Abstract
The western Transverse Ranges are a tectonically active mountain belt in southern California (USA) characterized by fast rates of shortening and rock uplift. Large drainages at the western end of this mountain belt, including the Santa Ynez River and its tributaries, transect regional west–northwest-striking reverse faults and folds. We used fluvial strath terraces within the Santa Ynez River watershed as geomorphic markers for measuring Quaternary rock uplift and deformation across these structures. Mapping, surveying, and numerical dating of these strath terraces in both hanging-wall and footwall blocks of the major reverse faults allow us to separate regional uplift from localized uplift along individual structures.
Luminescence dates from 18 sites within the Santa Ynez River watershed show that the three prominent terrace levels present throughout the area formed between ca. 85 ka and 95 ka, 55 ka and 75 ka, and 30 ka and 45 ka. All three fluvial terrace straths grade into marine paleo-shore platforms along the coast that formed during sea-level highstands. The fluvial straths were formed as a result of lateral erosion during warm, dry climate intervals when vertical incision was temporarily arrested. Incision of the terraces followed during intervening periods of wet climate.
Mapping and valley-long profiles of the terraces document deformation by faults and folds, and we infer minimum rock-uplift rates from the amount of incision below the terrace strath surfaces. Rock-uplift rates range from 0.3 mm/yr to 4.9 mm/yr, with faster rates in the hanging-wall blocks of the major reverse faults and slower rates in the footwall blocks. Rock-uplift rates calculated from strath terraces in the footwall blocks range from 0.3 mm/yr to 1.6 mm/yr, which indicates a regional component of uplift that results from deeper deformation. Higher rates of rock uplift in the hanging-wall blocks (0.5–4.9 mm/yr) are superposed on this regional component. Incremental rock-uplift rates calculated over three time intervals and differences in terrace deformation with age suggest that deformation rates across some structures have decreased over the past 85 k.y.
We conclude that topographic growth of the western Transverse Ranges results from a combination of localized uplift along individual structures that varies both spatially and temporally and a more constant regional uplift that likely results from deeper ductile deformation or slip along detachment faults that have been inferred to underlie the area.
INTRODUCTION
Motivation for the Study
One of the main goals of tectonic geomorphology is to understand the mechanisms that control the topographic expression of active mountain belts. Where active rock uplift is the result of crustal shortening, knowledge of how this crustal shortening translates into uplift is important for understanding how a mountain belt develops. Contractional mountain belts are typically composed of multiple structures that contribute to the overall elevated topography, including reverse or thrust faults that mark the fronts of the mountain belt, as well as faults and folds within the mountain belt (e.g., Jackson et al., 1996; Allmendinger et al., 1997; Thompson et al., 2002; Schmid et al., 2004; Sepehr and Cosgrove, 2004). But there is often a regional component of uplift in contractional mountain belts that cannot be attributed to specific structures seen at the surface (e.g., Ferranti et al., 2007, Calabrian Arc; García et al., 2003, Betic Cordillera; Page et al., 1998, California Coast Ranges). Separating this regional component of uplift from localized uplift along individual structures is necessary to determine the relative influence of different structural processes controlling the development of active mountain belts.
Where an active mountain belt is composed of multiple ranges bounded by faults, how much of the large-scale topography can be attributed to slip on those individual faults? How much uplift is due to deeper structures? How is crustal shortening translated into topographic development across a mountain belt? In this paper, we address these questions by evaluating the causes and patterns of rock uplift in the western Transverse Ranges of California, USA (Fig. 1), where high rates of shortening and uplift have been documented by both geologic and geodetic studies (Namson and Davis, 1988; Donnellan et al., 1993; Huftile and Yeats, 1995; Hager et al., 1999; Marshall et al., 2013; Johnson et al., 2020). We use fluvial strath terraces as markers for rock uplift and deformation over the past 100 k.y. and demonstrate how these markers can be used to separate a regional component of rock uplift from localized uplift and deformation along individual structures.
Shortening and Uplift in the Western Transverse Ranges
The western Transverse Ranges comprise an active fold-and-thrust belt in southern California that deforms Cretaceous through Quaternary sedimentary rocks (Fig. 1). Active uplift of the mountains is evident from flights of marine terraces along the coast, strath terraces along the major rivers within the mountain belt, and the relatively steep topography that mirrors the underlying structure in most places. Geodetic and geologic data from the Ventura area of the western Transverse Ranges show that uplift and shortening rates along the Ventura coast are some of the fastest in North America. Geodetic estimates of shortening rates typically range from 7 mm/yr to 15 mm/yr with an east-to-west decrease in rates (Donnellan et al., 1993; Hager et al., 1999; Marshall et al., 2013; Johnson et al., 2020), while measurements from geologic studies range from 6.5 mm/yr to 25 mm/yr since Miocene time (Yeats, 1983; Huftile and Yeats, 1995; Levy et al., 2019). Geodetic uplift rates average ~2 mm/yr in the interior of the western Transverse Ranges and along the San Andreas fault but show subsidence in the coastal areas (Hammond et al., 2018). This present-day subsidence signal along the coast disagrees with geologic uplift rates of 1–7 mm/yr from marine terraces (Gurrola et al., 2014; Rockwell et al., 2016; Morel, 2018), and Johnson et al. (2020) hypothesized that the geodetic pattern of uplift most likely reflects interseismic flexing above locked faults that accommodate north–south contraction in the western Transverse Ranges.
Most of the geologic work that documented Quaternary uplift rates and fault-slip rates has been concentrated at the southern front of the western Transverse Ranges in the Ventura and Santa Barbara areas (Fig. 1; Huftile and Yeats, 1995; Rockwell, 1988; Rockwell et al., 2016). Recently, rock-uplift rates as high as 2–5 mm/yr have also been documented at the western end of the western Transverse Ranges in the Santa Maria Basin (McGregor and Onderdonk, 2021). In both areas, the rapid geologic uplift rates were measured in the hanging walls of active reverse faults and probably reflect vertical motions on individual structures. Although valuable, these measurements tell us little about the mountain belt-scale uplift processes and may not represent the Quaternary uplift mechanisms and rates of the western Transverse Ranges as a whole. Fluvial strath terraces along rivers within the western Transverse Ranges, however, present a unique opportunity to evaluate the larger-scale topographic development. Strath terraces are present in the elevated hanging walls of the major reverse faults, as well as the footwall valleys, which suggests that deeper processes are causing regional uplift in the western Transverse Ranges.
GEOLOGIC SETTING
Tectonic History of the Western Transverse Ranges
The western Transverse Ranges fold-and-thrust belt trends east–west, cutting transverse to the dominant northwest-striking structural grain of California (Fig. 1). This transverse orientation is the result of two tectonic episodes related to the development of the San Andreas transform system: a Miocene episode of transtension and vertical-axis rotation, and late Miocene to present north–south shortening due to ongoing rotation and the nearby restraining bend in the San Andreas fault.
A Mesozoic to early Cenozoic convergent margin along the western edge of the North America plate transitioned to a transform margin when a spreading center (the East Pacific Rise) encountered the subduction zone around 30 Ma (Atwater and Stock, 1998). As the transform boundary developed, complications in the evolving fault system caused the western Transverse Ranges block to rotate ~90° clockwise while being translated to the northwest along the plate boundary (e.g., Luyendyk et al., 1980; Nicholson et al., 1994). This rotation is recorded in paleomagnetic data from Cretaceous through Oligocene forearc basin rocks exposed in the western Transverse Ranges that were originally deposited in a north–south-oriented basin on accretionary prism and batholithic basement rocks (e.g., Hornafius et al., 1986; Kamerling and Luyendyk, 1979; Onderdonk, 2005). Stratigraphic, structural, and paleomagnetic data from Miocene rocks in the western Transverse Ranges show that rotation was accompanied by normal faulting, basin development, and volcanic activity (Kamerling and Luyendyk, 1979; Crouch and Suppe, 1993). Early Miocene transtension along the southern California plate boundary evolved into transpression and north–south shortening toward the end of Miocene time, most likely related to a change in plate-motion vectors around 8 Ma and the development of the Big Bend of the San Andreas fault (e.g., Atwater and Stock, 1998). Stratigraphic evidence indicates that by Pliocene time, regional shortening and uplift had begun in the western Transverse Ranges and Coast Ranges of California and has continued through Quaternary time (Dibblee, 1982; Page et al., 1998). Many of the present-day reverse faults that deform the western Transverse Ranges are likely re-activated normal faults that formed during earlier Miocene extension (Namson and Davis, 1988; Seeber and Sorlien, 2000; Sorlien et al., 1999).
The Santa Maria Basin
The topography of the western Transverse Ranges is characterized by 1000–2000-m-high mountain ranges with lower-elevation (100–300 m) ridges and hills at the western end of the western Transverse Ranges and along the southern coast (Fig. 1). The western half of the western Transverse Ranges includes a triangularshaped area of hills and valleys called the Santa Maria Basin that lies between the higher Santa Ynez Mountains to the south and the San Rafael Mountains to the northeast (Figs. 1 and 2). This study focused on the Santa Maria Basin because well-preserved Pleistocene deposits and fluvial terraces in the area can be used to document deformation and uplift throughout the area. This is in contrast to the higher elevations of the western Transverse Ranges to the east, where such deposits are rare.
