Corresponding author: C. Sorlien, Earth Research Institute, University of California, Santa Barbara, 93106, CA, USA. (email@example.com)
 The Palos Verdes anticlinorium (PVA) is a 70 km-wide, NW-trending Pliocene-Quaternary transpressional high that separates the onshore Los Angeles basin from the offshore San Pedro basin. The PVA southwest limb underlies a 70 km-long SW facing seafloor escarpment. This escarpment is separated into two segments by a 6 km right step. It overlies two levels of NE dipping low-angle thrust faults. These faults project beneath Los Angeles toward known SW verging blind thrust faults and may be the same regional faults. These conclusions are based on a detailed geometric representation constructed from a regional analysis of 5000 km of multichannel seismic reflection profiles, logs, and paleontology from 19 wells, published seafloor geology, and multibeam bathymetric data. Patterns of sedimentation and erosion indicate that folding of PVA initiated during Pliocene time and propagated southward, with folding of part of the southern PVA occurring during the Quaternary. Structural relief continues to grow with regional long-term rock uplift of the PVA crest and ongoing subsidence in the adjacent basins. The low-angle NE-dipping faults beneath the PVA must be late Quaternary active in order to maintain hanging wall rock uplift above subsiding footwalls. Structural geometry requires that the underlying faults have slipped at 1.1–1.6 mm/yr over Quaternary time.
 The Los Angeles metropolitan region features more than 5 km of post-Miocene structural relief, from subsiding basins to uplifting ridges (Figure 1), as is characteristic of continental borderlands forming along diffuse transform plate boundaries. Many of the uplifts adjacent to Los Angeles are anticlinoria, complex long-wavelength anticlines with subsidiary folds, which are caused by (oblique) thrusting on faults beneath them [e.g., Davis et al., 1989]. Some of the more internal ones have formed mountains, while the external ones are submarine ridges and islands. As is common with continental tectonics, preexisting faults have been reactivated, with paleohighs now corresponding to subsiding basins and paleobasins inverting into anticlines [Yeats and Beall, 1991; Bohannon et al., 2004; Sorlien et al., 2006].
 One of these structures, the 70 km-long Palos Verdes anticlinorium (PVA), separates the submarine San Pedro basin from the deeper but sediment-filled Los Angeles basin onshore. The PVA structural high is mostly close to sea level along the shelf and is mostly drowned under sediment or water, except for its emerged culmination at the Palos Verdes Hills (Figure 1b). This partial morphological expression may explain why structural and seismotectonic investigations have focused on secondary elements of the PVA but have tended to overlook the structure as a whole. Until industry multichannel seismic reflection data were made available by the U.S. Geological Survey [Hart and Childs, 2005], the offshore part of PVA was difficult to study because there are no wells in San Pedro Basin and wells in Santa Monica Bay are few and drilled in the 1950s and 1960s. Although this major structure is directly relevant to earthquake hazards in Los Angeles, its extent, activity, and causative faults remain controversial [Nardin and Henyey, 1978; Davis et al., 1989; Ward and Valensise. 1994; Shaw and Suppe, 1996].
