A Comprehensive Assessment of Submarine Landslides and Mass Wasting Processes Offshore Southern California

It is critical to characterize submarine landslide hazards near dense coastal populations, especially in areas with active faults, which can trigger slope failure, subsequent tsunamis, and damage seabed infrastructure during earthquake shaking. Offshore southern California, numerous marine geophysical surveys have been conducted over the past decade, and high‐resolution bathymetric and subsurface data now cover about 60 percent of the total region between Point Conception and the United States‐Mexico border from the California coast out to the base of Patton Escarpment ∼200 km offshore. In a comprehensive compilation and interpretive mapping effort, we find evidence of seafloor failure throughout offshore southern California with nearly 1,500 submarine landslide‐related features, including 63 discrete slide deposits with debris and >1,400 slide‐related scarps. In our analysis, we highlight new mapping of submarine landslides in Catalina Basin, the Del Mar slide, the San Gabriel slide complex, and the 232 km2 San Nicolas slide, the largest area of any known submarine landslide mass offshore southern California. Analysis of the spatial distribution of submarine landslide features suggests that most mapped slide features are located relatively near coastal sediment sources, particularly during sea‐level lowstand conditions, which underscores the importance of sediment supply and sediment accumulation on low‐gradient slopes as failure preconditioning processes. Tectonically driven uplift at shelf edges and along basin flanks is another key preconditioning factor, and our results also suggest that earthquakes along active faults trigger mass wasting, especially for repeated, small‐scale failures on tectonically steepened slopes.

The region offshore southern California is composed of two complex and geologically active physiographic provinces-the Western Transverse Ranges in the north and the California Continental Borderland (hereafter, "Borderland") to the south, with the northern Channel Islands lying along the boundary of these provinces (Figure 1).A summary of Borderland slide features was previously published by Lee et al. (2009); however, the limited resolution of bathymetric data and seismic reflection profiles at the time precluded an accurate and full accounting of the submarine landslide features in the Borderland, their detailed morphological expressions, and their potential hazards.Dedicated seafloor mapping expeditions during the last two decades have provided nearly complete multibeam echo sounder (MBES) coverage of the Inner Borderland basins (i.e., those located inboard of the Channel Islands, e.g., Dartnell et al., 2015Dartnell et al., , 2017Dartnell et al., , 2021)), and a handful of surveys have acquired new data in the Outer Borderland basins (e.g., Santa Cruz Basin, San Nicolas Basin) and have remapped the Santa Barbara Basin in the northern Western Transverse Ranges physiographic province (Figure 1).These new data provide an opportunity to systematically map and characterize submarine mass transport features offshore of southern California from Point Conception in the north to the United States-Mexico Border in the south (Figure 1).The increase in available high-quality, high-resolution bathymetric and sub-bottom data has also led to new local characterizations of submarine landslides offshore of southern California, including analysis of the Santa Cruz Basin slide complex (Brothers, Maier, et al., 2019), the Santa Barbara Basin slides (Fisher et al., 2005;Greene et al., 2006;Kluesner et al., 2020), and the Palos Verdes debris avalanche (Locat et al., 2004;Normark et al., 2004).
In this study, we explore several hypotheses related to slide preconditioning and failure processes, namely that: (a) significant Plio-Pleistocene sediment deposition is a key preconditioning process and required to generate large slides; (b) uplift along Quaternary faults causes oversteepening and slope instability; (c) earthquake triggering is widespread, especially near Quaternary faults; (d) open-slope versus shelf-edge failures differ in their preconditioning contributions from depositional and fault-related processes; (e) increasing slide size correlates Figure 1.Regional land area/bathymetric map including submarine landslides, headwall scarps, and mass wasting zones compiled and mapped offshore of southern California for this study.High-resolution bathymetry (e.g., Dartnell et al., 2015Dartnell et al., , 2017Dartnell et al., , 2021) ) is in color; low-resolution bathymetry is in light grayscale (National Geophysical Data Center, 2012) with the 130 m depth contour noted with a purple line.Abbreviations: WTR-Western Transverse Ranges physiographic province; SB-Santa Barbara; LA-Los Angeles; SD-San Diego; DM-Del Mar; PC-Point Conception; MEX-Mexico; SBB-Santa Barbara Basin; SMB-Santa Monica Basin; GSC-Gulf of Santa Catalina; SCB-Santa Cruz Basin; CB-Catalina Basin; SNB-San Nicolas Basin; SClB-San Clemente Basin; SPB-San Pedro Basin; SCI-Santa Catalina Island; CR-Catalina Ridge; SNI-San Nicolas Island; SClI-San Clemente Island; SCR-San Clemente Ridge; PE-Patton Escarpment.The black boxes approximate the extents of Figures 3-6 (see labeled arrows), which highlight the San Nicolas slide, the Catalina Basin slides, the Del Mar slide, and the San Gabriel slide complex, respectively.Other major slides discussed in the text are also labeled.
positively with lower failure slopes (and perhaps with other regional depositional or tectonic factors); and (f) large slides are relatively infrequent in the late Holocene.We present a suite of new marine geophysical and geological observations and statistical morphometric analysis of submarine landslide features throughout the southern California offshore region.Our analysis includes the characterization of a number of slides, scarps, and debris aprons for the first time, among them relatively large slides located in San Nicolas Basin, proximal to the San Gabriel submarine canyon, along the continental shelf edge near Del Mar just north of San Diego, and in Catalina Basin (Figure 1).Our results indicate that there are two primary types of geological substrates that are weak and prone to failure: (a) prograding Quaternary shelf-edge deposits that fail along foreset beds (clinoforms) located above active transpressional fault zones (e.g., Fisher et al., 2005), and (b) onlapping and uplifted Plio-Pleistocene basin deposits that fail along basin margins (e.g., Brothers, Maier, et al., 2019).In both cases, sediment accumulation preconditions relatively low-gradient (<5°) slopes for failure.Additionally, tectonic uplift near active faults seems to promote a higher density of slope failures, implying that earthquake shaking promotes repeated small-scale failures along steep slopes.

Tectonic Setting
The southern California offshore region is a unique geologic area that spans from Point Conception to the south of the United States-Mexico border.The physiographic region, which includes the Borderland and the offshore extension of the Western Transverse Ranges to the north (Figure 1), developed along the plate boundary between the North American and Pacific plates that has evolved over the last 20 million years to become the San Andreas transform fault system today (e.g., Atwater, 1970;Jackson & Molnar, 1990;Nicholson et al., 1994;Sorlien et al., 2000;Yeats et al., 1988).Complex Neogene and Quaternary geologic history characterize this region, and a variety of structural styles and tectonic deformation histories exist along the proximal continental margin (e.g., Atwater & Stock, 1998;Bohannon & Geist, 1998;Crouch & Suppe, 1993;Legg, 1991;Legg et al., 2007;Moore, 1969;Nicholson et al., 1994;ten Brink et al., 2000;Vedder, 1987).Tectonic deformation offshore of southern California has created more than twenty fault-bounded basins and ridges that have defined the physiographic setting of the region during most of the Pliocene and Quaternary (e.g., Bohannon & Geist, 1998;Kamerling & Luyendyk, 1985;Legg, 1991;Moore, 1969;Shepard & Emery, 1941;Teng & Gorsline, 1989), with the current San Andreas tectonic regime becoming established with the slowing of Transverse Range rotation at about 3.5 Ma (Wright, 1991).
As the subduction of the Farallon plate ceased during the early-middle Miocene and the Pacific plate came into contact with the North American plate, a broad transform plate boundary began to form offshore southern California (e.g., Crouch & Suppe, 1993;Ingersoll & Rumelhart, 1999;ten Brink et al., 2000).Until the late Pliocene, when the majority of plate motion had shifted ∼200 km eastward to form the present-day San Andreas fault system, most of the tectonic deformation was accommodated by strike-slip and oblique-slip faults in the Borderland region.These fault systems developed the high-relief ridges, steep slopes, and fault-bounded basins that persist today offshore southern California (Bohannon & Geist, 1998;Chaytor et al., 2008;Field & Richmond, 1980;Greene et al., 1975;Kamerling & Luyendyk, 1985;Legg, 1991;Moore, 1969;Shepard & Emery, 1941;ten Brink et al., 2000;Teng & Gorsline, 1989;Vedder, 1987).