The stratigraphy of the Santa Maria Basin consists of Miocene through Pleistocene sedimentary rocks that were deposited on Mesozoic accretionary wedge rocks known as the Franciscan Complex (Woodring and Bramlette, 1950). The Miocene rocks are mostly biogenic shale, chert, and diatomite deposited in a deep basin that developed during regional transtension (Woodring and Bramlette, 1950; Namson and Davis, 1990; Crouch and Suppe, 1993; Behl et al., 2000). A change to transpression in Pliocene time initiated basin inversion, and the Pliocene sedimentary rocks consist of shale, sandstones, and minor conglomerate that record shallowing and uplift of the basin (Woodring and Bramlette, 1950; Behl and Ingle, 1998). Pleistocene sediments are entirely terrestrial and include the Paso Robles Formation, the Orcutt Formation, and localized terrace deposits in the major drainages. The Paso Robles Formation consists of fluvial sand and gravel that is mostly unlithified to weakly lithified. The Paso Robles Formation name has been applied to similar deposits throughout the southern Coast Ranges, and they are inferred to record Pleistocene uplift of the modern ranges (Page et al., 1998). The Orcutt Formation consists primarily of fluvial sand and gravel that was deposited unconformably on an extensive fluvial strath surface that truncates folded and faulted Paso Robles and older formations (Woodring and Bramlette, 1950). The fluvial deposits of the Orcutt Formation grade into marine terrace deposits near the coast, where the fluvial strath at the base of the Orcutt transitions into a paleo-shore platform (McGregor and Onderdonk, 2021; Woodring and Bramlette, 1950).
The inverted Santa Maria Basin is deformed by a series of folds and reverse faults that includes the Santa Ynez and Santa Ynez River faults that bound the southern edge of the basin and the Little Pine fault that bounds the northeastern edge (Fig. 2; Woodring and Bramlette, 1950; Sylvester and Darrow, 1979; Namson and Davis, 1990). Unconformities within the sequence of sediments in the basin show that shortening and uplift along these structures began as early as late Miocene time and intensified in Pleistocene time (Namson and Davis, 1990; Behl et al., 2000). The major folds within the basin are expressed as linear topographic highs, including the Casmalia Hills, Purisima Hills, and Santa Rita Hills. These folds formed above blind faults after deposition of the Orcutt Formation, which is folded over the elevated topography (Sylvester and Darrow, 1979; Namson and Davis, 1990; Kelty, 2020; McGregor and Onderdonk, 2021). Uplift along these structures caused the fluvial system to localize and incise, creating the current system of drainages that cross the structural grain in some places and parallel it in others.
In this study, we use the fluvial terraces developed along the Santa Ynez River and its major tributaries to evaluate late Quaternary deformation and uplift. The Santa Ynez River drains ~2000 km2 of the western Transverse Ranges, the southern Coast Ranges, and the Santa Maria Basin. The river flows westward along the northern edge of the Santa Ynez Range and empties into the Pacific Ocean near the town of Lompoc (Figs. 1 and 2). The drainage basin is asymmetric with a majority of the basin area to the north of the Santa Ynez River. All of the major tributaries of the Santa Ynez River within the Santa Maria Basin flow southward and have multiple levels of fluvial strath terraces. The modern Santa Ynez River, and its major tributaries that we evaluated in this study, have mixed bedrock-alluvial channels and are currently cutting laterally into bedrock in many places.
METHODS
Terrace Mapping and Descriptions
We mapped fluvial terraces along the Santa Ynez River and three of its tributaries: Zaca Creek, Santa Aqueda Creek, and Santa Cruz Creek (Figs. 2–5). We focused our efforts on three strath-terrace levels that are laterally continuous along the drainages and can be correlated among drainages based on the tread elevation above the active channels, and with luminescence ages. We did not completely map other less well-defined terraces along the drainages. These other terraces include: (1) older terrace fragments that are sparsely preserved at higher elevations in some places, (2) terraces that are only locally developed at the mouths of smaller tributaries, and (3) low terraces that are a few meters or less above the active channel and probably not above 500-year flood levels.
Terrace identification and mapping was done both remotely and in the field. Remote mapping was done first using digital elevation models (DEM; 1-m U.S. Geological Survey [USGS] LiDAR data and 3-m National Oceanographic and Atmospheric Administration [NOAA] interferometric synthetic aperture radar data), Google Earth, and historic air photos. We then checked and refined our maps in the field. The composition, stratigraphy, thickness, and weathering of the terrace deposits were also recorded in the field. These data are presented in detail in several California State University, Long Beach, masters theses (Farris, 2017; Kelty, 2020; McGregor, 2019; Tyler, 2013). We did not use soil development for terrace correlations because the source rock, and hence weathering style and rate, vary for each drainage, and because Farris (2017) showed that soil profiles do not correlate with numeric terrace ages in this area. We named the terraces using increasing numbers for increasing age (and height above the active channel), such that Qt1 is our youngest and lowest strath terrace. We adopted this convention because this study only deals with the prominent strath terraces that are correlative across the area, of which Qt1 is the youngest, and there are additional higher strath terraces in some areas that we did not map in detail or numerically date that could be included and evaluated in the future.
Terrace Elevation Profiles
We constructed profiles of the terrace surfaces along each drainage to aid in correlations and to document any deformation (Figs. 6 and 7). The profiles were made parallel to the average trend of the river valleys on parts of the terrace treads that are relatively well preserved and are not significantly altered by erosional, depositional, or anthropogenic features. We noted the thickness of the terrace deposits where the terrace straths were not hidden by soil cover or landslides.
The altitudes of most terraces were surveyed in the field using a Trimble GeoXH GPS receiver, which resulted in elevation data with a horizontal and vertical accuracy of between 10 cm and 75 cm after post-processing. For areas not accessible due to property restrictions, we used a USGS LIDAR 1 m DEM to produce profiles of fluvial terrace treads. GPS-derived elevations agreed with the LIDAR 1 m DEM within 0.5 m or less. The active channels of Zaca Creek and Santa Cruz Creek were also surveyed in the field with the Trimble GeoXH, but we used DEM data to generate a profile of the active channel of the Santa Ynez River because it is inaccessible in most locations due to thick brush and/or private property.
Luminescence Dating
We used luminescence dating to determine the numeric ages of the prominent terrace levels in each drainage. Luminescence dating measures the amount of time since quartz or feldspar grains in a sedimentary deposit were last exposed to sunlight (e.g., Aitken, 1998; Rittenour, 2008). An age is determined based on measurements of the rate at which the grains accumulate radiation from the surrounding environment (dose rate) and the amount of radiation accumulated in the grains since they were deposited and buried (equivalent dose). This method is well-suited for dating late Quaternary fluvial terraces because the age range extends to ca. 300 ka, and the sand and silt-sized particles needed for luminescence dating are common in sedimentary deposits (e.g., Rittenour, 2008).
We collected and dated 21 samples from 18 different sites. Multiple samples were dated at three sites to check for repeatability of the dates. Samples were collected at depths of 1–3 m below the terrace treads. These depths are close to the terrace surface to best approximate the age of the terrace tread, while still being deep enough to prevent light exposure or anthropogenic surface disruption that could affect the equivalent dose measurements of the samples. Samples were collected by removing the outer 40–50 cm of outcrop, driving a metal pipe (capped on the outer end) into the exposure wall, excavating around the pipe, and then capping the inner end while under a tarp to prevent light contamination.
Equivalent dose measurements were made at the California State University, Long Beach, Luminescence Lab. We analyzed feldspar grains using the post-infrared–infrared stimulated luminescence (pIR-IRSL) protocol of Rhodes (2015) on polymineral multi-grain aliquots (typically 20–24) from each sample. Samples were sieved and subjected to liquid settling to isolate the 1–8 µm fraction and were treated with HCl and H202 to remove carbonates and organics before settling the sample to 26 disks (aliquots) for analysis. Sample aliquots were analyzed in a Risø thermoluminescence (TL)/optically stimulated luminescence (OSL) reader with a BG-39 filter to isolate the feldspar signal. The pIR-IRSL protocol on feldspars was used instead of OSL on quartz because several studies have shown that quartz can have poor luminescence characteristics in young orogens (e.g., Rhodes, 2015; Trauerstein et al., 2014), and our previous attempts to date terraces along the upper Santa Ynez River with OSL resulted in unreasonably young ages. The pIR-IRSL protocol also corrects for the anomalous fading problem that affects traditional IRSL analysis by adding lowtemperature bleach steps prior to each measurement of the luminescence signal (Rhodes, 2015). Probability density functions of aliquot data were produced and evaluated to identify evidence of incomplete bleaching or bioturbation that could result in incorrect age calculations. The equivalent dose for each sample was then calculated using the central age model derived from radial plots of the aliquot data at 1 sigma uncertainty (Vermeesch, 2009; Galbraith and Roberts, 2012). Probability density functions and radial plots for each sample are presented in File A of the Supplemental Material1.
Dose rates for each sample were calculated using the concentrations of radioactive isotopes in the deposit, the cosmic radiation contribution, and the full range of possible moisture content. Sediment samples were collected from each sample site, and concentrations of potassium, uranium, and thorium in these samples were determined by gamma spectrometry at the USGS Luminescence Lab in Denver, Colorado, USA. The cosmic radiation contribution is based on the sample elevation, depth, and latitude (Prescott and Hutton, 1994), and in-situ water content was determined in the lab. In-situ water contents were typically 5–10%, but to account for the full range of possible water content over the lifetime of the deposits we used 2% and 20% as minimum and maximum values, respectively, to calculate dose rate uncertainties for each sample. We note that this large allowance for variable water content history results in larger error ranges for our terrace ages (3–5 times larger) than if we used the in-situ moisture contents measured, but this is a critical uncertainty that is frequently neglected in published luminescence ages (Nelson and Rittenour, 2015).
