S4). Heightened predicted strain in the preferred model coincides with warping and increased uplift within the La Honda Block in cross sections B–J along the SCM bend. In this region, binned mean uplift in the La Honda Block exceeds that in the Ben Lomond Block by an average of 280%, and binned mean maximum principal strain in the La Honda Block exceeds that in the Ben Lomond Block by 615% (Fig. 13).The maximum Purisima reconstruction also assumes that the Purisima Formation was deposited throughout the SCM southwest of the SAF in variable thicknesses, though this model describes the maximum plausible depositional thickness and extent of the Purisima Formation, as informed by geologic and thermal constraints. Though the general structure and distribution of uplift and deformation resemble the preferred model, this model requires that uplift and exhumation have been far more extensive throughout the SCM since 4 Ma than that predicted by the preferred model. Uplift that concentrates along the SAF from cross sections B–J reaches a maximum of 5 km (the prescribed limit). Likewise, uplift along Butano Ridge, the northern slopes of Ben Lomond, the Pilarcitos fault, and the northwestern slopes of Montara Mountain approaches 3 km. Uplift throughout the majority of the Ben Lomond Block is still minimal, however. The maximum Purisima Formation reconstruction predicts, at maximum, ~5000 m of uplift in the La Honda Block (the imposed limit; Fig. 12C), and a block-wide mean of 1644 m. In the Ben Lomond Block, the maximum Purisima Formation reconstruction predicts a maximum uplift of 3199 m, and a block-wide mean uplift of 1016 m. As in the preferred model, heightened predicted strain in the maximum model coincides with warping and increased uplift within the La Honda Block in cross sections B–J within the SCM bend, and, to a lesser extent, along the SAF in sections I–P. We also observe heightened predicted maximum principal strains along the northern slopes of Ben Lomond Mountain in the Ben Lomond Block, and along the northern slopes of Montara Mountain. Within the SCM bend, binned mean uplift in the La Honda Block exceeds that in the Ben Lomond Block by an average of 150%, and binned mean maximum principal strain in the La Honda Block exceeds that in the Ben Lomond Block by 379%.Regardless of the choice of a minimum, preferred, or maximum scenario, uplift and maximum principal strain fields and distributions incurred since 4 Ma (Figs. 12 and 13) illustrate that (1) uplift increased within the SCM bend for both blocks and that (2) deformation and uplift concentrated in the La Honda Block relative to the Ben Lomond Block. This principally reflects the observation that the Ben Lomond Block is characterized by thin, broadly warped sedimentary units deposited atop crystalline Salinian basement >10–20 km from the SAF, while the La Honda Block contains thick, Cenozoic sedimentary progressions that are deformed into a series of roughly fault-parallel folds adjacent to the SAF. The contrast in recent (younger than 4 Ma) deformation across structural block boundaries is particularly evident in the southern SCM within the SCM bend, where the two blocks are juxtaposed against one another in close proximity to the SAF (e.g., Fig. 8). In this area, the folded Purisima Formation beds preserved adjacent to the SAF and within the Glenwood syncline of the La Honda Block contrast with relatively flat-lying Purisima Formation beds atop the Ben Lomond Block.We reconstructed the three top-of-Purisima horizons to conform to permissible bounds based on surficial exposure of the Purisima Formation, well data, and thermal modeling constraints. Indirect geologic predictions associated with these reconstructions vary substantially, and, in the case of the minimum model, present a number of inconsistencies with geologic observations in the SCM. We use these inconsistencies to support our adoption of the preferred Purisima reconstruction. For example, the minimum model assumes that substantial paleobathymetry and emergent paleotopography is fairly widely distributed and in close proximity to active depositional settings during the deposition of the Purisima Formation. While this is plausible, widespread conglomeratic deposits in the Purisima Formation within the SCM are largely absent (Touring, 1959; Powell et al., 2007), which appears indirectly inconsistent with minimum model predictions. Both the preferred and maximum models assume that the Purisima is present, in some thickness, throughout the SCM at 4 Ma, and that paleobathymetry is minimal at this time, which appears consistent with Purisima stratigraphy. Additionally, the northeastern, elongate ridge of paleotopography predicted in the minimum model would presumably hinder the delivery of sediment across the SAF during deposition of the Purisima Formation. The minimum model also predicts widespread exhumation on the order of ~10 m/m.y., which is exceedingly low for crust that resides within a tectonically active plate boundary and adjacent to a demonstrably active restraining bend (Hecht and Oguchi, 2017). Even though geologic evidence does not refute the maximum model reconstruction, we present this model primarily to bound the realm of plausible predictions for uplift and deformation in the SCM.In this work, we primarily use the suite of AHe ages and associated thermal models to constrain recent deformation that has developed in the Ben Lomond and La Honda blocks since 4 Ma. However, these data may also provide useful constraints on tectonic events that occurred throughout the history of the strike-slip fault system. Whereas young ages have been substantially reheated and exhumed from depth (thus resetting any sensitivity to prior tectonic events), older AHe ages in the dataset record unroofing signatures from older than 10 Ma, and could be used to inform geologic structural reconfigurations during the evolution of