Sedimentary Setting
Although few slides have been tied to specific sediment sources, most of Quaternary sediment delivered to the inner basins derive from fluvial and coastal sediment that is funneled across the shelf and slope via submarine canyon-channel systems (Covault & Graham, 2010;Covault & Romans, 2009;Covault et al., 2007Covault et al., , 2010Covault et al., , 2014;;Maier et al., 2018;Normark & McGann, 2004;Normark et al., 2004;Romans et al., 2009;Teng & Gorsline, 1989;Wei et al., 2020).Other smaller terrigenous sources reach some of the outer basins through similar mechanisms, albeit from much smaller fluvial catchments on the Channel Islands.For example, the Santa Cruz Canyon captures coarse-grained terrigenous sediment sourced from Santa Cruz and Santa Rosa Islands, which is transported downslope to a ∼30 km 2 distributary fan and channel system on the basin floor (Barnes, 1970;  (Clarke et al., 1987;McCulloch, 1989;Vedder et al., 1986Vedder et al., , 1987) ) with mapped discrete submarine landslides (polygon outlines colored by slope type) and Quaternary faults from Walton, Papesh, et al. (2020;thin black lines).Extents and rotation similar to part a. Q-Quaternary sedimentary rock; Qsp-San Pedro Formation; QTs-Quaternary and Tertiary sedimentary rock; QTt-Quaternary and late Tertiary sedimentary rock; Tp-Pliocene sedimentary rock; Tpr-Pliocene and late Miocene sedimentary rock; Tm-Miocene sedimentary rock; Tmv-Miocene volcanics; Tmu-undifferentiated Miocene volcanic and sedimentary rock; Tmp-Miocene plutonic and hypabyssal rock; Te-Eocene sedimentary rock; Tep-Eocene and Paleocene sedimentary rock; Ku-Late Cretaceous sedimentary rock; TMz-undifferentiated Miocene igneous rock and pre-Late Cretaceous metamorphic rock; Mz-pre-Late Cretaceous metamorphic rock; m-unknown rock type; gr-Mesozoic granitic rock.Felsher, 1971;Schwalbach et al., 1996).Similarly, Catalina Canyon funnels terrigenous sediment from Santa Catalina Island to the floor of the Catalina Basin (Walton, Brothers, et al., 2020).Sedimentation rates and sources are variable along the coast (e.g., Warrick & Farnsworth, 2009), generally highest in Santa Barbara Basin and lower southward into the Gulf of Santa Catalina.

Geophysical Data
We used a suite of recently acquired high-resolution marine geophysical data combined with seismic reflection datasets to characterize the seabed morphology, map the distribution of submarine landslides, examine slide morphometrics, and interpret the shallow substrate architecture of the slide source regions offshore of southern California.
Raw MBES data from numerous surveys across the study region were obtained from the NOAA National Centers for Environmental Information repository (NCEI; http://www.ngdc.noaa.gov/mgg/bathymetry/relief.html) and edited and combined into mosaicked grids at 25-m cell spacing (Dartnell et al., 2015(Dartnell et al., , 2017(Dartnell et al., , 2021)).In some areas, higher-resolution MBES data are available (e.g., 10-m resolution data in Catalina Basin; Dartnell et al., 2017).Monterey Bay Aquarium Research Institute (MBARI) bathymetry data are available in Santa Barbara Basin (http://sirocco.mbari.org/data/mapping/Santa_Barbara_Basin/).Where higher-resolution MBES data do not exist or are sparse, we utilized the NCEI southern California Coastal Relief Model (CRM; National Geophysical Data Center, 2012), a regional scale, 90-m grid resolution that integrates publicly available singlebeam and multibeam coverage.The Coastal Relief Model was also used to query regional physiographic properties (e.g., water depth, distance to coast, distance to 130-m isobath) to obtain consistent results for these values.
The shallow stratigraphy (upper few hundred meters) in several of the slide areas has recently been imaged with numerous high-resolution multichannel seismic (MCS) and single-channel sparker surveys (data from Balster-Gee et al. (2017, 2020) and Sliter et al. (2017aSliter et al. ( , 2017bSliter et al. ( , 2017c) ) shown in this study), as well as subbottom Chirp available from the Marine Geoscience Data System (MGDS; survey EW9709, https://www.marine-geo.org/tools/search/entry.php?id=EW9709) and Rolling Deck to Repository (R2R; survey SR1701, https://www.rvdata.us/search/cruise/SR1701). Two-way travel time to depth conversions in high-resolution profiles assume a 1,500 m/s seismic velocity.Interpretations of seismic stratigraphy and analysis of the relationships between substrate architecture and seabed morphology were conducted using IHS Kingdom Suite (https://ihsmarkit.com/products/kingdom-seismic-geological-interpretation-software.html).
Linkages between the shallow, Quaternary stratigraphy and the underlying pre-Quaternary basin architecture were interpreted using regional-scale legacy MCS reflection profiles obtained from the USGS National Archive of Marine Seismic Surveys (Triezenberg et al., 2016; https://walrus.wr.usgs.gov/NAMSS/).Select profiles from legacy surveys (e.g., line 124 from L490SC in San Nicolas Basin; https://walrus.wr.usgs.gov/namss/survey/l-4-90-sc/)are presented in this study.These data generally provide 30-50 m vertical resolution and up to 6-s twoway travel time (∼6 km assuming 2,000 m/s seismic velocity) of sub-bottom penetration across the basin depocenters.Navigational offsets within the legacy profiles, arising from the limited positioning capabilities at the time of acquisition, were corrected by shifting profiles to align with correlative seafloor features in the highresolution multibeam bathymetry data.

Slide Identification
Esri ArcGIS Desktop was used to produce hillshade and slope rasters of all bathymetric data and to interpret surficial slide features.Although all data were used to inform interpretations, we generally utilized slope rasters (e.g., Figure 2a) instead of hillshade maps when picking detailed slide morphology such as slide perimeters, scarps, evacuation zones, debris aprons, and slide-prone regions.The interpretive submarine landslide maps have been released as GIS shapefiles in an associated data release (Papesh et al., 2023).After mapping a potential slide feature based on seabed morphology, we used elevation profile graphs generated from the bathymetric data to corroborate our interpretations.Some slide debris aprons (e.g., the Del Mar slide, San Gabriel slide complex) were mapped using subbottom (MCS/Chirp) data in cases where the slide debris apron was not readily mappable from the seafloor expression.The resolution of each mapped feature is based on the dataset with the highest resolution constraining that feature; because of this, the resolution of the features varies within the dataset.We have documented the resolution of the constraining data in the slide shapefile attribute tables (Papesh et al., 2023).
A combination of sediment failure regions and associated debris deposits provides the perimeters for the 63 discrete submarine landslides in our compilation.We refer to the exposed failure surface itself within each discrete submarine landslide as the evacuation zone (e.g., Brothers, Maier, et al., 2019;Clare et al., 2018;Hampton et al., 1996) and use the term debris apron following nomenclature in Brothers, Maier, et al. (2019) to describe the displaced sedimentary mass (i.e., debris) deposited during a slide event (e.g., Hampton et al., 1996).The distal toe of a debris apron thus defines the extent of displacement and runout for a particular submarine landslide (e.g., Clare et al., 2018;Hampton et al., 1996).Debris aprons often persist as topographic highs if emplaced on gentle, smooth slopes, contain hummocks or blocks, and have definable topographic signatures at the edges of the displaced material (e.g., Hampton et al., 1996).We define headwall scarps (following terminology from Varnes, 1978) at the edges of the uppermost extents of the slides, outlining the exposed failure surface where the sediment mass has been displaced.Headwall scarps often exhibit abrupt, concave inflections in the seabed slope with some meters of relief and geometrically form U-shaped curves on the seafloor where the curve of the "U" is topographically higher than the limbs.We mapped headwall scarps as continuous features and note that these can include both headwall and sidewall areas.Each of the 63 discrete submarine landslides in our dataset has headwall scarps, and we also map (as lines) 1,393 additional scarp features that lack identified debris deposits due to being buried and/or unresolvable in available bathymetry data.Lastly, we identify 36 broader areas that contain numerous scarps and/or debris deposits that we interpret as generally slide-prone areas (mass wasting zones).The mass wasting zones were interpreted subjectively in areas where discrete, individual scarps were not mappable due to being too numerous, discontinuous, and/or difficult to identify individually because of locally coarse data resolution.
We include morphometric and metadata attributes for each mapped slide, headwall scarp, evacuation zone, debris apron, and mass wasting zone feature, adapting the guidance of Clare et al. (2018) for our dataset (Papesh et al., 2023).All attribute data were extracted using the available bathymetry (generally the NCEI CRM for depthrelated statistics; National Geophysical Data Center, 2012), published literature, and the 2019 USGS Quaternary Faults Offshore of California database (Walton, Papesh, et al., 2020).We include a brief description of the attributes included in our dataset in Table 1.See Papesh et al. (2023) for complete metadata regarding each of the attributes outlined in Table 1.
Slide volumes reported in our study were calculated based on a wedge geometry following Clare et al. (2018) and McAdoo et al. (2000).The calculation uses the headwall scarp height and the mapped evacuation area of the slide in the formula volume = ½ area × height.Total percent volume uncertainty reported by Papesh et al. (2023) includes propagated uncertainty from both headwall scarp height measurements and area mapping.We note that, while calculated in a consistent manner, our volume estimates may differ from previously published values due to different calculation parameters.