RESULTS
Terrace Mapping and Descriptions
Terrace mapping shows that three prominent terrace levels correlate among the major drainages within the Santa Ynez River watershed. These include a lower terrace (Qt1) whose surface is typically 12–15 m above the active channel, an intermediate terrace (Qt2) at 35–50 m above the active channel, and a higher terrace (Qt3) at 60–75 m above the active channel (Figs. 2–5). These heights above channel values are based on the least-deformed sections of the drainages. Terrace elevations can be significantly higher than these averages in places where the terraces are deformed by faults and folds (described in Terrace Deformation and Fault-Slip Rates section below).
Terrace deposits typically vary from 2 m to 8 m thick and are deposited on straths that truncate Eocene to middle Pleistocene sedimentary rocks. There are several locations, however, where terrace deposits are considerably thicker, and the contact along which terrace deposits overlie Eocene and Pleistocene sedimentary rocks is poorly exposed. These relatively thick terrace deposits exist in two separate areas along the lower Santa Ynez River, where it flows through a syncline in the footwall of the Santa Ynez River fault west of Solvang and north of Lompoc (Fig. 2). The terrace deposits in this area are at least 15 m thick in the lower two terraces (Qt1 and Qt2) and at least 45 m thick in the higher terrace (Qt3; Fig. 6). We attribute the relative thickness of the terrace deposits to be the result of localized, tectonically driven subsidence in the footwall of the Santa Ynez River fault during deposition of the terrace deposits. A similar relationship is present along Santa Agueda Creek, where the Qt2 terrace is 18 m thick in the footwall on the north side of the Baseline fault but only 5 m thick in the hanging wall on the south side (Fig. 4).
Terrace deposits along the Santa Ynez River and Santa Cruz Creek are composed of interbedded gravel, sand, and silt. Gravel clasts are primarily well-rounded sandstone derived from Eocene formations but also include shale clasts from the Miocene Monterey Formation, various igneous rocks recycled from Oligocene, Eocene, and Cretaceous conglomerates, and chert from the Mesozoic Franciscan Complex. In general, terrace exposures show basal gravel layers, with clasts of up to 1 m in diameter that are overlain by sand and silt. Terrace deposits in Zaca Creek differ from the other two drainages in that they are composed almost entirely of Monterey Formation shale clasts that do not exceed ~10 cm in diameter. These clasts are derived from the upper reaches of Zaca Creek, which is underlain entirely by the Monterey Formation, and from the Pleistocene Paso Robles Formation that underlies the terrace straths. The Paso Robles Formation is mostly conglomerate that is also comprised primarily of Monterey Formation clasts with smaller amounts of chert from the Franciscan Complex.
In addition to these three dominant terrace levels, a lower terrace level is also present between 1 m and 5 m above the active channel in the drainages studied (Qt0 in Figs. 3–5). Because the variation in tread height above the active channel is high relative to the average height, it is hard to confidently correlate this terrace level along the drainages. The low heights also imply that this terrace may still be flooded in some areas during 100- or 500-year flood stages. Therefore, we mapped only the larger expressions of this terrace level that we could tentatively correlate and label as Qt0. These low terraces are fill terraces in some places and strath terraces in others. Higher terraces are also present in some locations along the upper Santa Ynez River and Santa Cruz Creek at elevations greater than 80 m above the active channel (Qt4 in Figs. 3 and 4). Although we include some of these higher terraces in our mapping, we did not date or use them in our measurements of uplift and deformation.
Our mapping of Qt3 includes several incised alluvial fans present along the north side of the Santa Ynez Mountains south of Lake Cachuma. The surfaces of these fans grade into fluvial terrace treads along the south side of the Santa Ynez River. We include these fans in our mapping because we infer that their upper surfaces are age-equivalent to the fluvial terrace surfaces into which they grade (for example, García et al., 2003; Garcia and Mahan, 2009). The fans are also faulted and folded in some places, which makes them useful for documenting deformation and uplift in the area.
Active Channel and Terrace Tread Profiles
Longitudinal (“long”) profiles of the active channels mapped in this study are generally concaveupward, but their shapes vary with drainage size. The Santa Ynez River is the main trunk draining much of the southern Santa Maria Basin and the mountain ranges to the northeast and south (Figs. 1 and 2). The modern Santa Ynez River is a mixed bedrock-alluvial channel and is cutting laterally into bedrock along its entire length. Although the bedrock is covered in many places, we infer that this cover is less than a few meters thick due to the presence of bedrock in some parts of the thalweg and the exposures of bedrock near the modern channel level. The Santa Ynez River profile is smooth and concaveupward (Fig. 8A) with a concavity index (ø) of 0.45, which is expected for a river that is in equilibrium (e.g., Wobus et al., 2006; Kirby and Whipple, 2012). Profiles of the terrace treads parallel the active channel, except where the terraces are locally deformed by folds (Fig. 6).
Santa Cruz Creek, the largest tributary evaluated in this study, is also a mixed bedrockalluvial channel with bedrock exposed where the active channel is cutting into the sides of the canyon (as opposed to the low fill terraces). Santa Cruz Creek has an overall concave long profile (ø = 0.5) but a conspicuous knickpoint exists ~250 m upstream of the Little Pine fault (Fig. 8D). This knickpoint coincides with a large landslide complex within metasedimentary rocks of the Jurassic Franciscan Complex. Strath terraces along Santa Cruz Creek are only present in the lower reach of the drainage, where the active channel profile has a smooth, concave shape. The average slopes of the terrace treads are slightly steeper than the slope of the active channel, which suggests that there may be a slight southward tilting in addition to the obvious folding and faulting of the older terraces (Fig. 7).
Zaca Creek and Santa Agueda Creek are also mixed bedrockalluvial channels with bedrock exposed in the active channel cutbanks. Their long profiles are less concave than those of the larger drainages (Figs. 8B–8C). Both Zaca Creek and Santa Agueda Creek have a concavity index of 0.22, which is lower than the generally accepted 0.45 concavity index for streams in equilibrium (0.4–0.6). These lower concavities are due to slightly convex reaches where the drainages cross the Baseline–Los Alamos fault. These convexities suggest active or recent fault slip, but the lowest inset fill terraces within these reaches are undeformed where they cross the fault. We infer that these convexities are instead due to a marked difference in rock strength along these drainages, with less resistant Pleistocene sediments of the Paso Robles Formation upstream of the Baseline–Los Alamos fault, and more resistant Miocene rocks in the hanging wall downstream of the fault. Profiles of the older terraces in Zaca Creek are noticeably steeper than those of younger terraces and the active channel (Fig. 7) and suggest a southward tilt of the footwall between the Little Pine and Los Alamos faults.
Terrace Ages
Our pIR-IRSL dates provide numeric ages for the three prominent terrace levels in the study area. Ages for each terrace level vary slightly between drainages, but representative ages for the three terrace levels are 85–95 ka for Qt3, 55–75 ka for Qt2, and 35–45 ka for Qt1 (Table 1, Fig. 9). Aliquot distributions for most of the samples were tightly grouped and did not show evidence of partial bleaching or bioturbation (see File A, footnote 1), and we used all of the measured aliquots to determine sample ages for those samples. Four samples, however, contained several obvious outliers in the aliquot distribution, and those aliquots were excluded in the determination of the central age from the radial plot (see Cachuma Camp samples, Highway 154, and Santa Agueda Qt3 in Table 1 and File A). Two samples exhibited distinct peaks in their aliquot distributions that required some interpretation. Using the full aliquot range for our sample from Qt3 in Santa Cruz Creek resulted in an unreasonably old age based on mapping and survey correlations with other terraces dated in the area (Fig. 3, Table 1). We, therefore, interpret that most of the grains in the sample were partially bleached, and we prefer the youngest peak in the aliquot probability density function as the closest representation of the true age (Fig. S8 in File A, see footnote 1; Table 1). One sample from Qt3 in Zaca Creek (sample: Zaca Mesa-A2) exhibited two separate peaks in the distribution of aliquots, and we interpreted that the older peak was likely more representative of the true age of the sample because it resulted in an age that agreed with the other sample from the same site (sample: Zaca Mesa-A1) and fit better with the ages and elevations of the other dated terraces in Zaca Creek.
Dose rates varied from 3.5 Gy/ka to 5.8 Gy/ka, with most between 4.5 Gy/ka and 5.5 Gy/ka. At sites where we dated two samples, the sample ages varied by 10% or less and the full age ranges overlapped, which indicates repeatability in the age determinations.
Quaternary Terrace 3 (Qt3)
Dates from Qt3 samples varied from 79 ± 7 ka to 125 ± 10 ka (Table 1, Fig. 9). This time span corresponds with deposition of the Orcutt Formation between 120 ka and 80 ka (McGregor and Onderdonk, 2021). The Orcutt Formation is a regionally extensive fluvial deposit that overlies a planation surface that once extended across most of the Santa Maria Basin and grades into a paleo-shore platform at the coast (McGregor and Onderdonk, 2021; Woodring and Bramlette, 1950). The highest fluvial terraces along the lower Santa Ynez River, and the highest marine terrace that constitutes the extensive geomorphic surface north of the mouth of the Santa Ynez River, were previously correlated to the Orcutt Formation and dated with pIR-IRSL to be 93.5 ± 7.5 ka and 85 ± 6 ka, respectively (Kelty, 2020). It is important to reiterate that the Qt3 fluvial terrace strath along the lower Santa Ynez River can be directly traced into the marine isotope stage (MIS) 5a paleo-shore platform at the coast, which shows that the Qt3 terrace graded to sea level during deposition (for example, Leopold and Bull, 1979). On the basis of landscape position and geochronology, Qt3 terrace surfaces along the upper and middle Santa Ynez River correlate downstream with the Orcutt terrace deposits along the lower Santa Ynez River. The Qt3 surfaces in Zaca Creek project southwestward into Orcutt deposits along the north side of the Los Alamos fault (Fig. 2). These ages and projections lead us to interpret that the Qt3 terraces we mapped are part of the regional Orcutt Formation, as are the fan deposits in the Lake Cachuma area that grade into Qt3 terraces. We note that the age of the Orcutt Formation (and Qt3) corresponds to the MIS 5a–e highstands and is equivalent to two extensively preserved marine terrace levels along the central coast of California that are ca. 80 ka and 120 ka in age (Hanson et al., 1992; Lettis et al., 2004).