 Competing sets of models have been proposed to account for the PVA. The first set proposes a series of distinct anticlinoria associated with right-lateral faults. The prominent Palos Verdes Hills, for example, are thought to be the result of local uplift related to a restraining segment of the right-lateral Palos Verdes fault [Nardin and Henyey, 1978; Ward and Valensise, 1994; Legg et al., 2004a] (Figure 2a). A second set of models recognizes the PVA as a single regional structure and accounts for it by blind thrust faulting (Figures 2b, 22c, and 3). One long-standing hypothesis couples the PVA with a SW-dipping roof thrust [Davis et al., 1989; Shaw and Suppe, 1996]; (Figures 3b and 3c). An alternative hypothesis has the PVA rooted on a NE-dipping thrust-reactivated normal fault system (Figure 3a) [Broderick, 2006]. In contrast to the thrust models, a SW-dipping normal fault, parallel to the right-reverse Palos Verdes fault, was proposed to explain the offshore structural relief along San Pedro Escarpment [Bohannon et al., 2004]. Distinguishing among these models has implications for maximum earthquake size, ground motion modeling, and tsunami generation for the Los Angeles metropolitan area. We test these models by documenting the structure of the PVA in unprecedented detail. The data show the PVA to be a single large continuous structure and supports the existence of two structural levels of a regional NE-dipping thrust system rooted beneath Los Angeles
 This study analyzes over 5,000 km of multichannel seismic reflection profiles, logs and paleontology from 19 offshore petroleum exploration wells, seafloor and land geology, and multibeam bathymetric data (Figure 1d). This analysis entails reconciling the extensive geological and geophysical data set available from industry, U.S. Geological Survey (USGS), the Bureau of Ocean Energy Management (BOEM, formerly the Minerals Management Service), and academia. When combined, these data afford an unprecedented view of the PVA, both detailed and synoptic, and provide a powerful basis to test structural growth models. Critical for our analysis is also the submarine growth sedimentation on the PVA flanks. Our strategy is to first assess structural growth by documenting rock uplift of the PVA, relative to rock subsidence of the flanking basins. We can thus identify the PVA as a single regional shortening structure that grew coherently during Pliocene-Quaternary time. Then we compare this observed growth with the deformation expected from known and possible faults subparallel to the PVA, according to the differing models. We consider both thrust motion on faults dipping NE or SW and also oblique-reverse motion on steep dextral faults. We find no evidence for a large SW-dipping normal fault, and such an interpretation does not explain the regional SW-dipping fold limb beneath San Pedro Escarpment (Figure 1). Our conclusion is that the PVA stems from both strike-slip and blind thrust faults, and our preferred model includes NE dipping basal low-angle (oblique) thrusts at two levels rooted below the Los Angeles Basin. These faults are blind but reach close to the seafloor beneath Santa Monica and San Pedro basins. The Palos Verdes fault contributes structural relief from oblique right-reverse slip on its onshore and inner shelf part, but little relief elsewhere.
2 Tectonic and Stratigraphic Evolution
 The southwestern California margin has undergone a complex and diverse tectonic evolution that controlled the present distribution and types of rocks. The latest of a series of reorganizations led to transpression reactivating Miocene normal and oblique normal faults in the study area. Therefore, we focus here on Neogene tectonics, denudation of metamorphosed subduction rocks, volcanism, and sedimentation.
 Crustal block rotation, tectonic denudation from extreme extension, fault reactivation, and basin inversion have all occurred during the Neogene within the greater Los Angeles area. These events are consistent across the margin of southwest California and derive from Neogene plate tectonic reorganizations. Subduction of the Farallon Plate and associated microplates ceased during early Miocene time, as spreading centers died out offshore or were subducted beneath the western continental margin [Lonsdale, 1991; Nicholson et al., 1994; Wilson et al., 2005]. Resulting slab windows produced voluminous early and middle Miocene volcanism (18–14 Ma) in the western Transverse Ranges, Los Angeles basin, and Inner California Continental Borderland [Luyendyk et al., 1998; Wilson et al., 2005].
 Regional major extension and oblique extension followed subduction as the Pacific Plate diverged from the North American plate during Miocene time [e.g., Atwater and Stock, 1998]. This extension facilitated over 90° of vertical axis clockwise rotation of the western Transverse Ranges crustal block since ~19 Ma, from a position alongside the modern coastline between Los Angeles and San Diego, to its current position as an east-west belt of faults and folds located north and northwest of Los Angeles [Kamerling and Luyendyk, 1979; Hornafius et al., 1986; Luyendyk, 1991]. This extension resulted in extensive tectonic denudation and the emplacement of metamorphic basement rocks, the Catalina Schist, at shallow crustal levels [Yeats, 1976; Crouch and Suppe, 1993; Bohannon and Geist, 1998; ten Brink et al., 2000] Displacement on certain of these large displacement low-angle faults may have been oblique-right-normal slip [Legg et al., 2004b]. One of the type areas for such emplacement offshore is the Catalina tectonostratigraphic terrane, where Miocene marine sedimentary and volcanic rocks directly overlie the Catalina Schist (Figure 4). This terrane makes up much of the California Inner Continental Borderland, including the PVA study area [e.g., Howell and Vedder, 1981].