Relevant Sampling and Age Control
The vast majority of slide deposits offshore southern California are undated; only 12 of the nearly 1,600 slide features we document here have been reliably dated by radiocarbon or other techniques.Our submarine landslide data compilation (Papesh et al., 2023) includes previously dated slides (n = 8) and slides dated for this study (n = 4; the Del Mar slide, San Nicolas slide, and two slides in Catalina Basin).
Radiometric ages are sparse and concentrated in areas along the coast for relatively young (<50 ka) slides.Targeted sediment analysis using ship-or ROV-based coring equipment (e.g., gravity, piston, or vibra-cores), which samples the uppermost few meters (<5 m) of seafloor sediment, has provided radiocarbon ages for the Palos Verdes debris avalanche (Normark et al., 2004), rotational slides within the San Gabriel slide complex (Brothers et al., 2015) (Fisher et al., 2005;Greene et al., 2006;Kluesner et al., 2020).
The ages of underlying geologic units in our dataset derive from studies based on dredge and dart-core samples that were published in the 1980s (Clarke et al., 1987;McCulloch, 1989;Vedder et al., 1986Vedder et al., , 1987)).Previous studies have also described the stratigraphic framework using data from regional drilling program holes, including ODP 167 Site 1015, located in the southeastern Santa Monica Basin between San Pedro Basin and Santa Barbara Basin (Figure 2a; Shipboard Scientific Party, 1997), and ODP 146 Site 893 in Santa Barbara Basin to the north (Figure 2a; Kennett & Venz, 1995;Romans et al., 2009).Sites 893 and 1015 both document thick Quaternary sediment cover, sedimentary fans, Holocene turbidites, and slope failures, as well as provide the overall regional sedimentological framework for near-shore basins (e.g., Fisher et al., 2003;Kennett & Venz, 1995;Normark & McGann, 2004;Normark & Piper, 1998;Normark et al., 2006;Romans et al., 2009).

Results
Our dataset (Papesh et al., 2023) includes (a) compilation and thorough reassessment of morphometrics (Table 1) of previously recognized slides, (b) systematic and comprehensive new descriptions of discrete slides throughout the Borderland region, including five large slides that have been recently discovered, and (c) a completely new dataset of headwall scarps and mass wasting zones containing numerous headwall scarps.Below, we first summarize previous characterizations of significant submarine landslides, alongside new morphometrics from our study where relevant, and then describe new mapping of the San Nicolas slide, slides in Catalina Basin, the Del Mar slide, the San Gabriel slide complex, and other mass wasting features (scarps, mass wasting zones).Our dataset includes 63 discrete submarine landslide features with associated evacuation zones and debris aprons, 1,456 individually mapped scarps, and 36 mass wasting zones that exhibit geomorphology consistent with numerous small-scale slope failures.Approximately 10% of the ∼60,000 km 2 region between Point Conception and the United States-Mexico border out to Patton Escarpment (Figure 1) is covered by seafloor mass-wasting features between the 63 slides with debris (total area of ∼1,260 km 2 including evacuation and deposit areas) and the 36 mass wasting zones (covering ∼4,560 km 2 ).

Santa Cruz Basin Slides
Some of the largest translational open-slope failures in the study area exist within Santa Cruz Basin (SCB; Figure 1).The SCB area exhibits complex and diverse seabed morphology with ∼350 km 2 of its seafloor modified by mass transport processes, including 10 discrete slides, and has been previously described in detail by Brothers, Maier, et al. (2019).SCB is flanked by the steep east-facing Santa Rosa-Cortes Ridge with a mean downslope seabed gradient between 6°and 9°at 200-1,600 m water depths.The mid to lower slope (1,000-1,400 m water depth) is covered by a series of arcuate headwall and sidewall scarps ranging from 2 to 16 km in length with highly variable angularity and steepness.Debris aprons associated with the scarps are generally broad and lobate with positive-relief blocky and hummocky debris (Brothers, Maier, et al., 2019).Slides in SCB can be classified as predominantly open slope and translational, with initial basal slip occurring along stratigraphic contacts in the steep section of the lower slope (∼7°).Based on stratigraphic relationships observed in the subsurface, Brothers, Maier, et al. (2019) inferred that the failed sediment is Pliocene with a failure surface close to the Miocene-Pliocene boundary and failures occurring incrementally throughout the Quaternary.No significant failures were observed on shallower slopes composed of Miocene sedimentary rocks above the onlapping Pliocene wedge (Brothers, Maier, et al., 2019; Figure 2b), suggesting that the upslope limit of the failures may be a function of Pliocene sediment thickness and yield strength along the failure surface, which here is exceeded for sediment thickness >50 m (see Brothers, Andrews, et al. (2019) and Brothers, Maier, et al. (2019) for detailed analysis).

Santa Barbara Basin Slides
Santa Barbara Basin (SBB; Figure 1) contains eight discrete slides in total with some of the largest shelf-edge failures in the study area.SBB is a foreland basin that lies along the propagation deformation front of the rotated, east-west trending WTR fold and thrust belt that is currently undergoing tectonic convergence and uplift (e.g., Bohannon & Geist, 1998;Luyendyk et al., 1980;Namson & Davis, 1988;Nicholson et al., 1994).West of Santa Barbara, California, SBB hosts multiple well-studied submarine landslide failures as well as a rich fluidflow system including the petroleum-producing Miocene Monterey Formation, tar mounds, structurally controlled oil and gas seeps, mud volcanoes, pockmarks, seafloor fissures, bacterial mats, and precipitates of authigenic carbonate (e.g., Eichhubl & Boles, 2000).Kluesner et al. (2020) showed that the compaction and porosity reduction in SBB from convergent tectonics drives pore fluids updip along structural trends and preconditions slopes for failure.The largest of the SBB failures is the Goleta slide complex (Figure 1), one of the best-studied slides offshore of southern California (e.g., Kluesner et al., 2020), which is about 15 km long, 11 km wide, extends downslope to depths of about 580 m, and has displaced between 0.3 km 3 (calculation from accompanying data release; Papesh et al., 2023) and 1.5 km 3 of Holocene sediment from the upper slope region (Greene et al., 2006;McCulloch, 1989;Vedder et al., 1986).The headwall scarps sit at an average depth of ∼144 m and the slope in the scarp area is as steep as 40-45° (Fisher et al., 2005).Greene et al. (2006) used surficial geomorphology from multibeam bathymetry to identify distinct lobes; the earliest (now buried) Goleta slide is as old as 200 ka (as derived from seismic imaging and sedimentation rates), but the three most recent slides (those included with our compilation) have been dated at 6-8 ka, 8 ka, and 10 ka (Fisher et al., 2005;Greene et al., 2006).The smaller (0.01-0.2 km 3 ) neighboring Gaviota landslide occurred due to similar processes (tectonic compaction, porosity reduction) as the Goleta complex (Fisher et al., 2005;Greene et al., 2006;Kluesner et al., 2020).