Two of the Qt3 samples we collected were from deposits that cannot be directly related to the Santa Ynez River (Santa Agueda Qt3 and Circle JB Polo Ranch, Table 1, Fig. 4). These samples also returned ages that were closer to MIS 5e, whereas samples from Qt3 terraces along the Santa Ynez River, Santa Cruz Creek, and Zaca Creek are closer to MIS 5a. We infer that these two older Qt3 samples are from earlier deposition of the Orcutt Formation, as it was deposited across much of the Santa Maria Basin before the modern drainage system began to form. We do not include these two samples in our measurements of incision.
Quaternary Terrace 2 (Qt2)
Along the upper and middle Santa Ynez River, Qt2 sample ages vary from 61 ± 6–76 ± 7 ka (Table 1). These ages are similar to a pIR-IRSL date of 60.5 ± 4.5 ka from the same terrace level along the lower Santa Ynez River (Kelty, 2020). The Qt2 terrace can be traced to the coast (interpolated between isolated remnants), where it merges with either the second or third emergent marine terrace north of the Santa Ynez River mouth (Kelty, 2020). This suggests that, like the Qt3 terrace, it is directly tied to a paleo-shore platform that was carved during sea-level rise in an interval of warming climate.
In Zaca Creek, three samples in Qt2 deposits returned younger ages, between 49 ± 5 ka and 59 ± 6 ka (Fig. 5). Projection of the Qt2 surface along Zaca Creek to the Qt2 surface along the Santa Ynez River makes it possible to correlate this terrace level between the two drainages. We, therefore, hypothesize that the younger ages are likely the result of delayed incision and abandonment of the Qt2 surface in Zaca Creek relative to incision and abandonment of the Qt2 surface along the Santa Ynez River. The mechanism that led to delayed abandonment of the Qt2 surface along Zaca Creek is stream antecedence within the Santa Ynez River drainage network. Zaca Creek is an antecedent stream that flows through the eastern Purisima Hills, which began to grow shortly after deposition of the Orcutt Formation/Qt3 terrace alluvium (McGregor and Onderdonk, 2021). There is a marked difference in erodibility of the rocks underlying the upper and lower reaches of Zaca Creek. The lower reach of Zaca Creek incises Miocene sandstone, chert, and shale, while the upper reach cuts down into largely unconsolidated Pleistocene conglomerate, sand, and silt of the Paso Robles Formation. We hypothesize that the increased time required for incision of the lower Zaca Creek though the rising topography and harder bedrock likely led to delayed abandonment and incision of the Qt2 terraces in the upper Zaca Creek part of the drainage basin.
Quaternary Terrace 1 (Qt1)
We sampled Qt1 sediments in two locations along the Santa Ynez River; samples from the uppermost reach yielded dates of 43 ± 4 ka and 46 ± 4 ka (Fig. 3), and a sample from the middle Santa Ynez River yielded a date of 32 ± 3 ka (Fig. 4, Table 1). Near the mouth of the Santa Ynez River, Qt1 can be traced into the lowest emergent marine terrace at the coast, which was dated to 41 ± 6 ka (Kelty, 2020). This ca. 40 ka marine terrace has also been mapped and dated at several locations along the coast south of the Santa Ynez River to Point Conception (Tutterow, 2021), and east of Point Conception in the Gaviota area (Morel, 2018; Morel et al., 2022). The ages of these marine terraces, and the Qt1 terraces, correspond with the MIS 3a highstand.
Incision and Rock-Uplift Rates
We calculated incision rates at each location where we dated a terrace (Fig. 2, Table 2) using the terrace ages and amount of incision represented by the difference in elevation between the terrace strath surface and the adjacent active channel. We infer that long-term incision of the Santa Ynez River and Santa Cruz Creek has kept pace with rock uplift such that the incision we measured is equal to rock uplift. This interpretation is supported by the following observations: (1) the active channel of the Santa Ynez River has a smooth and concave upward longitudinal profile (Fig. 8); (2) the Santa Ynez River has been generating strath terraces since at least Qt3 time; (3) the slope of the active channel is approximately the same as the adjacent terrace profiles where those terraces are not significantly deformed by intersecting structures, which is consistent with a stream that is in steady-state and incising at the same rate as rock uplift (Pazzaglia and Brandon, 2001); and (4) much of the drainage basin of the Santa Ynez River is formed in relatively weak rocks that are easily denuded. Streams draining relatively large catchment areas that formed in weak rocks, and that have strath terraces adjacent to the channel and an upwardly concave longitudinal profile, are able to incise vertically at the same rate as rock uplift (for example, Pazzaglia et al., 1998; Pazzaglia and Brandon, 2001). The premise of this model is that if the stream is able to maintain an upwardly concave equilibrium profile and also erode laterally as it cuts straths into bedrock, then there is sufficient stream power for the stream to incise at the same rate as rock uplift.
As mentioned previously, the long profiles of Zaca Creek and Santa Agueda Creek are slightly less concave (ø = 0.22) and have straight and/or convex reaches. This suggests that incision of these channels may not be keeping pace with rock uplift due to more resistant rocks in the hanging wall of the Baseline–Los Alamos fault at the lowermost reaches of these drainages. Zaca Creek and Santa Agueda Creek also have smaller drainage basin areas (62 km2 and 85 km2, respectively) than the larger Santa Cruz Creek drainage area (~200 km2) and that of the Santa Ynez River (~2000 km2). The reduced discharge (and stream power) from the smaller drainage basin size may result in slower incision rates (for example, Merritts and Vincent, 1989). Consequently, our uplift rates from these two drainages should be regarded as minimum rates.
Our time-averaged incision rates vary between 0.23 mm/yr and 0.95 mm/yr (Table 2). Figure 2 shows the spatial distribution of rock-uplift rates calculated in this study, as well as those determined from previous studies in the area (Morel, 2018; Kelty, 2020; McGregor and Onderdonk, 2021; Tutterow, 2021). Rockuplift rates are higher on average (0.7–4.9 mm/yr) at sites in the elevated hanging walls of the reverse faults, including anticlines that formed over blind thrust faults in the western part of the basin (McGregor and Onderdonk, 2021). Rock-uplift rates are lower in the footwalls of the major faults, with the lowest rates from footwall sites in a syncline that parallels the Santa Ynez River fault along the middle Santa Ynez River (Fig. 2).
Calculating incision (and rock uplift) rates based on the modern channel elevation may not accurately represent the long-term rates (Wegmann and Pazzaglia, 2002; Finnegan et al., 2014; Gallen et al., 2015). Incision of the Santa Ynez River is likely not constant over shorter periods of time due to climatic variations, and the modern channels of the river and its tributaries may currently be higher than average due to late Holocene warming and drying that temporarily arrest incision. So we also calculated incremental incision rates for three time periods defined by the ages of the three main terrace levels (Fig. 2, Table 3) to provide a more complete picture of uplift patterns though time and to identify and account for any bias in incision rate due to use of the active channels as a reference elevation. We used the difference in terrace ages and strath elevations to calculate incision rates from ca. 85 ka to 65 ka (Qt3 to Qt2) and from ca. 65 ka to 40 ka (Qt2 to Qt1). The active channel elevation was used to calculate a rate from ca. 40 ka to present (since Qt1). Although the errors associated with these incremental rates are high for some locations (Table 3), this approach suggests that there were variations in uplift rates during the past 100 k.y. In the hanging-wall blocks of the Santa Ynez River fault and the Baseline fault, incision rates appear to be faster (0.86–3.4 mm/yr) between 85 ka and 40 ka, with rates slowing down during the youngest time interval. Incision rates in Santa Cruz Creek are higher (0.9 mm/yr) between 85 ka and 65 ka and have decreased since 65 ka. In all of these locations, profiles of the fluvial terraces also show a greater degree of deformation of the older terraces and no measurable deformation of the younger terraces (Figs. 6 and 7). Conversely, no significant decrease in incision rates or degree of terrace deformation is observed in the footwalls of the active faults.
To test whether the slower incision rates we calculate for the younger terraces is due to a real change in rock-uplift rate through time, or to bias introduced when using the modern channel as a reference, we evaluated how the modern stream channel compares to its long-term average. If the modern stream channel is currently aggrading or eroding laterally, it may be higher than the long-term average and result in bias toward lower incision rates for younger terraces.