 Crustal thinning leads to subsidence, and deep marine middle and late Miocene sedimentary rocks overlie shallow marine or nonmarine sedimentary rocks or a major unconformity throughout large parts of the southwest California margin [Dibblee, 1966; Vedder, 1987; Wright, 1991; Lee, 1992]. In the Los Angeles region, the marine middle to upper Miocene Monterey Formation and the time-equivalent Puente Formation were deposited regionally within a broad depocenter [Wright, 1991]. On the coastal edge of the Los Angeles Basin, Monterey Formation rests on Catalina Schist basement, separated in places by a thin Miocene schist conglomerate or sandstone (Figure 4) [Blake, 1991; California Division Oil, Gas, and Geothermal Resources, 1992]. Jet and dart cores acquired across the Shelf Projection and inner San Pedro Shelf and onshore geologic mapping indicate middle to upper Miocene rocks (Relizian through Mohnian foraminiferal stages) outcrop near the crest of PVA (Figures 1a and S1) [Nardin and Henyey, 1978; Dibblee, 1999]. The Miocene outcrop on San Pedro Shelf is flanked by Pliocene rocks downplunge to the southeast along the crest of the PVA (Figure 1a).
 The rate of extension between the Pacific and North American plates decreased during late Miocene time because of a gradual Pacific Plate motion change [Cande et al., 1995; Atwater and Stock, 1998]. By the end of the Miocene, the plate boundary shifted into the Gulf of California, thus assuming the current shape of the San Andreas Fault. Misalignment between the San Andreas Fault in southern and central California resulted in a regional restraining double bend, called the Mojave segment or Big Bend [Crowell, 1979; Atwater, 1989]. Associated contraction initiated at the beginning of Pliocene time for most, if not all, of the western Transverse Ranges, as well as areas northwest, southwest, and south of this province [Clark et al., 1991; Schneider et al., 1996; Sorlien et al., 1999, 2006]. As a result, many preexisting normal separation faults have been reactivated as oblique, reverse, and thrust faults. This reactivation has, in some cases, caused basin inversion, whereby thick Miocene sections became uplifted into present-day mountains, islands, and submarine banks [Yeats and Beall, 1991; Schneider et al., 1996; Seeber and Sorlien, 2000; Sorlien et al., 2006; Brankman and Shaw, 2009; Schindler, 2010]. These widespread basin inversions record not only the reversal from extension to shortening but also the tendency for current thrust faults to reactivate preexisting normal faults rather than cut through them.
 The onset of transpression near the beginning of Pliocene time increased relief and the supply of terrigenous sediments. The Pliocene “Repetto” and upper Pliocene–lower Pleistocene “Pico” intervals are distinguished by benthic foraminiferal stages, not rock type, and are members of the Fernando Formation [Blake, 1991; Yeats, 2004]. While locally Repetto and Lower Pico might be identified as patterns on electric logs from wells, they are generally calibrated to the Repettian and Venturian benthic foraminiferal stages [Blake, 1991]. These stages are known to be time transgressive in California [Ponti, 1989]. These stages were calibrated to diatoms and other micropaleontology by Blake  for several sections around Los Angeles basin. The most applicable sections for our study are from near Palos Verdes Hills and also Newport Bay. Top Venturian Stage is 1.95 ± 0.15 Ma, top Repettian Stage is ~2.5 Ma, and top Delmontian Stage (=base Repettian) is 5.0 ± 0.2 Ma (Newport Bay) or 4.42 ± 0.57 Ma (Malaga Cove in Palos Verdes Hills) [Blake, 1991]. On San Pedro Shelf northeast of the Palos Verdes fault, we interpret the top Lower Pico of Wright  and Rigor , rather than the deeper top Venturian Stage given in the paleontology reports listed in supporting information Appendix 1. We have correlated our top Lower Pico to correlate with “Pliocene B” of Ehman and Edwards  on seismic reflection profiles just outside the outer breakwater entrance of Los Angeles harbor. Pliocene B is “middle Long Beach” in McDougall et al. , which is >1.76 Ma from magnetostratigraphy and could be as old as 2.4 Ma from paleontology. We will use the Blake  1.95 ± 0.15 Ma age for the Wright  top Lower Pico in this paper. This is discussed in more detail in supporting information Appendix 1.
 Quaternary strata include and overlie the Lower Pico and include rapidly deposited turbidites that cover the floor of Santa Monica basin and that onlap its flanks. A major onlap surface and unconformity beneath Santa Monica Bay was proposed to date from Oxygen Isotopic Stage 10 [Normark et al., 2006], which peaked at 341 ka [Lisiecki and Raymo, 2005]. This age estimate was based on extrapolation of post ~ 32 ka sedimentation rates at ODP site 1015 and seismic stratigraphy (located on Figure 1) [Shipboard Scientific Party, 1998; Normark et al., 2006]. This “Hueneme” unconformity is deepest in western Santa Monica sedimentary basin, with onlap of sediments of Hueneme Fan toward the east [Normark et al., 2006; Fisher et al., 2005]. In the deepest part of the Quaternary Santa Monica sedimentary basin, our Pico unconformity is ~200 m below Hueneme unconformity.