the plate boundary. Within the Ben Lomond Block, long-term time-temperature histories associated with sample SC4 (Fig. 8; Fig. S1vi) predict reheating associated with burial of 2–3 km beneath the Butano and San Lorenzo Formations until ca. 30 Ma. Dip-slip motion across the Zayante fault associated with transtension across the plate margin (Fig. 4C) may have caused subsequent unroofing. We observe similar cooling histories predicted for sample SC3 (Fig. S1v). Samples SC1 and SC2_Re, between sections F and G on Ben Lomond Mountain, predict rapid unroofing at ca. 20 Ma (Figs. S1iii and S1iv), which is broadly consistent with the timing of the unconformity observed at the base of the Lompico Sandstone and the Monterey Formation (Fig. 2), and have subsequently been minimally reheated. The difference in unroofing histories between samples SC3 and SC4 and samples SC1 and SC2_Re may represent exhumation associated with substantial dip-slip motion along the Ben Lomond fault at ca. 20 Ma. Cooling histories in the La Honda Block are dominated by recent exhumation, though sample time-temperature paths associated with older ages generally predict residence within ~1 km of Earth’s surface since ca. 10 Ma, assuming geothermal gradients of 25–35 °C. This observation is consistent with the unconformable deposition of the Santa Margarita Sandstone and Santa Cruz Mudstone atop deformed sedimentary rocks of Oligocene and Eocene age along the western slopes of Butano Ridge, inland along cross section H, and east of the Zayante fault in cross sections E–F.Additional AHe ages that fall outside of the Purisima reconstruction extents also divulge information regarding uplift and deformation in the SCM. West of the SGHF, the AHe age for sample JG_PB5 and the associated thermal models (Fig. S1xxviii) permit 0.5–1.75 km of exhumation within Pigeon Point Block since 10 Ma. Recent inferred exhumation is substantially higher between the SAF and Pilarcitos faults, where thermal models associated with sample SG6 (Fig. S1xxxi) predict 2–3 km of unroofing since 4 Ma. This exhumation may result from shearing and rotation of rocks within the Pilarcitos Block as slip from the SGHF partitions onto the SAF-oblique Pilarcitos and La Honda faults. South of the SCM, exposures of the gabbro of Logan (the inferred basement of the La Honda Block) that crop out adjacent to the SAF and are onlapped by the Purisima Formation predict exhumation to within ~1.5 km of Earth’s surface by 6 Ma (see sample CB_SCM_LGN1; Fig. S1xviii).The combination of apatite (U-Th)/He ages and structural cross sections that were used to compute post-Pliocene uplift and strain fields suggest that deformation within the SCM localizes within particular lithotectonic blocks as the plate boundary evolves. While previous work has been successful in recreating the long-wavelength attributes of the restraining bend tectonics of the SCM using relatively simple models (Anderson, 1990, 1994; Bürgmann et al., 1994; Baden et al., 2022), this analysis shows that the specific way in which these long-wavelength characteristics are manifest at particular locations depends on the distribution and material properties of lithotectonic terranes present prior to 4 Ma. These pronounced changes in material properties between adjacent lithotectonic terranes may reflect the fact that the thick sedimentary successions atop formerly extended crust in the La Honda Block and the sediment-veneered crystalline granitic rocks of the Ben Lomond Block (Fig. 3) have substantially different vertically integrated strengths. This contrast is particularly distinct in cross sections A–E, where cylindrical folding of the Cenozoic sedimentary progression that generates deformation and uplift in the La Honda Block within 5–10 km of the SAF contrasts with minimal deformation resolved in the Ben Lomond Block immediately across the Zayante fault (e.g., Fig. 8). In the SCM, we assert that crustal heterogeneity augments the effect of fault proximity in localizing deformation near the SAF to produce the observed contrasts in deformation and rock uplift across lithotectonic block boundaries.There are several uncertainties in the time-temperature histories for each AHe sample that need acknowledgment. First, the temperature sensitivity of He diffusion in apatite varies depending on the effective uranium content (eU) in an analyzed apatite crystal (Shuster et al., 2006; Flowers et al., 2009; Shuster and Farley, 2009; Willett et al., 2017). This temperature sensitivity (the HePRZ) ranges from 30 °C (complete He retention in low eU crystals) to 90 °C (complete He diffusion in high eU crystals), with a typical range of ~40–80 °C (Willett et al., 2017). (U-Th)/He measurements in apatite thus provide minimal information pertaining to time-temperature predictions that reside outside of the plausible HePRZ. For example, the Purisima Formation in the minimum thickness scenario is <0.25 km in places, which may not generate a resolvable thermal signature in the presence of substantial subsequent exhumation. Second, best-fitting thermal histories did not always produce modeled AHe crystal ages that closely corresponded to observed crystal ages. For relatively poor model fits (e.g., JG_SCM1, Fig. S1xx; CB_SCM_V; Fig. 9), we relied more heavily on surrounding samples and adjacent cross-section interpretations. Finally, we use constant geothermal gradients of 25 °C/km and 35 °C/km to map temperature to depth to calculate exhumation histories. In this context, regional variations in the geothermal gradient are almost certainly influenced by differences in the thermal properties of rock across lithotectonic terranes (Lachenbruch and Sass, 1980), and the geothermal gradients evolve as erosion exhumes warm, formerly buried crust (e.g., Ehlers, 2005). However, we assume that the range of 25–35 °C/km adequately bounds the range of plausible gradients at ca. 4 Ma (prior to