We followed the approach of Gallen et al. (2015) and compared terrace ages to cumulative incision (strath elevations above the active channel or the lowest strath) on log-log plots to determine the power-law relationship between cumulative incision and terrace formation intervals (File B, see footnote 1). The power-law exponent of a line fit to the data will be 1 if the modern channel is at its long-term average, less than 1 if it is below, and more than 1 if it is above (Gallen et al., 2015). This enables one to account for the dynamic reference frame of strathterrace formation resulting from unsteady incision of a river channel. The power-law exponent for the terraces along the Santa Ynez River is 1.9 for terraces in the upper Lake Cachuma area, 2.4 for terraces in the Santa Rita Hills, and 2.1 for terraces along the entire Santa Ynez River (Figs. B1, B3, and B6, see footnote 1), which indicates that the modern channel is higher than its long-term average. When we do the same analysis using the youngest strath terrace (Qt1) as a datum to remove the bias of the modern channel, the power-law exponent is even higher (Figs. B2 and B4). This implies that uplift rates have indeed changed and that incision of the Santa Ynez River has slowed since deposition of the youngest terrace ~40 k.y. ago. In addition, a log-log plot of the three terrace levels in the Solvang area, which is in the footwall of the Santa Ynez River fault that was subsiding for much of the past 85 k.y., results in a power-law exponent less than 1 (Fig. B5). This indicates that the river in the footwall block is incising faster than the 85 k.y. average. The incremental uplift rates from the Solvang area also show an increase in uplift rate during the younger time increments (Fig. 2). We interpret these results to indicate that the uplift rate of the hanging-wall blocks, and correlative subsidence of the footwall blocks, has decreased through time such that the active channel is now incising slower than average in the hanging-wall blocks, and faster than average in the footwall blocks.
Terrace Deformation and Fault-Slip Rates
All of the major faults and folds in the southeastern Santa Maria Basin deform terraces along the primary drainages, which shows that these structures have been active in the past 100 k.y. Here, we describe deformation of the terraces and calculate fault-slip rates where possible.
Displacement and Slip Rate across the Santa Ynez Fault
In the upper reaches of the Santa Ynez River, all three terrace levels are vertically displaced by the Santa Ynez fault. At the Manzanita Road site, the Qt1 terrace is displaced by two strands of the Santa Ynez fault, both of which are expressed as scarps on the terrace tread and are exposed in a stream cut at the eastern edge of the terrace (Figs. 10–12). The terrace surface is dropped down between the fault strands, forming a 50–70-m-wide sag in the terrace tread. Where it is exposed at the eastern extent of the terrace, the southern fault strand is expressed as a 10–40-cm-wide gouge zone that dips 40° ± 2° to the south and places Eocene sandstone over Qt1 terrace deposits (Fig. 11). The terrace strath is vertically displaced 12–14 m across the fault, which corresponds to 19.5–20.9 m of dip-slip displacement. We dated two samples from the upper silty terrace deposits in the footwall of the fault, with the youngest sample returning an age of 43 ± 4 ka. Using this age, we calculate a dip-slip rate of 0.48 ± 0.06 mm/yr (Table 4).
The same stream cut exposes the northern fault strand, which is expressed as multiple gouge zones and striated surfaces that dip 75–85° north and separate Oligocene mudstones on the south from Miocene shale on the north (Fig. 12). The zone of deformation is ~3 m wide and is truncated upward by the Qt1 terrace gravels. One fault surface extends upward into the terrace deposits, however, and the terrace gravels are dropped down on the south side by 1.5 ± 0.1 m. Sub-horizontal striations on the numerous secondary fault surfaces suggest strike-slip along the fault. There is a possible ~220 m leftlateral separation of a stream crossing both strands (Fig. 12), but this hypothesis requires further evaluation and possibly subsurface investigation, which was not done during this study.
Where the Santa Ynez fault crosses Hot Spring Canyon 5 km west of the Manzanita Road site (Fig. 3), the Qt2 and Qt3 terraces are vertically displaced across the fault. Neither the fault nor the terrace strath surfaces are well exposed here due to dense vegetation and soil cover. In addition, the Qt3 terrace surface on the steeper hanging-wall block of the fault is incised by small drainages and has been distorted by landslides. Therefore, at this site we can only make tentative estimates of the amount of vertical displacement of the Qt2 terrace based on vertical separation of the terrace tread projected across the fault and an estimated fault dip of 40° ± 10° south (File C, see footnote 1) based on the observed fault dip of 40° at the Manzanita Road site. The resultant dip-slip rate is 0.31 ± 0.11 mm/yr since 72 ± 6 ka (Table 4). We emphasize again that this is a tentative estimate of fault-slip rate based solely on inferred fault dip and apparent separation of the terrace treads. The error ranges from the dip-slip rate from this site and the Manzanita site overlap around 0.42 mm/yr and suggest that the dip-slip rate across the Santa Ynez fault has not changed significantly over the past ~73 k.y. We note that these rates are also similar to the 0.44 mm/yr dip-slip rate reported by Morel (2018) based on a 40 ka marine terrace that is vertically displaced by the western end of the Santa Ynez fault near Gaviota (Fig. 2).
Displacement and Slip Rates across the Eastern Baseline and Little Pine Faults
Both the Baseline and Little Pine faults deform terraces in Santa Cruz Creek, but the fault surfaces are not clearly exposed. Again, we can only make tentative slip-rate estimates for these faults using our terrace ages, vertical separation of the terrace treads, and fault-dip angles measured at other locations along the fault. The Baseline fault is a southdipping reverse fault that vertically displaces the Qt3 and Qt2 terraces. The fault dips 30–40° south, where it was exposed in trenches near Zaca Creek (Guptil et al., 1980), and we assume that a slightly steeper dip is possible at Santa Cruz Creek based on 3-D projection of the inferred location of the fault at various elevations closer to Santa Cruz Creek. We therefore infer a fault dip of 30–50°. We used projections of the terrace treads for the Qt3 and Qt2 terraces across the fault to measure vertical separation and dip-slip displacement (File C, see footnote 1). The resulting dip-slip rates are 0.17 ± 0.08 mm/yr since 107 ± 8 ka (Qt3), and 0.07 ± 0.04 mm/yr since 76 ± 6 ka (Qt2, Table 4). The Qt3 terrace is also vertically displaced by the Baseline fault to the west, where this terrace is present on the west side of Santa Agueda Creek, but the strath of this terrace is not well exposed, and the tread of this terrace has been significantly eroded, so an estimate of displacement is not possible. However, the tread of Qt2 is also vertically displaced 7 ± 1 m by the Baseline fault in Santa Agueda Creek. We dated Qt2 in Santa Agueda Creek both upstream and downstream of its intersection with the Baseline fault (Fig. 4, Table 1). The downstream site is located in the elevated hanging wall of the fault, where the terrace deposits are ~5 m thick and a clear terrace strath is exposed. Our pIR-IRSL date here was collected 3 m above the terrace strath and was 72 ± 6 ka. The upstream site is located in the footwall of the fault, where the terrace deposits are considerably thicker (18 m), and the terrace strath is only ~4 m above the active channel (Table 2). Our pIR-IRSL date here was collected near the top of this deposit and was 47 ± 4 ka. The thicker terrace deposits in the footwall of the fault, and the younger age from the footwall site, suggest that the fault was active during terrace formation and that sedimentation continued upstream of a scarp that developed along the Baseline fault between 72 ka and 47 ka. Projecting the terrace tread across the fault and using 47 ± 4 ka as an approximate age of the Qt2 tread, we estimate a dip-slip rate of 0.22 ± 0.11 mm/yr since 47 ± 4 ka. This is slightly higher than the rate calculated using the offset of Qt2 in Santa Cruz Creek, but these calculations based on profiles of the terrace treads should be regarded as speculative and only used to infer that the fault has slipped in the latest Quaternary, but at a slow rate.
The Little Pine fault is a north-dipping reverse fault that vertically displaces the Qt2 terrace by 6.5 ± 1 m in Santa Cruz Creek. We infer a dip of 35–55° based on nearby measurements (Dibblee and Minch, 2005) and calculate a slip rate of 0.13 ± 0.05 mm/yr since 76 ± 6 ka (Table 4; File C, see footnote 1). The Qt3 terrace is not preserved in the hanging wall of the Little Pine fault, and the Qt1 terrace is not present in Santa Cruz Creek, so we have no way to assess the amount or rate of slip over any other time periods.
The two older terraces (Qt4 and Qt3) in Santa Cruz Creek are also folded across the San Marcos anticline (Fig. 7). The Qt2 terrace does not show any deformation across the fold. The slightly steeper slope of the terraces compared to the active channel, however, suggests that there may be ongoing regional deformation in this area in the form of southward tilting.
Folding in the Middle and Lower Santa Ynez River
Along the Santa Ynez River, the Qt3 and Qt2 terraces are folded where they approach or cross the major faults. In the middle Santa Ynez River, between Lake Cachuma and the town of Solvang, the Qt3 and Qt2 terraces are lifted in the hanging wall of the Baseline fault, where they are in close proximity to the fault (Figs. 2 and 6). Farther downstream, the terraces are lifted and broadly folded along a 10 km stretch of the river where it passes through the Santa Rita Hills (Fig. 2) in the hanging wall of the blind Santa Ynez River fault (Kelty, 2020). Conversely, the terrace straths are dropped down in the footwall of the Santa Ynez River fault, where the river parallels the fault downstream of Solvang. The terrace straths are also below the present-day active channel in the Lompoc area farther downstream. Both of these locations lie in a large syncline that is present in the footwall of the Santa Ynez River fault between Solvang and Lompoc (Fig. 2). We infer that this syncline developed at the same time as the adjacent Santa Ynez River fault and the anticline in its hanging wall, and it folded the terraces down in the footwall as they were being lifted in the hanging wall. Down warping of the Qt3 and Qt2 terrace treads show that the syncline continued to subside after terrace formation. The Qt1 terrace, however, is not deformed where it crosses the Santa Ynez River fault or the Baseline fault, which indicates that these two faults have not been active in the past ~40 k.y.