The Palos Verdes Debris Avalanche
The Palos Verdes debris avalanche is another well-studied shelf-edge slide complex in the San Pedro Basin (Figure 1), and represents two of the three discrete slides mapped in that basin.The Palos Verdes debris avalanche was first discovered by Emery and Terry (1956) and interpreted as a submarine landslide by Gorsline et al. (1984).
The slide complex lies on the southwest side of the Palos Verdes anticlinorium and consists of a gullied headwall that follows the shelf break for about 13 km, a slump deposit on the lower slope, a debris field on the floor of San Pedro Basin that extends up to 18 km from the headwall, and a total area of ∼91 km 2 .The estimated volume of the slide is between 0.34 and 0.73 km 3 (Bohannon & Gardner, 2004), which is potentially tsunamigenic (Locat et al., 2004).An age of 7,500 ka was determined for the most recent debris-avalanche deposit (Normark et al., 2004).While Normark et al. (2004) considered the deposit to be a single event, the presence of numerous headwall scarps, isolated debris fields in San Pedro Basin, and mass transport deposits (MTD) in the subsurface indicate that numerous slope failures have occurred (e.g., Bohannon & Gardner, 2004).The Palos Verdes debris avalanche is a somewhat unusual southern California slide in that it features blocky debris (e.g., Locat et al., 2004), whereas most other discrete submarine landslides in our study are translational sediment failures.
The uplift of the Palos Verdes Hills has served to isolate this part of the coast from significant littoral sediment supply; thus, slide material is likely sourced from Miocene-Pliocene sedimentary rocks and/or bedrock on the southwest limb of the Palos Verdes anticlinorium (Bohannon & Gardner, 2004;Vedder et al., 1986).

The San Nicolas Slide
The San Nicolas Basin in the California Outer Borderland (Figure 1) contains two discrete slides, one of which is the largest submarine landslide by area that is known offshore of southern California, first recognized in new bathymetry data collected by the R/V Sally Ride in 2017 (Rolling Deck to Repository, 2017).The failure occurred along the southwestern slope of the San Nicolas Basin, with a runout direction to the ENE and a runout distance of 22 km (Figure 3).The San Nicolas slide has a morphology typical of a single discrete translational submarine landslide occurring on an open slope with a dip of ∼4-5°.The slide has a continuous ∼27-km-long, ∼100-m-high combined headwall/sidewall scarp (Figures 3a and 3b) at a water depth of 1,109 m.The total area of the slide is 232 km 2 , including both the scarp area and the lobate debris apron on the San Nicolas Basin floor; the debris field represents about 196 km 2 of this total area (Figure 3a).According to shallow geologic mapping of the San Nicolas Basin area using dart cores (Vedder et al., 1986; Figure 2b), the headwall scarp is located in undifferentiated Pliocene and Quaternary sedimentary rocks downslope of mapped Miocene deposits.The nearest land mass, San Nicolas Island, is located ∼41 km northwest of the headwall scarp; the nearest mapped Quaternary fault system, the San Clemente fault, is ∼48 km east of the headwall scarp.The Ferrelo fault is the only potential Quaternaryactive feature in the vicinity with the potential to trigger slides, but also does not exhibit any significant Quaternary activity within ∼40 km of the San Nicolas slide (Legg et al., 2015).
In the subsurface (Figures 3b and 3c), the San Nicolas slide debris apron is an acoustically chaotic package up to ∼300 m thick (Figure 3b) and draped by ∼15 m of stratified sediment (Figure 3c).Upslope, the basal failure surface is the same as the base of an onlapping sedimentary package, suggesting that the onlapping package failed along its basal contact (Figure 3b).The failed onlapping sediment is similar to the stratigraphic relationship observed in slide evacuation areas in the neighboring Santa Cruz Basin to the north, first characterized by Brothers, Maier, et al. (2019).
ODP 167 Site 1013 (Shipboard Scientific Party, 1997) is located just southeast of the San Nicolas slide (Figure 3a).The 146.1-m drill hole contains a relatively simple sedimentary sequence of siliciclastic clay and other minerals that decrease steadily downhole, where calcareous minerals are more abundant (Shipboard Scientific Party, 1997).The age model, based primarily on magnetostratigraphy and calcareous nanofossil and foraminifera biostratigraphy, reaches the uppermost Pliocene (2.7 Ma) at the base of the drill hole (Figure 3c).The sedimentation rate (average of 65 m/Myr) remains relatively constant throughout the sequence without any sharp lithologic or age transitions (Shipboard Scientific Party, 1997).We map the slide debris-capping surface at the base of the ∼15 m drape over to the Site 1013 drill hole using Chirp data (following Janik et al., 2004; Figure 3c), correlating the debris-capping surface to the age model at ∼350 ka.The average sedimentation rate in the upper 35.66 m of Site 1013 is slightly higher than average at 77.5 m/Myr, which suggests the ∼15 m drape has accumulated since ∼200 ka.These age constraints place the San Nicolas slide at 200-350 ka at the youngest.Due to stratigraphic relationships, the depth of the Site 1013 hole, and data coverage, we were unable to map the onlap failure surface at the headwall scarp (Figure 3b) to Site 1013.

Catalina Basin Slides
Catalina Basin, located in the Inner Borderland (Figure 1), contains three total discrete slides, two of which are significant failures in the northeastern part of the basin first documented by Walton, Brothers, et al. (2020) (Figure 4).The slides are located at just under 1,000 m water depth and downslope of the shelf break offshore of northwestern Santa Catalina Island along the south-facing slope of Catalina Ridge (Figure 4a).The total area of the larger southern slide is ∼10 km 2 and the northern slide is ∼3 km 2 .The two discrete slides each have arcuate, connected headwall and sidewall scarps (relief of up to ∼40 m) located on an open slope area dipping at approximately 8-12°in the headwall region (Figure 4a).Both slides appear to be translational with SW-trending lobate debris aprons; the debris apron for the larger southern slide has a runout of ∼7 km and appears to terminate at the ∼5-m-relief Catalina fault (Walton, Brothers, et al., 2020; Figure 4a).The nearest land is Santa Catalina Island, and the closest shoreline to the slide headwalls is 6-8 km away.The nearest fault system, the Catalina fault, crosses through the slide area with the headwall scarps located a kilometer or less from the primary strand of the Catalina fault system (Figure 4a).Oblique Chirp (Figure 4a) and MCS (Figure 4b) crossings of the slide depict a ∼40-m-high sidewall scarp and a thin downslope debris deposit flanked by faults and indicate that the failure occurred on an onlap surface similar to the failure surfaces mapped in San Nicolas Basin (Figure 3b) and Santa Cruz Basin (Brothers, Maier, et al., 2019).The failure surface correlates with a regional unconformity mapped by Walton, Brothers, et al. (2020), who ascribed the failure to the Pliocene based on geologic mapping (Vedder et al., 1986; Figure 2b).A ∼5 m drape layer atop the slide debris (Figure 4a), along with an assumed hemipelagic sedimentation rate of 10 cm/kyr (Maier et al., 2018;Normark, McGann, & Sliter, 2009), suggests that the slides occurred in the late Pleistocene, ∼50 ka at youngest.

The Del Mar Slide
The Del Mar slide lies about 3 km north of the La Jolla submarine canyon offshore of the northern San Diego beach cities of Del Mar and Solana Beach (Figures 1 and 5) and is one of 16 discrete slides located in the Gulf of Santa Catalina.The Rose Canyon fault zone trends along the outer shelf parallel to the coast in the vicinity of the Del Mar slide (Figure 5a; Conrad et al., 2019) and has slip estimates from onshore fault exposures of about 1.5 mm/yr, suggesting an average earthquake recurrence interval of about 4 kyr (Lindvall & Rockwell, 1995).The ∼9-km-long headwall scarp of the Del Mar slide lies 3-4 km offshore and is coincident with the shelf break for ∼7 km of its length (Figure 5a).The steep, gullied headwall is about 200 m high, with an average slope of about 13°just past the shelf break.The location of the slide was revealed when new high-resolution bathymetry data acquired as part of the California State Waters Seafloor Mapping Program (Seafloor Mapping Lab, 2018) showed the gullied headwall at the shelf edge.Except for a small area of rough seafloor suggesting blocky debris extending 1-2 km downslope from the headwall, an extensive debris field is not obvious on existing bathymetry data (Figure 5a).Single-channel minisparker profiles (Sliter et al., 2017b), however, show slide debris up to 20 m thick that extends at least 10 km downslope to the west underlying a layer of about 3 m of hemipelagic sediment (Figures 5b and 5c; Conrad et al., 2019).The minisparker data suggest that the debris field covers an area of approximately 57 km 2 .Radiocarbon ages from a core through the hemipelagic drape indicate a failure age of approximately 16,400 years before present for the most recent failure (Table 2, Figure 5d).Seismic reflection profiles also show at least 6 buried MTD underlying the slope to the west of the slide headwall, indicating numerous older slide bodies (Figure 5c).