SCM-bend-induced tectonism and exhumation), when we use time-temperature and inferred time-depth models to extract exhumational bounds for the Purisima Formation reconstructions. The relatively modest ~20% increase in regional geothermal gradients in the SCM associated with rock uplift and exhumation (Ehlers, 2005; Baden et al., 2022) is unlikely to influence model predictions at and before 4 Ma that predate range formation.Our 18 structural cross sections are subject to geologic uncertainty associated with the complex plate margin. Numerous geologic and geophysical studies have reconstructed the tectonic and stratigraphic evolution of the plate boundary (Cummings et al., 1962; Graham, 1978; Stanley, 1984; Graham et al., 1989, 1997; Brabb et al., 1998; Page et al., 1998; Wentworth et al., 1999; Wagner et al., 2002; Jachens et al., 2004; McLaughlin and Clark, 2004; Sharman et al., 2013; Gooley et al., 2021); however, the large-scale tectonic histories, regional fault zone structures, and unit thicknesses and depositional extents remain coarsely constrained. In our geologic model, subsurface interpretations are poorly constrained in the absence of surface exposure (for buried structures and units) and well control. Additionally, well compilations and reinvestigations suggest that certain horizon designations in existing wells may be susceptible to misinterpretation (Brabb et al., 2001). In these cases, we rely on geological inference, continuity between adjacent cross sections, and interpretations of previous investigators. Notably, cross-section interpolations used to construct the geologic model and Purisima reconstructions also omit localized structural features, and generalize geologic relationships from section to section. However, the high density of cross sections, which are spaced approximately every ~5 km parallel to the SCM, bolster confidence in the 3-D geologic model derived from these sections.In portions of the La Honda Block where constraints on uplift, exhumation, and structural attitude are minimal or absent, our geologic reconstructions conservatively minimize the amount of deformation that may have accumulated. Along the SAF in cross sections D through J, where structural constraints on deformation within the (now absent) Purisima Formation have been removed by erosion, and AHe thermal models predict relatively large exhumational magnitudes, our Purisima reconstructions conservatively limit the extent of plausible folding and associated deformation of the Purisima Formation adjacent to the SAF. Although the omission of plausible folding in the Purisima Formation may serve to reduce modeled predictions for maximum principal strain adjacent to the SAF in the La Honda Block (Fig. 12, cross sections D–J), predictions for strain derived from our conservative reconstructions of the Purisima Formation in the La Honda Block are already larger than those for the neighboring Ben Lomond Block (Figs. 12 and 13). Incorporating additional SAF-adjacent structural folding in Purisima reconstructions would therefore increase the contrast in deformation accommodated by the two structural blocks and further highlight the importance of strain-partitioning that occurs during mountain building within complex plate margins over millions of years.Our reconstructions and analyses do not explicitly incorporate regional strike-slip fault motion that occurs along faults residing between the SAF and the SGHF in the southwestern SCM (i.e., the Butano, La Honda, Pilarcitos, Ben Lomond, and Zayante faults). Along these faults, slip magnitudes are assumed to be small, albeit poorly constrained, and the timing and distribution of fault slip is uncertain (McLaughlin and Clark, 2004). For all regional faults excluding the Zayante fault, we omit the effect of strike-slip motion because plausible magnitudes of strike-slip motion along these structures since 4 Ma (likely <5–10 km) would likely only minimally affect our results. Although predictions for rock uplift in the SCM may concentrate along these regional faults in certain locations (discussed below), along-fault variations in predicted uplift and exhumation are minimal, reducing the artifactual impact of omitting minor strike-slip offsets during model retrodeformation. Similarly, we do not explicitly account for strike-slip motion along the Zayante fault, but we unfold the 3-D models for the top-of-the-Purisima reconstructions for the Ben Lomond and La Honda blocks separately (see Methods section), such that the distribution and magnitude of deformation and resultant rock uplift in each crustal block are independent from one another and are insensitive to slip along the Zayante fault. The Zayante fault lacks distinct geomorphic expression sporadically along its length (McLaughlin and Clark, 2004), which may indicate minimal Holocene slip, and gradual increases in uplift with distance along the SCM bend in both the Ben Lomond and La Honda blocks (Fig. 13B) appear spatially in sync and consistent with contemporaneous SCM-bend-induced uplift and advection (Baden et al., 2022). Thus, we favor a model in which relative displacement between the Ben Lomond and La Honda blocks accommodated by strike-slip motion along the Zayante fault since 4 Ma has been relatively minimal.In this work, we assume that the top of the Purisima Formation represents an approximate sea-level datum that records deformation and uplift associated with the formation of the SCM that began 4 million years ago. Here, we use stratigraphic relationships, age correlations, and inferred depositional environments of the Purisima Formation summarized in Powell et al. (2007) to evaluate this assumption, and to contextualize our three top-of-the-Purisima reconstruction models in a geologic and tectonic framework. First, we note that the Santa Cruz and San Mateo coast sections described in Powell et al. (2007) have been only loosely correlated with