Zaca Creek, Purisima Hills, and Casmalia Hills
Terrace profiles in upper Zaca Creek show no evidence of deformation (Fig. 6). We note, however, that the Qt3 and Qt2 terraces in Zaca Creek have slightly steeper slopes than the Qt1 terrace and the active channel, which may be due to southward tilting resulting from slip on the Little Pine and/or Los Alamos faults at either end of the mapped reach. None of the terraces continue into the hanging wall of the Little Pine fault in Zaca Creek, so we are not able to test for late Quaternary displacement on the fault at this locality. In lower Zaca Creek, where the creek crosses the Los Alamos fault and cuts through the Purisima Hills, the lower terraces (Qt2, Qt1, and Qt0) are not well-preserved, which makes mapping and correlation tentative. A few kilometers to the west, however, the Qt3 terrace is truncated by the Los Alamos fault, which bounds the northern side of the eastern Purisima Hills. The Qt3-equivalent Orcutt Formation is deformed across both the Purisima Hills and Casmalia Hills, which shows that these hills developed after deposition of the Orcutt Formation (McGregor and Onderdonk, 2021).
DISCUSSION
Our results reveal two interesting aspects of topographic growth in the western Transverse Ranges. First, rock uplift is occurring in both the footwall and hangingwall blocks of the major reverse faults, which shows that there is a regional component of uplift in addition to localized uplift along individual structures. Second, apparent changes in incremental incision rates through time suggest that rock uplift may not have been constant over the past 100 k.y., which could be the result of changes in fault-slip rates through time. In this section, we first present a model for the genesis of terraces that we use as markers for deformation and uplift. We then describe these regional and local components of uplift and the possible variations in uplift and deformation during the past 100 k.y. We finish by evaluating possible causes and implications for regional tectonics and mountainbuilding processes in general.
Models of Fluvial Terrace Genesis and Stream-Longitudinal Profile Development
Coastal River Response to Sea-Level Fluctuation
To use terraces to calculate incision rates driven by tectonism, the influence of sea-level fluctuation on stream incision and terrace production must be assessed. A fundamental factor controlling stream-channel response to sea-level fluctuation is the gradient of the continental shelf adjacent to the river mouth (Pazzaglia, 2022, and references therein). If the gradient of the continental shelf is relatively low, particularly on shelves adjacent to areas like the Western Transverse Ranges, where climate and topography ensure abundant sediment supply to streams (Langbein and Schumm, 1958; Gray et al., 2018), then sea-level fall leads to sedimentation on the subaerially exposed shelf rather than stream incision (Schumm, 1993; Harvey et al., 1999). Consequently, in coastal areas adjacent to gently sloping continental shelves, sea-level fall does not influence rates of stream incision into bedrock (Pazzaglia and Gardner, 1993, 1994; Schumm, 1993; Pazzaglia and Brandon, 2001; García et al., 2003, 2004; Wegmann and Pazzaglia, 2009; Pazzaglia, 2022).
The slope of the continental shelf where the Santa Ynez River flows into the Pacific Ocean is relatively gentle (Marsaglia et al., 2019). Measured perpendicular to the local trend of the coastline at the Santa Ynez River mouth, the continental shelf extends 115.6 km offshore along a smooth surface that has an overall slope of 0.016 (data is from the Point Conception to Point Sur 1:216,116 scale [at latitude 35° 20′] nautical navigational chart published by NOAA in 2013). If incision by the Santa Ynez River occurs when sea level falls, a submarine canyon would be cut into the adjacent continental shelf (Schumm, 1993), and highresolution bathymetric data indicate that no such submarine canyon exists (Marsaglia et al., 2019). Researchers have concluded that sealevel fall does not contribute to stream incision into bedrock by the Santa Ynez River because adjacent to the river mouth, the shelf lacks a submarine canyon (Schumm, 1993) and slopes gently for over 115.6 km (for example, Pazzaglia and Gardner, 1993; Schumm, 1993; Pazzaglia and Gardner, 1994; Pazzaglia and Brandon, 2001; García et al., 2003; García et al., 2004; Wegmann and Pazzaglia, 2009; Pazzaglia, 2022, and references therein).
The spatial and temporal relationships of the straths carved by the Santa Ynez River to uplifted marine terrace paleo-shore platforms at the coast indicate that since Orcutt Formation time, the Santa Ynez River was graded to sea level, or to a base level of <1 km inland from the coastline, during sea-level highstands. The strath underlying Qt3 terraces is physically continuous with the Orcutt Formationtime marineterrace paleoshore platform. This indicates that during Qt3/Orcutt Formation time, the lowermost reach of the Santa Ynez River adjusted to sea-level rise, was graded to the sea-level highstand, and was eroding laterally into bedrock. The Qt2 and Qt1 straths are exposed within 1.0–0.5 km inland of temporally correlative marine terrace paleo-shore platforms, and the fluvial straths can be traced into the paleo-shore platforms (Kelty, 2020). This requires that during Qt1 and Qt2, the Santa Ynez River eroded laterally into bedrock and was graded to a base level that was less than 1 km from the coastline. This relationship strongly suggests that, as it was during Orcutt Formation (Qt3) time, the Santa Ynez River was graded to the Qt2/Qt1 sea-level highstands.
Along the lower 10 km of the Santa Ynez River, the modern stream channel is meandering and locally eroding laterally into bedrock-channel walls, which indicates that straths are currently being carved and that, as during Qt3 time, the Santa Ynez River is adjusted to base level at the river mouth during a sea-level highstand. To produce the relationships we observe between fluvial straths and paleo-shore platforms in the lowermost Santa Ynez River Basin, base-level rise due to sea-level rise must be overpowered by tectonically driven base-level fall. Moreover, the exposure and stream-erosion of bedrock at and near the channel level in the lower part of the Santa Ynez River is consistent with minimal, spatially restricted eustatic influence on stream- and terrace-forming processes (Pazzaglia and Brandon, 2001). In the context of our model of strath-terrace production, what is most significant is that the influence of sea-level fluctuations on stream channel and terrace-forming processes, if any, is limited to a few kilometers or less from the coastline.
Climatic Factors and Geochronology
Conditions that generate high sediment yield from hillslopes currently and in past times predominate in the Santa Ynez Valley as well as in the surrounding area. Relatively high sediment yield is due to the current “Mediterranean” to semiarid climate of the study area, which has been identified as optimal for high-magnitude sediment yield from hillslopes (Langbein and Schumm, 1958; Gray et al., 2018), coupled with active mountain building and related steep topography. For example, sedimentation rates in the Santa Barbara Channel (Fig. 1), which is a depocenter for many streams that drain the western Transverse Ranges, are among the highest known on Earth since at least 60 ka (Hendy and Kennett, 1999).
The ages of fluvial terrace deposits in the Santa Maria Basin correspond entirely or partly to periods of relatively dry climate and high sediment yield that led to stream-terrace alluviation within the Transverse Ranges south and east of the Santa Ynez River drainage basin (DeVecchio et al., 2012). These warm, dry climate intervals that led to aggradation and lateral erosion of straths identified by DeVecchio et al. (2012) also correlate to times of stream-terrace sediment deposition (García and Mahan, 2014) as well as alluvial-fan progradation (Stokes and Garcia, 2008) northwest of the Santa Ynez River drainage basin in the central Coast Ranges of California. The role of abundant sediment supply during times of sea-level highstands in our terrace genesis model is discussed in the following section.
Sedimentation, Strath Carving, and Terrace Production
In a broad perspective, fluvial terraces form in tectonically active areas because over graded time scales, stream incision is fundamentally unsteady (Pazzaglia, 2022). The cause of unsteady vertical incision in the Santa Ynez River system is climatically driven fluctuations of the relationship between the carrying capacity of the channel and the sediment load delivered to the channel. During dry periods of high sediment yield within rising hanging-wall blocks, the carrying capacity (“available stream power”) of the Santa Ynez River equals the sediment load delivered to the channel (“critical stream power”), lateral erosion predominates, and straths are carved (for example, Pazzaglia and Brandon, 2001). During these same dry periods of high sediment yield, the sediment load delivered to the channel exceeds the carrying capacity of the Santa Ynez River in subsiding footwall blocks, and aggradation predominates. The different along-channel response of the Santa Ynez River to a sediment load generated by the same climate-forcing event is due to relatively low, slope-controlled available stream power in the subsiding footwall blocks, and relatively high, slope-controlled available stream power in rising hanging-wall blocks.
Although not commonplace, concurrent development of strath terraces and fill terraces within a single drainage basin, as a consequence of climate forcing coupled with tectonically driven base-level changes, has been documented in distinctly different climatic and tectonic settings such as the Olympic Mountains of Northwestern North America (Pazzaglia and Brandon, 2001), the Tien Shan in Kyrgyzstan (Burgette et al., 2017), and the Betic Cordillera of southern Spain (García et al., 2003). Carving straths within rising hanging wall blocks while aggradation occurs in subsiding footwall blocks is, in part, made possible by the abundant supply of sediment in the Santa Ynez watershed.