The San Gabriel Slide Complex
The San Gabriel slide complex is located in the Gulf of Santa Catalina along the south-facing margin of the San Pedro Shelf and closely parallels the San Gabriel submarine canyon (Figures 1 and 6).The San Gabriel canyon  2. itself aligns with the trend of the active Palos Verdes fault zone (Figure 6a; Brothers et al., 2015) and has not had any documented Holocene sediment flows (Maier et al., 2018).The primary San Gabriel slide, while one of the largest in our study by area at 164 km 2 and the largest by volume, has uncertain dimensions due to the debris apron being reworked by sediment flows down the San Gabriel canyon (Figure 6a).The depths of the slides in the San Gabriel canyon area range from a shallow shelf depth of about 60 m to the base of the slope at about 809 m, covering a maximum runout distance of about 33 km (Figure 6a).Downslope from the San Gabriel canyon head at depths of about 430-610 m, a series of 8 mapped slumps with arcuate headwalls up to 75 m high and roughly 1-5 km across are located mostly west of, and partly incised by, the San Gabriel canyon channel (Figure 6a; Brothers et al., 2015).Seismic reflection profiles show that these are headwall and sidewall scarps of rotational slumps that involve minor translation (several tens of meters) of blocks of relatively intact sedimentary sections (Brothers et al., 2015), similar to complex compound slides described by Greene et al. (2006).Late Pleistocene failure ages of about 31 ka and 17-24 ka were determined for two of the uppermost slumps, the headwalls of which are offset up to 55 m by right-lateral offset along a strand of the Palos Verdes fault zone (Brothers et al., 2015).

Headwall Scarps, Slide-Prone Regions, and Other Mass-Wasting Features
Many mass-wasting features in our study lack discrete, mappable debris-flow deposits.We include mapping of 1,456 headwalls and any connected sidewall scarps (of which only 63 have mappable debris deposits) in the study area, which range in length from 19 m to just over 27 km (Figures 1 and 2a).These occur at water depths up to 3,600 m, distances up to 96 km from coastlines (mean of 15 km; Figure 7), and as far as 91 km from mapped Quaternary fault systems (mean of 13 km; Walton, Papesh, et al., 2020) (Figure 8).In general, scarps are concentrated on the steep slopes (up to 80°-90°) of tectonically uplifted ridges bounding the Borderland basins and in areas where sediment supply is higher, such as along the edges of the continental shelf and in the upper reaches of most of the major submarine canyons such as the Hueneme, Santa Ana River, La Jolla, and San Gabriel canyons (Figures 1,2a,5a,and 6a).These canyons often feature numerous small headwall scarps near the shelf edge and/or canyon walls, suggesting routine small-scale mass wasting resulting in localized slumping or turbidity currents.
Fault-bounded basins flanked by prominent, steep ridges include Patton Escarpment bounding the western edge of the Outer Borderland as well as numerous ridges within the Borderland such as Catalina Ridge and San Clemente Ridge (Figure 1; Walton, Brothers, et al., 2020).These steep ridges are host to numerous small-scale mass-wasting features, particularly scarps lacking distinct debris deposits, where it is often difficult to interpret discrete individual features.In total, our study identified 36 distinct slide-prone regions (mass wasting zones) located along steep escarpments in the Borderland and adjacent regions (Figures 1  and 2a).The mean slope in these regions is 12°(but can reach up to 80°-90°) and the total area covered by these mass wasting zones is just over 4,500 km 2 .

Discussion
Together, interpretations of our submarine landslide data compilation (Papesh et al., 2023) suggest that a combination of sediment supply, tectonic uplift, and fault activity are the primary factors that have led to the landscape of mass wasting features we observe today offshore of southern California.Below, we present analysis of slide features' proximity to terrestrial sediment sources over time (a proxy for sediment accumulation), proximity to Quaternary faults (a proxy for earthquake shaking/triggering), and the role of tectonic uplift on slope gradient and sediment accumulation.We also examine the relationship of landslide size and depositional environment (open slope vs. shelf-edge) with each of these processes.Lastly, we present our interpretations' implications for sediment strength and regional hazard and outline next steps for improving our understanding of submarine landslide processes offshore southern California.

Submarine Landslide Distribution and Statistics
Our results suggest a positive relationship between proximity of terrestrial sediment sources (sediment supply) and mass wasting processes in the Borderland region and emphasize the importance of lowstand sedimentation for preconditioning subsequent slope failure (Figure 7).To investigate the role of sediment supply on slide occurrence in the Borderland, we compare the proximity of slide features (evacuation areas from discrete submarine landslides and headwall scarps) to the modern coastline (Figure 7a) and the 130 m depth isobath (Figures 1 and 7b).Higher sedimentation rates generally occur during sea-level lowstands (e.g., Covault & Graham, 2010;Normark, McGann, & Sliter, 2009;Normark, Piper, et al., 2009), so we use the 130-m isobath to represent the approximate Last Glacial Maximum (LGM) lowstand paleoshoreline in our analysis; shelf break depths range from ∼120 to ∼140 m in the region (e.g., Chaytor et al., 2008) and eustatic sea level rise since the LGM is also estimated between 120 and 140 m (Lambeck et al., 2002;Peltier, 2002;Shackleton, 2000;Yokoyama et al., 2000).The results indicate that 78% of mapped slide evacuation features occur within 20 km of the modern shoreline, and 88% of the same features occur within 20 km of the 130-m isobath (Figure 7).Because more slide features occur closer to the approximate LGM lowstand coastline than the modern coastline, we interpret our results to suggest that sediment supply and higher lowstand sedimentation rates contributed to preconditioning the ∼60%-70% of failures that occurred in Quaternary sedimentary units (n = 43/63 discrete failures, n = 854/1,456 scarps).Mapped features farther than 20 km from the modern shoreline are largely located in the Outer Borderland basins (e.g., Santa Cruz Basin, San Nicolas Basin), where ∼80% of slide activity also occurred in Pliocene and older units; therefore, lowstand sedimentation is likely not a significant preconditioning factor in the Outer Borderland.Lowstand sedimentation may also be less important than other preconditioning factors in Arguello Canyon in the northern study area (Figure 1), which hosts the majority of the mass wasting features located farther than 20 km from the 130-m isobath (e.g., Marsaglia et al., 2019).(National Geophysical Data Center, 2012), an approximation for the lowstand Last Glacial Maximum (LGM) coastline (e.g., Chaytor et al., 2008;Lambeck et al., 2002;Peltier, 2002;Shackleton, 2000;Yokoyama et al., 2000).Blue bars show headwalls (binned at 1 km) and pale orange bars show landslides (binned at 10 km) from Papesh et al. (2023).The data show that 88% of these slide features occur within 20 km of the approximate LGM coastline.
Our results also indicate a positive relationship between slides and Quaternary faulting (Figure 8), suggesting that shaking from earthquakes along active faults triggers nearby slides in the Borderland and adjacent regions.We examine the relationship between slide features (evacuation areas from discrete submarine landslides and headwall scarps) and their proximity to mapped Quaternary faults (Walton, Papesh, et al., 2020) and find that approximately 75% of mapped slide evacuation features occur within 10 km of a mapped Quaternary fault (Figure 8).As with the coastline proximity data, the majority of mapped slide features located farther than 10 km from Quaternary faults are generally located in Outer Borderland basins (e.g., Santa Cruz Basin, San Nicolas Basin), which contain some of the largest failures, and in the Arguello Canyon area.Some Quaternary fault activity has been hypothesized in the Outer Borderland area (e.g., Ferrelo fault; Legg et al., 2015); however, the amount of slip on Outer Borderland faults is not well constrained.
The numerous scarps in our dataset, in addition to 36 mass wasting zone areas, demonstrate widespread small-scale mass wasting processes along steep, uplifted slopes throughout the Borderland region.Of the 1,456 scarps in our dataset, 1,393 lack associated debris aprons mappable in available bathymetric data.We suggest this may be because associated debris has been buried and/or eroded, has been evacuated completely via a turbidity current, is too small or thin to be resolved by existing bathymetric data, or because some mapped scarps may be related to non-slide conditions and processes such as faulting or local geology (e.g., natural overhangs due to differential erosion of bedding layers).In general, we interpret the majority of the scarp features as small-scale slope failures that occur frequently in areas with relatively steep slopes, where sediment cannot settle as easily.Areas with steep slopes are also generally coincident with Quaternary faults (e.g., Figure 2), which generate local uplift and cause shaking that can trigger slope failure.Because scarp features that lack mappable debris are similar to the larger slides with debris deposits in terms of proximity to coastal sediment supply (Figure 7) and Quaternary faults (Figure 8), we suggest that similar processes promote the occurrence of slope failures at all scales, that is, increased sediment supply, tectonic uplift, and repeated earthquake shaking along active faults, which seems to be particularly important for the numerous small-scale failures we map as scarps.
Our results support the interpretation that uplift and transpression along active fault systems play a significant role in preconditioning slope failure for sedimentary units of all ages in the Borderland.In our dataset, about half (n = 27) of the 63 discrete submarine landslides failed in Quaternary sediment, only 4 of which occurred in late Pleistocene-Holocene units.The other ∼half (n = 36) of discrete slides occurred in Pliocene or older deposits, with Pliocene being one of the most common (n = 28) failed sedimentary unit ages.The ages of the failed sedimentary units are consistent with timelines of regional sediment deposition and deformation suggested by previous work (DeMets & Merkouriev, 2016;Ingersoll & Rumelhart, 1999;Luyendyk et al., 1980;ten Brink et al., 2000;Walton, Brothers, et al., 2020), with sediment deposition during the Pliocene (accelerated in some areas during Pleistocene sea-level lowstand) and continued deposition and regional transpressional uplift during the Pleistocene, leading to increased slopes and subsequent failures at that time.