one another, since the middle and upper portion of each section lacked diagnostic diatoms for age control. Variability in the depositional environments through time in each section, in addition to substantial differences in integrated stratigraphic thicknesses, further obfuscates the relationship between these sections. However, we propose that the structural configuration of the SCM during the deposition of the Purisima Formation from 6.9 Ma to 4 Ma, and the subsequent initiation of uplift of the SCM during the final stages of Purisima deposition, may offer an explanation for the relationship between these two sections, and provide context for their utility in this study.The preservation of the Purisima Formation from Point Reyes to onshore Monterey Bay, California, demonstrates that the minimum latitudinal range over which this unit was deposited (~180 km) stretches beyond the northern and southern extents of the SCM. The preservation of correlative submarine units east of the SAF in the southwestern San Joaquin Valley (the cross-SAF piercing point and sediment route for the Purisima Formation; Powell et al., 2007) suggests that the Purisima Formation may have been deposited over the entirety of the present-day SCM. During the earliest stages of Purisima deposition (i.e., ca. 7–5 Ma), lower portions of both the Santa Cruz and San Mateo Coast sections were deposited in marine shelf conditions of <100 m water depth. Baden et al. (2022) suggested that the reorganization of plate motion between 6 Ma and 4 Ma initiated uplift in the vicinity of the SCM bend. In the Santa Cruz Section in the southern SCM, the hiatus in deposition between 4.9 Ma and 3.65 Ma may correspond to initial uplift and associated cessation of deposition and onset of erosion in this area. This change in plate motion may have also contributed to local extension in the present-day La Honda Basin, which led to syndepositional normal faulting and deepening across SAF-subparallel faults as inferred in cross section M–O and in industry exploration wells (Horn, 1983; Hector, 1986; Wright, 1990). Inferred paleo-depositional water depths increased from <100 m to between 150 m and 750 m during deposition of the Pomponio member from 5.3 Ma to 3.7 Ma (Powell et al., 2007). Shales and mudstones dominate the lithology of the Pomponio member; the lack of clastic material suggests that extensive paleotopography and/or high bathymetric relief in close proximity to the La Honda basin was likely absent during this time.Following the initiation of uplift in the SCM at 4 Ma, we assume that widespread deposition of the Purisima Formation ceased, while SCM-bend-related deformation continued and topography developed. The preservation of the Santa Clara Formation unconformably deposited atop Miocene and older units adjacent to sample CB_SCM_V along cross section K requires that this portion of the SCM was emergent (with any underlying Purisima deposits removed) by 3 Ma. In the San Mateo coast section, Powell et al. (2007) estimated that Purisima deposition in shallow waters (<100 m) continued until 2.8 Ma. We propose that this deposition was locally confined to the La Honda Basin, where additional deposition following range uplift at 4 Ma did not exceed ~200 m. Following the depositional hiatus in the Santa Cruz section, Purisima deposition also resumed in the southern SCM. Where present, upper portions of the Santa Cruz section are clast-inundated, and represent foreshore, shorefront, and potentially non-marine deposits. We propose that these deposits, which amounted to <200 m in thickness, represent the gradual cessation of post–4 Ma Purisima deposition south of the SCM as crust advected into and through the SCM bend. Therefore, it is likely that, in a broad sense, the top of the Purisima Formation provides an approximate sea-level datum.Predicted rock uplift fields for the preferred and maximum Purisima Formation reconstructions reveal multiple spatially distinct uplift domains since 4 Ma (Fig. 12). These domains include localized uplift adjacent to the SAF, broadly distributed uplift of Ben Lomond Mountain and within the central portion of the Ben Lomond Block, uplift of the northern slopes of Ben Lomond Mountain adjacent to the Zayante fault, uplift within the central portions of Butano Ridge, and uplift along the La Honda and Pilarcitos faults. SCM-bend-induced deformation produced the localized uplift adjacent to the SAF (Baden et al., 2022), which preferentially concentrated in the La Honda Block. SCM bend deformation likely also produced broad, regional warping and uplift of the Ben Lomond Block and of Ben Lomond Mountain as they advected into and through the SCM bend, as is suggested by the increase in predicted uplift in the Ben Lomond block alongside the La Honda Block within the SCM bend (Fig. 13).The absence of the Purisima Formation, in conjunction with sparse preservation of Miocene stratigraphy, atop Ben Lomond Mountain (the highest topography within the Ben Lomond Block) obfuscates the tectonic history of the granitic massif, and its relative position to sea level, since SCM bend uplift began at 4 Ma. However, minimal exhumation predicted by thermal models for AHe samples from the central and southern portions of Ben Lomond Mountain (i.e., SC1, SC2, SC3; Figs. S1iii–S1v) suggest that Ben Lomond Mountain was likely only minimally buried prior to and during Purisima Formation deposition. Foreshore and shorefront deposits at the top of the Santa Cruz section of the Purisima Formation (inferred a depositional age younger than 3 Ma; Powell et al., 2007) suggest that southern portions of Ben Lomond Mountain may have resided at or near sea level well after SCM-bend-induced uplift began at 4 Ma, likely due to the fact that this portion of the Ben Lomond Block had not yet passed into and through the zone of bend-induced uplift. Preserved Quaternary marine terraces along the coastal