In the Santa Ynez River system, incision below straths and resultant strath-terrace production occurred during times of relatively moist climate when the sediment load delivered to the stream channel was less than during times of relatively dry climate. In the hanging-wall blocks of reverse faults, vertical incision and strath abandonment are driven by rock uplift due to localized uplift along reverse faults as well as regional rock uplift. In footwall blocks, the effect of regional uplift offsets subsidence along reverse faults and folds, and over graded time scales vertical incision predominates. For instance, in the footwall-block reach between the towns of Buellton and Solvang (Figs. 2 and 4) the tread of a Qt2 terrace adjacent to the channel is ~40 m above the Santa Ynez River channel. The uppermost part of the 16-m-thick alluvial sequence that underlies the Qt2 tread yielded an OSL age of 61 ± 6 ka (Fig. 6B). Within the structural and geomorphic context of the Santa Ynez Valley, tectonically driven, regional rock uplift of the footwall block of the Santa Ynez River fault is the most plausible mechanism that could drive 40 m of incision below a paleofluvial level with 16 m of aggradation that preceded incision.
Regional versus Local Uplift
Our data show that rock uplift is occurring everywhere, in the footwalls of the reverse faults as well as in the hanging walls (Fig. 2). For example, rock-uplift rates in Zaca Creek averaged over the past 85 ka vary from 0.7 mm/yr to 1.0 mm/yr even though this area is in the footwall of two oppositely dipping reverse faults to the north and south. Likewise, rock-uplift rates between the Baseline fault and Little Pine fault in Santa Cruz Creek have been 0.3 mm/yr since folding ceased after Qt3 and before Qt2, and rock-uplift rates in the foot-wall of the Santa Ynez River fault have been ~0.5 mm/yr since deposition of Qt2. Uplift rates measured from marine terraces in the footwall of the Santa Ynez River fault near the mouth of the Santa Ynez River are even higher at ~1.5 mm/yr (Kelty, 2020). These rates at the mouth of the Santa Ynez River are higher than those at other footwall sites and just as high as those at some hanging-wall sites. We are unsure why the rates are higher here but note that this site is not located in a footwall syncline directly adjacent to one of the major faults as are the sites in the Solvang area or in Santa Cruz Creek. It’s possible that uplift rates calculated at the Santa Ynez River mouth site more accurately represent the true regional uplift rate, which is closer to 1–1.5 mm/yr, while the other footwall sites have lower uplift rates due to local subsidence in footwall synclines.
Our interpretation of 0.5–1.5 mm/yr of regional uplift in the Santa Maria Basin since 100 ka is supported by previous studies in the western Transverse Ranges. Rockwell et al. (1984) used fluvial terraces in the central western Transverse Ranges along the upper Ventura River to document uplift rates over the past 92 ka. Although the rates vary slightly through time, the average was 0.79 ± 0.15 (Rockwell et al., 1984). These rates were calculated north of the San Cayetano–Ventura–Pitas Point fault trend and in the footwall of the Arroyo Parida fault. Therefore, they likely reflect the regional component of uplift. At the northern edge of the western Transverse Ranges, incision of dated fluvial terraces in the Cuyama Valley (Fig. 1) also indicates a footwall uplift rate of ~1 mm/yr since 65 ka (DeLong et al., 2007, 2008).
The rock uplift documented in various footwall blocks indicates that there is a regional component of uplift in the western Transverse Ranges. We propose that this regional uplift is due to deeper crustal shortening. Deeper shortening may result from distributed ductile deformation, slip along deeper detachment faults, or a combination of both. Previous large-scale cross-sections across the western Transverse Range and seismic studies have hypothesized a north-dipping detachment at 12–15 km depth beneath the western Transverse Ranges (Huang et al., 1996; Levy et al., 2019; Namson and Davis, 1988; Seeber and Sorlien, 2000). These studies postulate that the major reverse faults of the western Transverse Ranges sole into this detachment system and that many of the large-scale folds at the surface result from bends or ramps in the detachment or are related to listric fault geometries at depth. Slip along a regional detachment, with a nonzero dip, would result in widespread uplift. Johnson et al. (2020) inferred that active uplift across the western Transverse Ranges observed in geodetic data is largely related to interseismic strain accumulation above north-dipping thrust faults that reach the surface south of the Santa Barbara–Ventura coast. In their model of geodetic and geologic data, much of the vertical motion is transient and will be focused along the Santa Barbara–Ventura coast during the next large earthquake on the Ventura–Pitas Point fault. However, their model also includes a component of permanent distributed uplift north of the coast within the western Transverse Ranges, which agrees with the data presented here.
Regional uplift may also be partly due to isostatic adjustment to erosion across the western Transverse Ranges. A complete evaluation of the contribution of isostatic uplift would require more uplift data from the interior of the mountain belt and is beyond the scope of this study, but our first-order observations suggest that the isostatic influence is probably minor compared to the influence of tectonic thickening. First, the higher-elevation areas of the interior of the western Transverse Ranges also have relatively high mean elevations; the average relief of the high interior is not significantly different from that of the lower elevations. Second, the long profile of the Sespe River that drains much of the eastern western Transverse Ranges is convex upward through the lower one-third of the river, which suggests that incision and denudation in that part of the mountain belt has not kept pace with uplift. Third, the relatively high rates of shortening and uplift documented in an area with average annual rainfall of only 40–70 cm/yr (Warrick and Mertes, 2009) suggests that tectonic uplift likely outpaces denudation.
Localized uplift along individual faults at the surface is superposed on this regional component. Total rock uplift in the hanging wall of the reverse faults is a combination of differential uplift across a fault (fault throw) and the regional background uplift. For example, at the Manzanita Road site, rock uplift in the hanging wall of the Santa Ynez fault for the past 45 k.y. has been 0.65 mm/yr, while uplift in the footwall has been 0.23 mm/yr. The difference is represented in the vertical component of the dip-slip rate across the fault, which we measured to be ~0.3 mm/yr.
Possible Changes in Rock-Uplift Rates and Fault-Slip Rates through Time
The incremental incision rates calculated over three different time intervals suggest that rock-uplift rates in the hanging walls of some faults may have changed over the past ~100 k.y. (Fig. 2, Table 3). Although the error ranges are high (as much as ~40% of the rate) for some of the older time intervals due to uncertainties in terrace ages, there is a consistent pattern of decreasing incision rates through time in the hanging-wall blocks. These apparent changes likely reflect changes in deformation rates on individual structures through time. In the hanging-wall blocks of the Santa Ynez River fault and the Baseline fault, incision rates were faster between Qt3 and Qt1 than they have been since Qt1 time (Table 3, Fig. 2). Conversely, incision rates in the footwall of the Santa Ynez fault in the Solvang area appear to increase during the younger time intervals. This suggests that uplift in the hanging wall and corresponding subsidence in the footwall have decreased during the past 100 ka as deformation rates have decreased. Slip rates have not been measured along the blind Santa Ynez River fault (Sylvester and Darrow, 1979), but in the Santa Rita Hills a decrease in deformation of terraces across the fault through time indicates that slip along the fault was fastest between 93 ka and 60 ka, slowed between 60 ka and 40 ka, and had stopped by ca. 35 ka (Kelty, 2020). Along the Baseline fault, the decrease in the hanging-wall uplift rate may be the result of the decrease in fault-slip rate through time that we estimated with offset terraces in Santa Cruz Creek (Table 4). In the western Santa Maria Basin, the Casmalia and Purisima Hills are anticlinoria that formed over a blind thrust system since the deposition of the Orcutt Formation, which occurred between 85 ka and 120 ka (McGregor and Onderdonk, 2021). Rock-uplift rates across these structures average around 2 mm/yr since deposition of the Orcutt Formation, but because younger terrace deposits are not present in this area, there is no way to evaluate changes in uplift rate through time. It is possible that uplift and deformation rates in the Casmalia and Purisima Hills were faster than 2 mm/yr initially and have also decreased in the past 40 k.y., as we see in the other parts of the Santa Maria Basin. Collectively, the data suggest that deformation rates along many of the faults within the Santa Maria Basin have decreased during the past ca. 85 ka.
The faster deformation rates between 85 ka and 40 ka were preceded by a period of relative quiescence. This is primarily based on the interpretations that the Orcutt Formation was deposited on a low-relief erosional surface that once extended across much of the Santa Maria Basin (McGregor and Onderdonk, 2021; Woodring and Bramlette, 1950), and that the Qt3 deposits that we mapped in the eastern Santa Maria Basin are part of the Orcutt Formation. A significant period of tectonic quiescence would be required for the angular unconformity at the base of the Orcutt Formation to form. The projections of the base of the Orcutt Formation over the Casmalia and Purisima Hills (McGregor and Onderdonk, 2021), the elevated hanging wall of the Santa Ynez River fault, and the elevated hanging wall of the Baseline fault indicate that these positive topographic features did not exist prior to 85 ka. Deposition of the Orcutt Formation ceased as the faults began to lift the present topography and what was previously a broad alluvial plain became localized as the rivers incised and formed the younger terrace levels. This interpretation implies episodic movement of the faults (between 85 ka and 40 ka) that is superposed on a more constant regional uplift rate of 0.5–1.5 mm/yr.