Characteristics of the Largest Submarine Landslides
In general, larger slides offshore of southern California occur on statically unstable, tectonically uplifted slopes of 5°-10°(Figure 9; McAdoo et al., 2000).In a comparison of slope versus discrete slide area (Figure 9), the majority (75%) of discrete slides in our dataset are smaller than 20 km 2 and occur on a wide range of slopes from 0.6°to 36.8°.The largest 25% of discrete slides by area (>20 km 2 ) in our compilation are similar to the other mass-wasting features in that they occur relatively near sediment supply sources at modern and paleoshorelines (within 20 and 10 km on average, respectively; e.g., Figure 7).Most of the largest slides occur in areas that have been tectonically uplifted along basin edges (e.g., Santa Cruz basin slides, ∼5°-10°slopes; Figure 9) or due to transpression near the shelf edge (e.g., Goleta slide complex, Palos Verdes debris avalanche), both factors that precondition slope failure.(Walton, Papesh, et al., 2020).Blue bars show headwalls (binned at 1 km) and pale orange bars show discrete slides (binned at 5 km) from Papesh et al. (2023).The data show that approximately 3/4 (75%) of these slide features occur within 10 km of mapped offshore Quaternary faults.
Our results emphasize the importance of sediment accumulation on low slopes to precondition large failures.The largest 25% of discrete slides generally occurred on the lowest slopes overall (<10°; Figure 9), and the two largest slides in our dataset (San Nicolas slide, San Gabriel slide complex) occurred on slopes <5°(Figure 9).Smaller discrete failures <20 km 2 (Figure 9) as well as scarp features, which we consider to be the smallest failures in our dataset, almost always occur on steep slopes at shelf breaks or along the edges of canyon systems (e.g., Figure 1).Intuitively, more sediment can accumulate on low slopes, allowing for high-volume slides in low-slope areas (Figure 9).The reverse applies as well; steep topography prevents significant sediment accumulation and necessarily leads to smaller failures in steep areas (Figure 9).
Our results suggest that proximity to fault-related processes also affects slide size (Figure 10).We find a loose correlation between increasing slide size and increasing Quaternary fault distance (Figure 10), with the largest 25% of discrete slides located twice as far (mean of 18 km) as the discrete slide average (mean of 9 km) from mapped Quaternary faults.The largest slide by area (the San Nicolas slide) is located 48 km from the nearest mapped Quaternary fault (Figure 10).Consistent with the correlation between slide size and fault distance, 75% of the smallest failures in our dataset (scarp features that lack debris deposits) are located within 10 km of Quaternary faults (Figure 8).Fault-related processes (oversteepening from uplift, earthquake triggering) can increase the spatial and temporal frequency of slides, but these processes can also inhibit sediment accumulation, affecting slide size.Larger slides located relatively far from faults may be less uplifted than areas closer to faults, allowing for more sediment to accumulate on lower slopes, and the largest slides in our dataset indeed occur on relatively low slopes (Figure 9).Away from active faults, less frequent shaking and longer earthquake recurrence intervals should also allow more time for sediment accumulation, thus preconditioning larger failures.Following that logic, earthquake triggering for the largest 25% of slides in our dataset may be infrequent compared to small slides.Areas instead prone to frequent earthquake shaking (i.e., closer to faults) generate frequent, smaller failures (i.e., scarps; Figure 8) and limit the buildup of thick sediment that could generate a single large failure.Our observations are similar to submarine landslide patterns observed along the seismically active Queen Charlotte fault margin in the northeastern Pacific, where relatively large slides are also located farther from the active fault trace in deeper water depocenters (Brothers, Andrews, et al., 2019;Greene et al., 2019).

Shelf-Edge Versus Basin Failures
We note that two primary modes of slope failure occur in the Borderland region: failure along shelf edges (e.g., Del Mar, San Gabriel, and Palos Verdes slides) and within basins along the flanks of uplifted ridges (e.g., Santa Cruz Basin, Catalina Basin, and San Nicolas slides).All of these slides share similar preconditioning factors (i.e., high sedimentation rates, tectonic uplift), but varying degrees of preconditioning and environmental differences in shelf edge versus basin settings lead to differences in slide failure planes, triggering, recurrence, and size.
The shelf-edge area (Figures 2b, 9, and 10) is a critical environment for slide generation due to the localized steepness of the upper continental slope, in places tectonically enhanced, and repeated sediment recharging by deposition of unconsolidated littoral deltas during Pleistocene sea-level lowstands.The shelf-edge environment is also prone to relatively small, repeated failures due  to seismic triggering.Numerous examples of shelf-edge slides in our dataset include the Santa Barbara Basin slides, the Palos Verdes debris avalanche, the San Gabriel slide complex, and the Del Mar slide.In each of these areas, repeated Pleistocene lowstands recharged sediment at the shelf edge, preconditioning repeated failures of Quaternary sediment along bedding planes within prograding delta sequences in susceptible sections (e.g., Palos Verdes debris avalanche, Brothers et al., 2015;Normark et al., 2004).There is evidence for recurrent failures in the subsurface for many shelf-edge slides (e.g., Palos Verdes debris avalanche, Bohannon & Gardner, 2004;Goleta slide complex, Fisher et al., 2005;Greene et al., 2006; Del Mar slide, Figure 5; San Gabriel slide complex, Figure 6).Shelf-edge deltas (clinoforms), primarily derived from Pleistocene lowstand littoral deposits (e.g., Normark et al., 2006), are also being uplifted along active transpressional faults, promoting instability.For instance, failures in the Palos Verdes debris avalanche area have been driven by ongoing uplift of the Palos Verdes anticlinorium caused by a restraining left bend of the Palos Verdes fault over the past 2-3 Ma (Ward & Valensise, 1994).
Relatively  2b).Open-slope failures in distal basins are generally located farther from terrestrial sediment supply, and many open-slope failures in our dataset occur within Pliocene or older sediment partly because of this.(Pleistocene sedimentation was still likely somewhat elevated in the distal basins, however; lower sea level increased shoreline length and subaerial land in the distal basin areas, enhancing erosion and sediment supply in regions that are otherwise sediment-starved at highstands).Tectonic uplift is another key slide preconditioning factor in the basin environments-regional extension, crustal block tilting, and basin formation starting in the late Miocene led to favorable failure conditions such as oversteepened slopes for any accumulated post-Miocene sediment (e.g., Walton, Brothers, et al., 2020).
Finally, open-slope failures in general are located farther from Quaternary faults (mean of ∼16 km) than shelfrelated failures (mean of ∼2 km), which makes open-slope environments more likely to have longer earthquake recurrence intervals (i.e., less frequent slide triggers) compared to shelf-edge settings (Figure 10).Our results suggest that shelf-edge failures are likely repeatedly triggered by earthquakes along the active faults.Shelf-edge failures tend to be relatively close to the active fault systems-the Del Mar slide headwall, for example, coincides with a strand of the Rose Canyon fault zone (Figure 5a; Conrad et al., 2019), and all discrete shelf-edge failures occur within ∼6 km of Quaternary faults.In open-slope settings farther from Quaternary faults, infrequent shaking leads to fewer triggers and also allows for more sediment accumulation between triggers, thus favoring relatively large slides as well.Most (10 of 16) of the largest 25% of discrete slides are located on open slopes within basins in the Outer Borderland (Santa Cruz Basin, San Nicolas Basin, and San Clemente Basin; Figure 10).As examples of these fault-related processes, the large open-slope San Nicolas slide seems to have been a rare event due to a lack of evidence for repeated failures in the subsurface (Figure 3), whereas shelf-edge failures tend to occur repeatedly (e.g., Del Mar slide, Figure 5; San Gabriel slide complex, Figure 6).