slopes of Ben Lomond Mountain constrain the rate, timing, and magnitude of uplift since ca. 600 ka (Lajoie et al., 1979; Anderson, 1990; Gudmundsdottir et al., 2013), and suggest that uplift rates along the coast of the Ben Lomond Block over this time period are highest in the southern SCM, consistent with the inferred peak uplift zone within the SCM bend southwest of the SAF (Anderson et al., 1990; Gudmundsdottir et al., 2013; Baden et al., 2022). We use these lines of evidence to infer that the northwestern portions of Ben Lomond Mountain began to uplift at ca. 4 Ma as a result of SCM-bend-induced deformation, and preserved, more recent signatures of uplift now concentrate along the southwestern slopes.Interestingly, the remaining uplift signals occur along the Zayante, Butano, and Pilarcitos faults, which all trend obliquely to the SAF as they span the distance between the SAF and the SGHF. Recent deformation, uplift, and exhumation along these structures suggest that slip from the SGHF (geodetically inferred slip = 2.4 ± 1.0 mm/yr; d’Alessio et al., 2005) may partition onto these oblique fault structures as the SGHF approaches and eventually merges with the SAF <50 km north of the SCM. Along the Pilarcitos and La Honda faults in the La Honda Block, dense thermochronologic sampling and associated geologic reconstructions also suggest that restraining and releasing structures along the length of the La Honda and Pilarcitos faults have generated resolvable changes in uplift and exhumation (Fig. 12). While Pilarcitos-fault-adjacent rock uplift in cross sections L–O increases, and is coincident with a left-bend in the right-lateral fault system (restraining structure), Pilarcitos-fault-adjacent rock uplift in cross sections P–Q decreases, and spatially corresponds to a subsequent right bend in the right-lateral fault system (releasing structure). The Zayante, Butano, and La Honda/Pilarcitos faults have been historically inactive, and are frequently geomorphically indistinct (McLaughlin and Clark, 2004). However, these faults also traverse steep, densely vegetated terrain hosted within sandstones and mudstones in the SCM, and are subject to erosive processes that may obscure geomorphic and geologic expressions of faulting. Despite the lack of direct evidence for recent fault slip, time-integrated rock uplift predictions suggest that these faults may thus pose a formerly unrecognized seismic hazard to surrounding communities.We infer that the drastic differences in tectonic history and resultant stratigraphic structure of the La Honda and Ben Lomond blocks establish a distinct contrast in vertically integrated crustal strength within the SCM. The stratigraphic differences that establish this contrast are observed in subsurface well data, inferred from surface outcrops and structural attitudes, and are expressed in regional seismic velocity profiles (Catchings et al., 2004; Zhang et al., 2018; Fuis et al., 2022), where an anomalous low velocity zone characteristic of sedimentary rock and/or fractured basement extends to a depth of ~8–10 km between the SAF and Zayante faults. Regional gravity and magnetic surveys also constrain geological interpretations of bulk stratigraphic thicknesses and basement composition in the SCM (Jachens et al., 2004). Despite the wealth of geologic and geophysical data, however, quantitative assessments of bulk crustal strength for the two lithotectonic blocks remain difficult to constrain.The strength of the upper crust is predominantly determined by its resistance to brittle failure. Rock strengths are often measured in the laboratory and parameterized as a function of cohesion and internal angle of friction (i.e., Mohr-Coulomb theory; Coulomb, 1776; Mohr, 1900; Labuz and Zang, 2012), but these assessments inadequately characterize the strength of fractured and/or heterogeneous layered rock at the scale of the upper crust. Instead, we compare approximations for uniaxial compressive strength (UCS) derived from regional geophysical observations to infer differences in strength of the upper crust between the Ben Lomond and La Honda blocks. To estimate UCS at the scale of the crust, we use empirical relationships between UCS and seismic p-wave velocities (e.g., Sharma and Singh, 2008; Minaeian and Ahangari, 2013; see Supplemental Text S2). In the epicentral region of the Loma Prieta Earthquake, differences in p-wave velocities at depths of ~2.5 km, 5 km, and 10 km (Lin and Thurber, 2012; Fuis et al., 2022) suggest that the UCS of Salinian granites in the Ben Lomond Block exceeds that of the Cenozoic sedimentary rocks in the La Honda Block by ~16%, 27%, and 12%, respectively (Table S5). Although rock strength also evolves as a function of the coefficient of internal friction of rock alongside UCS (i.e., the linearized form of the Mohr-Coulomb failure criterion; Colmenares and Zoback, 2002) these first-order approximations are consistent with the inference that differences in lithotectonic stratigraphy impart a contrast in crustal strength.Mechanisms such as widespread flexural slip and fractured basement rocks may provide an explanation for these differences in integrated strength. Flexural slip has been shown to amplify folding-related deformation, particularly when the spacing between sedimentary layers separated by weak bedding interfaces is small. Johnson (2018) showed that folds within sedimentary layers whose thicknesses are much less than the wavelength of the fold itself grow faster than folds in non-layered rock, and fold amplitudes increase by a factor of 4–5. In the sedimentary Cenozoic stratigraphy of the La Honda Block, the interbedded sandstones, siltstones, and mudstones within each stratigraphic unit may mechanically divide the rock mass into the closely spaced and mechanically contrasting layers required for drastic fold amplification and strength reduction. These mechanical subdivisions and flexural