We hypothesize that the temporal variations in rock uplift and fault slip are due to changes in how and where north–south shortening across the western Transverse Ranges is accommodated through time. The time scale of these variations is too long to be attributed to changes in faultrupture intervals (hundreds to thousands of years) and too short and local to be due to large-scale changes in plate motion or long-term development of the plate boundary (millions of years). Shortening in the higher elevation interior of the western Transverse Ranges has been occurring since middle Miocene time (Onderdonk, 2005), and reverse faults and uplift began in the Santa Maria Basin in the Pliocene (Woodring and Bramlette, 1950; Behl and Ingle, 1998). Previous studies have inferred that the locus of this deformation across the western Transverse Ranges has changed over millions of years based on large-scale forward modeling of geologic cross-sections (Levy et al., 2019) and thermochronometry (Townsend et al., 2018). We propose that strain partitioning and changes in deformation patterns are also occurring on time scales of tens of thousands of years. We note that an increase in uplift rates measured along the Santa Barbara coast (Gurrola et al., 2014) and near Point Conception (Tutterow, 2021) since ca. 45 ka appears to correspond in time with the observed decrease in uplift and deformation rates along the Baseline and Santa Ynez River faults in the Santa Maria Basin. This may indicate that deformation has shifted southward and that structures in the Santa Barbara Channel are now accommodating a greater amount of the shortening across the western parts of the western Transverse Ranges.
Implications for Tectonic Development and Topographic Growth
This study demonstrates how the topographic development of a compressional mountain belt can be the result of both localized uplift along individual structures and regional uplift from deeper mechanisms accommodating shortening. We show that documenting vertical motions in the footwalls of reverse faults is needed to detect and measure the regional component of rock uplift that would otherwise be missed if only structures that intersect the surface were evaluated. Recognition of the different causes of topographic development leads to a greater understanding of how total shortening across a mountain belt is accommodated at different depths within the upper crust. Measurements of shortening and rock uplift across a mountain belt from geodetic data often do not correspond with longer-term measurements from geologic data, and this mismatch is frequently hypothesized to be related to transient strain. However, discrepancies between rock-uplift measurements over multiple time scales may also result from an incomplete understanding of the factors that contribute to long-term uplift and topographic development.
Modeling of geodetic data across the western Transverse Ranges (e.g., Hammond et al., 2018; Johnson et al., 2020) shows regional uplift rates that are similar to the geologic rates we measured by considering both regional and local uplift. Hammond et al. (2018) noted uplift rates of 1–3 mm/yr in the northern inland parts of the western Transverse Ranges and subsidence along the coast. Because geologic uplift rates are 1–7 mm/yr along the coast, they assumed that most of the regional uplift they observed is due to interseismic strain and that this strain will eventually be expressed as permanent uplift focused along the San Cayetano–Ventura–Pitas Point fault trend at the southern front of the western Transverse Ranges. However, Johnson et al. (2020) constructed a deformation model of the western Transverse Ranges using both the geodetic data and available geologic uplift rates. They concluded that the interseismic arching across the western Transverse Ranges seen in the geodetic data is partially replaced by focused uplift along the southern front of the mountain belt, but that some uplift inland is permanent and is ~2 mm/yr across the Santa Ynez Range. This agrees with the data presented here, which indicate that slip along the Santa Ynez fault contributes ~0.5–1 mm/yr to the uplift of the Santa Ynez Range, leaving another 0.5–1.5 mm/yr of background regional uplift.
Another argument that supports the importance of regional uplift on the overall topography of the western Transverse Ranges comes from comparing the distribution of deformation rates to average elevations in the area. Fast rates of north–south shortening across the mountain belt (6.5–25 mm/yr) have been inferred from both geologic and geodetic studies (Yeats, 1983; Donnellan et al., 1993; Huftile and Yeats, 1995; Hager et al., 1999; Marshall et al., 2013; Levy et al., 2019). These studies generally conclude that the majority of this shortening is accommodated at the surface along the thrust faults that mark the southern boundary of the western Transverse Ranges, primarily the San Cayetano, Ventura, Red Mountain, and Pitas Point faults (Fig. 1). Most of the high topography of the western Transverse Ranges, however, is located farther north, which suggests that late Quaternary deformation at the southern front does not represent the long-term uplift and deformation patterns of the mountain belt. Our new data from the Santa Maria Basin area support this and show that late Quaternary reverse slip and associated topographic growth has occurred along all of the major faults in the western half of the mountain belt. In the higher eastern half of the western Transverse Ranges, however, only the Big Pine fault has documented late Quaternary slip, at a rate of 0.9 mm/yr (DeLong et al., 2007). The Pine Mountain fault and eastern Santa Ynez fault in the interior of the mountain belt accommodated significant amounts of shortening between late Miocene and early Pleistocene time (Onderdonk, 2005). The pronounced topographic gradients across these faults may be the result of ongoing reverse slip, but to date there has been no detailed evaluation of Quaternary slip along either fault.
The fast geologic uplift rates documented along the Santa Barbara–Ventura coastline (up to 7 mm/yr; Rockwell et al., 2016), and above the blind thrust system in the western Santa Maria Basin (up to 5 mm/yr; McGregor and Onderdonk, 2021), are lifting primarily Miocene and younger rocks in their hanging walls to elevations of ~300 m. However, the high-elevation ranges (~2000 m) in the interior of the western Transverse Ranges are in the hanging walls of faults with slip rates of 1 mm/yr or less (Big Pine fault, Little Pine fault, and western Santa Ynez fault). These higher elevations are also underlain by older rocks of Cretaceous through Eocene age. The one exception is the San Cayetano fault that bounds the southern edge of the Topatopa Range in the southeastern western Transverse Ranges. Slip rates there may be as high as 8.75 ± 1.95 at the eastern section of the fault and decrease to ~1 mm/yr toward the west (Rockwell, 1988). This apparent mismatch between fast uplift rates and high topography suggests to us that the present-day topography of the western Transverse Ranges is not a direct result of the rapidly slipping faults along the Santa Barbara–Ventura coast and in the western Santa Maria Basin. Previous studies have hypothesized that there was a southward migration of the thrust front through time such that faults interior to the western Transverse Ranges (the Pine Mountain, Santa Ynez, and Little Pine faults) were active earlier in Pliocene to Pleistocene time but have slowed or stopped as deformation has shifted southward to the San Cayetano–Ventura–Pitas Point fault trend (Rockwell, 1983; Levy et al., 2019). This idea cannot be fully tested because of the current lack of data from the Pine Mountain and eastern Santa Ynez faults, but it is supported by observations that the interior faults accommodated tens of kilometers of shortening from late Miocene to at least late Pliocene time (Onderdonk, 2005), and that the presently fast-slipping faults along the San Cayetano–Ventura–Pitas Point fault trend deform Pleistocene rocks that were shed from the rising western Transverse Ranges to the north (e.g., Rockwell et al., 1988). We hypothesize that in addition to, or instead of, a migration in the locus of deformation over time, the higher interior topography of the western Transverse Ranges may be the result of sustained regional uplift since Pliocene or early Pleistocene time with topographic differences across the mountain belt largely due to differences in rock type and corresponding erodibility. We anticipate that our ongoing work to measure rockuplift rates, timing of uplift, and fault-slip rates in the interior of the range will allow us to test these ideas.
CONCLUSIONS
Uplift, incision, and deformation of fluvial and marine terraces in the Santa Maria Basin area of the western Transverse Ranges record the amounts and rates of rock uplift and deformation over the past ~100 k.y. There are three prominent levels of fluvial strath terraces that correlate to three marine terraces along the coast. Dates from the upper terrace level range from 125 ± 10 ka to 79 ± 7 ka and correlate with the Orcutt Formation and an MIS 5 marine terrace level that is extensively preserved along the California coast. Ages from the middle terrace level range from 49 ± 5 ka to 76 ± 7 ka and correlate with a ca. 60 ka marine terrace near the mouth of the Santa Ynez River. Ages from the lower terrace level range from 32 ± 3 ka to 47 ± 6 ka and correlate with the lowest emergent marine terrace along the coast, which is ca. 45 ka. These strath terraces were laterally cut during warm and dry climactic periods that also led to sea-level rise, and they were subsequently incised during cooler, wetter periods. Using these terraces as markers, we make the following interpretations about rock uplift and fault-slip rates:
Rock uplift is occurring in both the footwall and hanging-wall blocks of the major reverse faults in the Santa Maria Basin, which indicates that there is a background component of regional uplift in addition to localized uplift along individual structures. This regional uplift has occurred at a rate of 0.5–1.5 mm/yr over the past ca. 100 ka.
Rock-uplift rates across some of the major structures may have been variable over the past ca. 100 ka due to variations in fault-slip rates and folding. Although there is significant uncertainty, rock-uplift rates appear to have been highest (1.4–4 mm/yr) between ca. 85 ka and 40 ka and have decreased since then.
The dip-slip rate along the Santa Ynez fault has been 0.48 ± 0.06 mm/yr since 43 ka, but there is likely an additional unknown component of strike-slip.
The dip-slip rate along the Little Pine fault has been 0.13 ± 0.05 mm/yr since 76 ka.
Three measurements along the Baseline fault from two locations and two time intervals suggest a dip-slip rate of ~0.17, with small variations in rate along the fault and through time.
The highest topography of the western Transverse Ranges does not correlate with the areas of fastest uplift and fault-slip rates, which suggests that the large-scale topography of the mountain belt is a result of a change in the locus of deformation over time and/or sustained regional uplift for the past several million years or more.
ACKNOWLEDGMENTS
This work was funded by a National Science Foundation award (#1839301) to N. Onderdonk, and by GSA Graduate Student Research Grants and Johnson-Conrey graduate student fellowships to C. Kelty and A. Farris. We thank Reed Burgette and Karl Wegmann for excellent reviews that greatly improved this paper. We also thank Shannon Mahan and Tammy Rittenour for valuable advice regarding luminescence dating during development of the Luminescence Lab at California State University–Long Beach over the past 5 years.