Implications for Sediment Strength
There is a clear relationship between earthquake triggers and smaller slides, especially very small failures (i.e., scarps) lacking mappable debris deposits (Figure 8), but our results may also imply that sediment offshore southern California has been seismically strengthened from repeated shaking along numerous active faults.Sawyer and DeVore (2015) found empirical evidence from drilling data suggesting that repeated earthquakes result in increased shear strength in the upper 100 m of sediment.Strengthened sediment can also lead to relatively large slides on low slopes, and likewise, small slides in cohesive sediment require steeper slopes to fail (Frattini & Crosta, 2013), which is consistent with our results (Figure 9).
As an active margin, southern California has far smaller slides and generally fewer slides with significant debris fields compared to passive margins (e.g., U.S. Atlantic margin; Chaytor et al., 2007).Of 106 slides analyzed for size offshore of the U.S. Atlantic margin (Chaytor et al., 2009), 70% of the evacuation areas are >10 km 2 , whereas only 13% (n = 8/63) of the southern California slide evacuation areas exceed 10 km 2 .Our results provide evidence that (a) sediment strengthening from earthquake shaking leads to relatively fewer discrete failures offshore southern California than passive margins, (b) stress from the earthquake shaking still sometimes exceeds the sediment shear strength failure threshold and leads to slope failures, and (c) failures offshore southern California and other active margins may lead to relatively small failures compared to passive margins due to repeated shaking.

Hazard Implications and Paths Forward
While Borderland area slides are a concern due to dense coastal California populations, our results suggest that large submarine landslides perhaps occur more often as a result of Pleistocene/lowstand and/or transgressional processes than as a result of Holocene conditions or triggers.Ages of dated slide deposits (n = 12 out of 63 discrete slide features) span a wide range (100 years-200 ka), with half (n = 6) in the late Pleistocene, half (n = 6) in the Holocene, and only one slide (the Gaviota slide) occurring in the past 6 ka.Although previous work suggests possible links between Pleistocene slide occurrence and lower sea level due to increased seismicity rates (Brothers et al., 2013), it is also possible that higher Pleistocene lowstand sedimentation rates, especially near the coastline, are a factor in preconditioning slopes for failure in the later Pleistocene and Holocene.
Our results underscore the need for better age control offshore of southern California to improve hazard assessment.More detailed studies of seismic stratigraphy aimed at establishing slide chronology, frequency, and recurrence (e.g., Del Mar slide area, Figure 5c) are required to fully understand the risk associated with regional submarine landslide hazards.Even for the Quaternary failures documented in this study, relatively few Borderland slides have been dated, and underlying regional sedimentary unit ages have not been reassessed since the late 1980s (Clarke et al., 1987;McCulloch, 1989;Vedder et al., 1986Vedder et al., , 1987)).Additionally, the ages we have for observed failures may well be a result of observational, sampling, and/or dating bias (e.g., Urlaub et al., 2014).
For instance, Holocene highstand conditions may serve to limit the erosion and/or burial of late Pleistocene and Holocene slides by subsequent sedimentation, making them observable with seabed imaging techniques.Furthermore, the prevailing dating methods (radiocarbon) favor sampling targets that are likely to be within the age range covered by that method (less than ∼50 ka).
To date, our work is the most complete and methodical regional compilation and mapping effort of Quaternary mass wasting features offshore southern California (Papesh et al., 2023) and has allowed us to identify compelling relationships between active faults, sediment supply, slope, and slide size.However, numerous questions remain about Borderland slide processes.For instance, regional shear strength analysis, sediment thickness, links between specific sediment sources and slide deposits, and quantitative comparisons between slope failure and seismicity have yet to be examined.We make inferences about sediment thickness and shear strength based on proximity to sediment sources and Quaternary faults, but do not directly measure or model these parameters.Additionally, processes related to subsurface fluids, gas, and seeps, which are prominent in several regions including Santa Barbara Basin (Kluesner et al., 2020), were not analyzed in this study and may provide further insights into the slide preconditioning and triggering processes.Finally, while several studies have noted the tsunami risk from Borderland slides (Borrero et al., 2004;Brothers, Andrews, et al., 2019;Brothers, Maier, et al., 2019;Fisher et al., 2005;Greene et al., 2006;Legg & Kamerling, 2003;Lee et al., 2009;K. J. Ryan et al., 2015), few have been quantitatively modeled, including the newly discovered San Nicolas slide.The slide characteristics determined in our compilation could be leveraged for future modeling studies aimed at regional geologic characterizations and as input for three-dimensional reconstruction (e.g., Alberti et al., 2022) or machine learning workflows to better understand and quantify controls on submarine landslide processes.

Conclusions
1. We provide an analysis of a new compilation of submarine landslide feature mapping offshore of southern California (Papesh et al., 2023) that includes detailed attributes such as slide size, type, geology, failure slope, failure age, and key metrics such as distance from the shoreline and distance from Quaternary faults.2. Our study provides previously unpublished observations, morphometrics, ages, and constraints on some of the largest slides in the Borderland region, including the San Nicolas slide, the Catalina Basin slides, the Del Mar slide, and the San Gabriel slide complex.3. Our results suggest that most slides in the Borderland region are controlled by three primary factors: (a) significant sediment deposition on low slopes, (b) uplift and oversteepening due to regional tectonics, and (c) triggering from earthquakes on Quaternary faults.
4. Repeated, small-scale (area <20 km 2 ), earthquake-triggered failure of Pleistocene sediment dominates slide processes in steep, shelf-edge settings.Long-term sediment accumulation on low slopes, followed by tectonic uplift, preconditions many larger failures of Plio-Pleistocene and older sediment in open-slope environments.5. Larger slide size loosely correlates with increased Quaternary fault distance.Farther from faults, less-uplifted slopes and/or infrequent earthquake shaking (i.e., longer recurrence intervals between significant earthquake triggers) may promote sediment accumulation, thereby preconditioning large failures.6.Only one failure age in our study (of n = 12 total ages) is in the past 6 ka, perhaps suggesting that present-day risk from submarine landslides offshore southern California is relatively low; however, our study also emphasizes a need for more age control for large slides and detailed recurrence studies throughout the region to better understand slide and tsunami risk.7. Our work provides a foundation for numerous potential follow-on studies aimed at better quantifying slope stability and failure preconditioning processes offshore southern California.Regional shear strength, sediment thickness, seismic modeling, tsunami modeling, subsurface fluids/gas, and age control studies are timely and would build off of our results to advance our understanding of submarine landslides offshore California and globally.

Figure 2 .
Figure 2. (a) Rotated location figure showing mapped submarine landslides features (as in Figure 1; mwz = mass wasting zones) and Quaternary faults from Walton, Papesh, et al. (2020; thin black lines) atop a background of semi-transparent bathymetric color overlaid on a grayscale slope gradient map.Locations of Ocean Drilling Program (ODP) Leg 167 sites 1013 and 1015 and ODP Leg 146 Site 893 are also highlighted in green.(b) Regional offshore geology(Clarke et al., 1987;McCulloch, 1989;Vedder et al., 1986Vedder et al., , 1987) ) with mapped discrete submarine landslides (polygon outlines colored by slope type) and Quaternary faults from Walton, Papesh, et al. (2020; thin black lines).Extents and rotation similar to part a. Q-Quaternary sedimentary rock; Qsp-San Pedro Formation; QTs-Quaternary and Tertiary sedimentary rock; QTt-Quaternary and late Tertiary sedimentary rock; Tp-Pliocene sedimentary rock; Tpr-Pliocene and late Miocene sedimentary rock; Tm-Miocene sedimentary rock; Tmv-Miocene volcanics; Tmu-undifferentiated Miocene volcanic and sedimentary rock; Tmp-Miocene plutonic and hypabyssal rock; Te-Eocene sedimentary rock; Tep-Eocene and Paleocene sedimentary rock; Ku-Late Cretaceous sedimentary rock; TMz-undifferentiated Miocene igneous rock and pre-Late Cretaceous metamorphic rock; Mz-pre-Late Cretaceous metamorphic rock; m-unknown rock type; gr-Mesozoic granitic rock.