slip planes within the 5-km-thick sedimentary succession reduce the effective elastic thickness (and resistance to bending and folding) of the deforming section significantly. Following calculations presented by Otsubo et al. (2020), the effective elastic thickness for a 5-km-thick section composed of individual bedding planes spaced 10 m apart that each accommodate flexural slip is less than 5% of an undivided, 5-km-thick sedimentary succession that does not allow for flexural slip. Extensive fold-driven exhumation predicted in cross-section reconstructions within the La Honda Block adjacent to the SAF (i.e., sections B–G) appears consistent with localized cylindrical fold amplification and strength reduction.Beneath the Cenozoic stratigraphy, potential pre-existing extensional fault structures in the gabbroic basement of the La Honda Block may have been reactivated due to reductions in frictional strength resulting from previous slip (e.g., Ruh and Vergés, 2018). Although damage within the gabbroic basement hasn’t been quantified, even moderate reactivation of basement faults beneath sedimentary cover can initiate folding in overlying strata (Butler et al., 2006). Additionally, analogue models show that contrasts in lithospheric strength at depth tend to localize deformation within relatively weak parcels of crust surrounding strong lithospheric domains, even when upper crustal rheologies are similar (Calignano et al., 2015). Those model results suggest that tectonically inherited weaknesses in the gabbroic basement of the La Honda Block associated with prior extension may uniquely contribute to strain localization. Numerical predictions also suggest that formerly extensional basins may preferentially accommodate deformation in contemporary contractional tectonic regimes (Vauchez et al., 1998). Although geothermal gradient can also strongly influence integrated lithospheric strength (e.g., Kusznir and Park, 1987), and heat flow measurements between the Ben Lomond and La Honda blocks differ substantially (67 mW/m2 versus 100 mW/m2, respectively; Lachenbruch and Sass, 1980), we assume that contrasts in heat flow at the surface between the two blocks is largely driven by differences in thermal diffusivity between the lithologies present in each block (Robertson, 2013), and differences in exhumation and resultant heat advection (e.g., Ehlers, 2005). We, therefore, assume that contrasts in crustal strength are driven by lithotectonically imparted mechanical weaknesses, as opposed to contrasts in geothermal gradient at depth. Inelastic deformation that develops in response to elevated stresses in crust surrounding the SCM bend may therefore tend to preferentially localize within folds in the Cenozoic sedimentary rocks and along potentially reactivated basement fault structures in the La Honda Block. It is important to acknowledge that these effects are poorly constrained and currently difficult to quantify. Nonetheless, there is reason to think that differences in the characteristics of the stratigraphic section and state of the underlying bedrock that were present throughout this area at 4 Ma provide a plausible explanation for the differences in rheological of the La Honda and Ben Lomond blocks, which may preferentially partition deformation into the lower-strength areas.Fault zone complexity and material heterogeneity have been shown to influence fault position, slip behavior, and the distribution of deformation surrounding shear zones over time scales ranging from seconds to millions of years, and spatial scales ranging from sub-millimeter crystalline microstructures to plate margins (e.g., Tesei et al., 2013; Niemi et al., 2013; Fitzgerald et al., 2014; Lindsey et al., 2014; Michie, 2015; Zielke et al., 2017; ten Brink et al., 2018; Huang, 2018; Townsend et al., 2021). In the SCM along the SAF, we show that deformation and uplift localize within crust dominated by thick sedimentary successions and formerly extended basement (Figs. 12 and 13). Our results suggest that crustal heterogeneity resulting from the juxtaposition of structural blocks of variable integrated crustal strength strongly influences the distribution of deformation along continental transform plate margins.Rheological heterogeneity has been invoked elsewhere along the SAF to explain interblock contrasts in deformation, uplift, and exhumation. In southern California, the large restraining bend in the SAF (the “Big Bend,” near Los Angeles, California) cuts through a complex arrangement of fault-bounded, geologically variable crustal blocks as it transects the Transverse Ranges (Blythe et al., 2000, 2002; Spotila et al., 2001). Previous investigations have suggested that fault proximity and local fault geometry within the Transverse Ranges predominantly control the distribution and magnitude of deformation surrounding the SAF, because rock uplift inferred from low-temperature thermochronology appears most pronounced within 20 km of the fault’s location at Earth’s surface (Spotila et al., 2007b). However, numerous crustal fragments along the SAF appear more deformed and more deeply exhumed than crustal blocks to which they are immediately adjacent. In one such location, Niemi et al. (2013) suggested that relative contrasts in strength between basement structures of adjacent blocks in the northern Transverse Ranges concentrated deformation and exhumation within the San Emigdio Mountains. In this region, weak, extended, and underplated lithosphere of the San Emigdio Block resides between relatively undeformed granitic basement and old, cold ophiolitic crust that is inferred to be strong (Chapman et al., 2010; Niemi et al., 2013). Interestingly, mafic basement assemblages within the deformed San Emigdio Block (i.e., gabbro of Eagle Rest Peak) have been correlated with the gabbro of Logan across the SAF (the inferred basement of the La Honda Block in the SCM), which resides northwest of the San Emigdio Mountains by >300 km (Ross, 1970). The fact that deformation has been shown to localize in both the La Honda and San Emigdio blocks, which host correlative weak, extended, mafic basement rocks, may indicate that these basement structures serve as pronounced weaknesses in the lithosphere along the SAF. Crustal heterogeneity and resultant lithosphere-scale strength contrasts may also dictate the distribution and magnitude of deformation in the western Transverse Ranges, where thick sequences of sedimentary marine basin deposits have been extensively thrust-faulted, uplifted, and exhumed (Atwater, 1998; Townsend et al., 2021).Rheology-driven strain localization may also be important elsewhere within the Transverse Ranges in locations where it has not been previously recognized. For example, AHe ages sampled within the Yucaipa Ridge Block in the San Bernardino Mountains are far younger than ages sampled from surrounding blocks, implying that exhumation rates within this narrow crustal sliver adjacent to the SAF have been much higher than in surrounding regions (Spotila et al., 1998, 2001). Spotila et al. (1998, 2001) attribute extensive exhumation of the Yucaipa Ridge Block to changes in fault activity and geometry within an evolving plate margin. However, differences in basement lithology reported in regional geologic maps and structural interpretations (Matti et al., 2003; Cromwell and Matti, 2022) suggest that rheological contrasts between neighboring crustal blocks may play a more important role in localizing strain within the Yucaipa Ridge Block than previously appreciated. Specifically, the foliated schists and highly fractured and sheared cataclastic granitic rocks (San Bernardino-type basement) within the Yucaipa Ridge Block differ from and are likely far weaker than the coherent granitic rocks (Peninsular and Mojave-type basements) within neighboring crustal blocks. Our results suggest that this inferred contrast in strength may help to explain the stark differences in observed AHe ages and associated exhumation rates in the Yucaipa Ridge Block relative to neighboring blocks. Importantly, the material properties of rocks near the fault zone and its evolution may not be independent of one another. Large deformations that accrue within geologically weak materials adjacent to a fault zone might change the geometry of the fault itself, which could alter the stresses and deformations along the deforming boundary as fault motion accrues.Broad trends in topography and exhumation along the Pacific–North American plate boundary have been attributed to large-scale transpression across the SAF (Spotila et al., 2007a), and other studies suggest that near-field uplift may be controlled by local structural effects and material anisotropy (Buscher and Spotila, 2007; Spotila et al., 2007b). However, our results, alongside previous studies (Niemi et al., 2013; Townsend et al., 2021), suggest that rheologically driven strain localization strongly influences the partitioning of deformation between crustal blocks along the SAF. We posit that the influence of rheologically driven strain localization operates over larger spatial scales as well, and may help to explain contrasts in topographic relief, fault density, and seismicity across the SAF at the scale of the Big Bend as a whole. West of the SAF within the Big Bend, the San Gabriel Mountains and the western Transverse Ranges are composed of geologically heterogeneous crustal blocks and displaced crustal fragments that have been rotated and displaced relative to one another (Atwater, 1998). Many of these crustal blocks host thick successions of sedimentary rock deposited within former marine basins (e.g., the western Transverse Ranges; Hornafius et al., 1986; Atwater, 1998; Townsend et al., 2021), and others host highly deformed crystalline rocks (e.g., the Yucaipa Block; Matti et al., 2003; Cromwell and Matti, 2022). This region hosts substantial topographic relief, is extensively faulted, and, in some locations, is highly exhumed. Topographic relief and thrust faulting east of the SAF in the Mojave Block, which consists of crystalline granitic rocks, is minimal in comparison (Hauksson, 2011). Admittedly, this supposition generalizes the complex configuration of the crust and lithosphere in this region, which is demonstrably complex at the scale of the Big Bend (Plesch et al., 2007; Shaw et al., 2015). Additionally, fault activity, topographic production, and exhumation west of the SAF often pre-dates restraining bend deformation surrounding the Big Bend at 5 Ma (Crowell, 1979), so topographic and exhumational signals in the San Gabriel Mountains and western Transverse Ranges may be inherited from previous tectonic configurations (Spotila et al., 1998; Blythe et al., 2000, 2002; Townsend et al., 2021). Despite these complexities, increased off-SAF seismicity in the San Gabriel Mountains and the western Transverse Ranges relative to seismicity observed east of the SAF (Allen et al., 1965; Hauksson, 2000, 2011) supports the hypothesis that weak, faulted crust west of the SAF may predominantly accommodate off-fault deformation surrounding the Big Bend in the current tectonic configuration.Strain localization resulting from rheologic heterogeneity may also influence deformation along other strike-slip fault systems around the world. Along the Pacific–North American plate boundary in Alaska (USA), ten Brink et al. (2018) showed that, in contrast to the SAF, the vast majority of fault slip and associated deformation along the Queen Charlotte fault concentrates within a narrow zone along the plate margin. Ten Brink et al. (2018) attribute the localized fault zone to strong contrasts in rheology between oceanic and continental crust in the region. Near Icy Point, Alaska, contrasts in AHe cooling ages on either side of a restraining bend in the Fairweather–Queen Charlotte