Figure 3 .
Figure 3. San Nicolas slide.(a) Enlarged slope gradient map of the San Nicolas slide with a semi-transparent bathymetric color overlay (extent shown in Figure 1).Map shows the locations of Chirp and MCS lines highlighted in panels (a and b), as well as the location of Ocean Drilling Program (ODP) Leg 167 Site 1013 (Shipboard Scientific Party, 1997).(b) MCS line 124 from USGS survey L490SC (https://walrus.wr.usgs.gov/namss/survey/l-4-90-sc/)crossing the San Nicolas slide.Image highlights a slide headwall scarp and downslope slide debris deposit.The location of the seismic line is shown in panel (a).(c) Left plots show smoothed log curves (10point moving average) of multisensor track (MST) data (density, natural gamma ray), porosity data (from index properties), and the age model from ODP Leg 167 Site 1013 (Shipboard Scientific Party, 1997).Colored lines tie the Site 1013 data to Chirp data from EW9709 (Rolling Deck to Repository, 2015) and SR1701 (Rolling Deck to Repository, 2017), linking the age model and sediment properties to the San Nicolas slide debris and the drape layer capping the debris.The locations of the Chirp data are shown in panel (a).

Figure 4 .
Figure 4. Catalina Basin slides.(a) 2016 R/V Thompson Chirp data (Balster-Gee et al., 2020) detailing the sidewall scarp, debris, and ∼5 m drape layer topping the slide debris.Inset shows an enlarged bathymetric image of the Catalina Basin slides (extent shown in Figure 1) which includes locations of Chirp and MCS data highlighted in panels (a and b).(b) Highresolution MCS data from a 2016 R/V Thompson survey (Balster-Gee et al., 2020) and a 2014 R/V Sproul survey (Balster-Gee et al., 2017) in Catalina Basin crossing a slide sidewall scarp and debris lobe.The locations of MCS lines are shown in panel (a) inset, location of Chirp data shown in 4a is superimposed on the coincident MCS data.CF-Catalina fault; SCF-San Clemente fault.

Figure 5 .
Figure 5. Del Mar slide.(a) Enlarged bathymetric image of the Del Mar slide area (map extent shown in Figure 1) showing the extent of the slide as mapped from seismic reflection imaging.Map also shows the location of panels (b and c), and core SD-1 shown in panels (b and d).(b) Chirp data from a survey crossing the Del Mar slide (extended trackline location labeled as OS-39 in panel (a); Sliter et al., 2017a).Image shows the Carlsbad fault zone (also shown on panel (c)) and the Del Mar slide debris topped by an acoustically transparent drape layer.Core SD-1 is enlarged in panel (d).(c) High-resolution MCS data (trackline location labeled as OS-39 in panel (a); Sliter et al., 2017b) crossing the Del Mar slide and showing multiple mass-transport deposits (MTDs) in the subsurface, as well as the Carlsbad fault zone which is enlarged in panel (b).(d) Enlarged Chirp image from panel (b) showing core SD-1 (map location shown on panel (a)) with radiocarbon ages in calibrated years before present (cal yr BP); ages are also shown in Table2.

Figure 6 .
Figure 6.San Gabriel slide complex.(a) Enlarged bathymetric image of the San Gabriel slide complex area (extent shown in Figure 1) showing the extent of the main slide as mapped from seismic reflection imaging.Map also shows the location of seismic profile OS-134 shown in panel (b).(b) High-resolution MCS data(Sliter et al., 2017c) crossing the debris field area of the San Gabriel slide complex.MCS data reveal evidence for several buried mass-transport deposits (MTDs) in the subsurface.

Figure 7 .
Figure 7. Statistical analyses of submarine landslides.(a) Histogram of slide feature proximity to the modern coastline.Blue bars show headwall scarps (binned at 1 km) and pale orange bars show discrete slides (binned at 10 km) from Papesh et al. (2023).The data show that over 3/4 (78%) of these slide features occur within 20 km of the modern coastline.(b) Histogram of slide feature proximity to the 130 m depth contour from the California coastal relief model(National Geophysical Data Center, 2012), an approximation for the lowstand Last Glacial Maximum (LGM) coastline (e.g.,Chaytor et al., 2008;Lambeck et al., 2002;Peltier, 2002;Shackleton, 2000;Yokoyama et al., 2000).Blue bars show headwalls (binned at 1 km) and pale orange bars show landslides (binned at 10 km) fromPapesh et al. (2023).The data show that 88% of these slide features occur within 20 km of the approximate LGM coastline.

Figure 9 .
Figure 9. Scatter plot showing slide failure slope versus slide area for the 63 discrete slides mapped in our dataset (Papesh et al., 2023).Major slides discussed in the text are labeled and different slide symbols denote slope environments.Results indicate that the largest slides occur on slopes of <10°, with smaller slides occurring on a much wider range of slopes.

Figure 10 .
Figure 10.Scatter plot showing the proximity of slides to Quaternary faults(Walton, Papesh, et al., 2020) versus slide area for the 63 discrete slides mapped in our dataset(Papesh et al., 2023).Slides plotted to the right and left of the vertical dotted line at 20 km 2 represent the largest 25% and smallest 75% of slides by area, respectively.Slides plotted above and below the horizontal dashed line at 10 km represent slides more than (27%) and less than (73%) 10 km distance from Quaternary faults.The largest slides discussed in the text are labeled and different slide symbols denote slope environments.Results indicate that the majority of the largest 25% of slides occur relatively far from Quaternary faults in predominantly open-slope environments, differing from the distribution of the rest of the slide features.
large, infrequent failures along uplifted basin flanks are characteristic of open-slope mass wasting environments offshore southern California (Figures 2b, 9, and 10), which host the other broad category of submarine landslides in our study.Examples of open-slope translational failures include the San Nicolas slide, the Catalina Basin slides, and the Santa Cruz Basin slides (Figures 1 and

Table 1
Summary of Attribute Information Included for Each Mapped Slide Feature The mean slope, associated uncertainty, and gradient measured from MBES just outside the headwall area (landslides) or the slope within the mapped area (mass wasting zones) Landslides, mass wasting zones Depth range dep_min_m, dep_max_m Minimum and maximum depths of the feature (from the NCEI CRM) Landslides, headwall scarps Proximity to coast prox_coast Proximity of the feature's evacuation zone to the NCEI CRM coast (0 contour) Landslides, headwall scarps Proximity to 130 m depth prox_130m Proximity of the feature's evacuation zone to the 130 m depth contour extracted from the NCEI CRM Landslides, headwall scarps Proximity to faults prox_qflts Proximity of the feature's evacuation zone to Quaternary faults in Walton, Papesh, et al. (2020) Landslides, headwall scarps Data constraints data_type Primary data type(s) and resolution used for mapping Landslides Slide type slide_type Categorization of landslide type following Varnes (1978) and Hungr et al. (2014) Landslides Slope type slope_type Categorization of slope type where the failure occurred (open slope, shelf, shelf edge, or canyon/channel) , and the Del Mar slide (this study; see Section 4.3.1).Other submarine landslide chronologies were derived using a combination of stratigraphic mapping from acoustic imaging, sedimentation rate data, and radiometric, biostratigraphic, or magnetostratigraphic dating techniques from Ocean Drilling Program (ODP) data.Three drilling program sites exist in our region of interest offshore southern California; in this study, we analyze data from ODP Expedition 167 Site 1013(Shipboard Scientific  Party, 1997)in San Nicolas Basin (Figures2a and 3) to provide new timing constraints on the San Nicolas slide (see Section 4.2.1).We also use a combination of sedimentation rate data and seismic imaging to newly constrain ages for n = 2 slides in Catalina Basin (see Section 4.2.2),similar to methods previously used to determine slide ages in Santa Barbara Basin