fault (Lease et al., 2021) may also be explained by contrasts in crustal strength. In this area, young AHe ages to the southwest of the Fairweather–Queen Charlotte fault reside within folded and faulted sedimentary rocks, whereas older ages northeast of the fault reside within metamorphic and plutonic rocks, which may be stronger. Along the Denali fault in the Alaska Range, Fitzgerald et al. (2014) invoke differences in lithospheric strength between tectonic terranes to explain the localization of deformation and the resultant creation of the highest mountain peaks in North America along the Mount McKinley and Mount Hayes restraining bends. Apatite fission track and zircon (U-Th)/He ages collected by Benowitz et al. (2022) from within the weak Alaska Range suture zone delineated by Fitzgerald et al. (2014) demonstrate that exhumation rates in this weak parcel of crust are also very high. Along the Alpine fault in the Southern Alps in New Zealand, patterns of uplift and exhumation do not vary between fault-bounded crustal blocks, but rather appear correlated with differences in fault slip rates and fault geometries at depth driven by rheological differences in the lithospheric mantle (Little et al., 2005; Eberhart-Phillips et al., 2022).In the SCM, strain partitioning exemplified by the modeled differences in post–4 Ma rock uplift within the Ben Lomond and La Honda blocks (Figs. 12 and 13) is not clearly expressed in the interseismic motions measured across the area over decades using satellite geodesy (d’Alessio et al., 2005; Bürgmann et al., 2006). Baden et al. (2022) found that the interseismic motions were likely dominated by elastic bending of the crust, while subtle interseismic plastic strain accumulation appeared to produce the geologic deformations observed within the SCM bend. In the present case, the fact that the large differences in long-term deformation observed between the two blocks appears invisible to interseismic surface motions suggests that these short-term motions are dominated by elastic bending of crustal blocks with similar stiffnesses, while long-term deformation reveals the contrast in the integrated yield strengths of the crustal blocks. We posit that the steady accumulation of inelastic flexural slip and accommodation of deformation along pre-existing basement fractures and faults within the La Honda Block may represent the manifestation of interseismic plastic deformation in the SCM predicted by Baden et al. (2022). Moreover, bedding plane rotation and disruption by fracturing and faulting could also serve to strain-harden the crust as deformation accumulates, as was invoked in the rheological prescriptions for the mechanical models in that study. These mechanisms provide a means of reconciling the observation that geologically observed deformation localizes in particular lithotectonic blocks, while geodetic measures of crustal deformation often fail to identify these differences. In this context, careful geologic studies may provide valuable insight into both short-term deformation and strain localization processes along plate boundaries that cannot be resolved geodetically.Rock uplift fields predicted from a 3-D geologic model of the SCM showcase how Earth’s crust accommodates earthquake cycle deformation heterogeneously over geologic time. Our results suggest the following:Deformation, rock uplift, and exhumation preferentially localize in parcels of relatively weak crust along geologically complex transform plate boundaries.Contrasts in integrated crustal strength between crustal blocks may be driven by differences in geologic histories and resultant stratigraphic architecture and lithotectonic structure. Inferred differences in the mechanical behavior of bulk rock masses establish these rheological contrasts.Rheology-driven strain partitioning appears to strongly influence the distribution of deformation surrounding complex strike-slip fault zones (Spotila et al., 1998, 2001; Niemi et al., 2013; Fitzgerald et al., 2014; Lease et al., 2021; Townsend et al., 2021). This effect may play a major role in localizing deformation in similar settings around the globe.Deformation, rock uplift, and exhumation preferentially localize in parcels of relatively weak crust along geologically complex transform plate boundaries.Contrasts in integrated crustal strength between crustal blocks may be driven by differences in geologic histories and resultant stratigraphic architecture and lithotectonic structure. Inferred differences in the mechanical behavior of bulk rock masses establish these rheological contrasts.Rheology-driven strain partitioning appears to strongly influence the distribution of deformation surrounding complex strike-slip fault zones (Spotila et al., 1998, 2001; Niemi et al., 2013; Fitzgerald et al., 2014; Lease et al., 2021; Townsend et al., 2021). This effect may play a major role in localizing deformation in similar settings around the globe.See Table A1 for the geologic units in the Santa Cruz Mountains, California, USA. For purposes of reproducing or extending the analysis, all data that support the findings of this study (i.e., geochemical measurements, thermal models, Move3D models and their associated products), in addition to all code necessary to reproduce the results of this work, are available from the following data repository: https://purl.stanford.edu/hs043jz7077.The authors thank Richard Lease for early comments on this manuscript and James Spotila for a careful and insightful journal review that significantly improved the quality of the manuscript. The authors acknowledge support from the U.S. Geological Survey Energy Resources Program. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. government. C.W. Baden and G.E. Hilley acknowledge support from National Science Foundation Career Grant EAR-TECT-105581. D.L. Shuster acknowledges support from the Ann and Gordon Getty Foundation.