 A figure in Normark et al.  (their Figure 17 of USGS-57) interprets time thickness of 0.2 s two-way travel time or more than 150 m of post-125,000 B.P. sediment in northern San Pedro Basin. This horizon is our “Pico unconformity” (also Figures 4-6, S2, and S3). Our “Hueneme unconformity” is just 30–40 m above the Pico unconformity in this area, and we suggest that strata between the two unconformities and below the lower one are much older than 125 ka. The Pico unconformity was also interpreted by Sorlien et al. , who suggested it might correlate to a 800–1000 ka unconformity in northern Los Angeles basin [Tsutsumi et al., 2001; Hummon et al., 1994]. A detailed seismic stratigraphic interpretation of Santa Barbara Channel indicates that sediment supplying Hueneme Fan may have reached Santa Monica basin by about 800 ka rather than ~330 ka [Marshall, 2012]. Our correlation through marginal seismic reflection imaging is permissive of the Pico unconformity being top Middle Pico at wells 5 and 6 (Figure 1, supporting information Appendix 1) [Sorlien et al., 2006]. The 975 ± 75 ka horizon [Yeats, 1981, 1983; Huftile and Yeats, 1995] in Santa Barbara Channel is close to top Middle Pico there [Marshall, 2012]. The strata between these two unconformities thin, but do not pinch out to the east. Strata beneath our Pico unconformity in northern San Pedro Basin may be as old as 1000 ka, and the strata above the Hueneme unconformity are time transgressive and could be locally younger than 125 ka (Figure 4).
 Over 5000 km of mainly industry and USGS multichannel seismic reflection (MCS) data are analyzed to produce 3-D digital fault and stratigraphic horizon representations covering the study area in Figure 1 [Hart and Childs, 2005; Sliter et al., 2005]. Logs from 19 wells and paleontological reports from BOEM and other sources provided stratigraphic control (supporting information Appendix 1). Sea floor maps utilizing hundreds of jet and dart cores published by Nardin and Henyey  (Figure 1a) constrained interpretations on the shelf. Another published seafloor geologic map indicates Catalina Schist and Miocene volcanic rocks are exposed on Redondo Knoll and the slope northeast of Catalina Island [Vedder et al., 1986]. Large displacements on regional-scale low-angle Miocene normal faults are suggested in order to tectonically denude the Catalina Schist exposed in their footwalls [Crouch and Suppe, 1993; Bohannon and Geist, 1998]. Seafloor data from NOAA point bathymetry and USGS and other high-resolution multibeam bathymetry [Dartnell and Gardner, 1999] constrain the location of seafloor-cutting faults and folds (Figures 1b, 1c, and S1). Onshore fault surface representations from the Southern California Earthquake Center (SCEC) Community Fault Model complement the new fault surfaces and the structural model produced here (Figures 1a and 1c) [Plesch et al. 2007]. Much additional subsurface published geologic information from onshore was incorporated (Figure 7) [e.g., Wright, 1991].
 Relocated earthquake hypocenters for Santa Monica Bay and adjoining areas [Hauksson, 2000] appeared scattered and were not clearly defining fault surfaces when visualized in 3-D [Sorlien et al., 2006]. Diffuse seismicity also characterizes the southern part of the study area [Brankman and Shaw, 2009], and thus seismicity was not used in the fault interpretations.
3.2 Subbottom Mapping
 A central contribution of this work are digital 3-D surfaces representing faults and stratigraphic horizons, which are primarily derived from the abundant seismic reflection data located on Figure 1d. The dense grids of seismic reflection data include numerous strike and dip profiles. Careful loop tying was systematically followed at all intersections between profiles and lends a high degree of confidence to the horizon maps. Limitations include reliance on industry picks for top Repetto and top Lower Pico in Santa Monica Bay, near pinchouts of top Miocene onto acoustic basement, and the known time-trangressive nature of post-Miocene benthic foraminiferal stages [e.g., Blake, 1991]. Bottlenecks in regional correlations include the narrow basin between Redondo Knoll and Palos Verdes Hills and poor imaging in places near wells 5, 6, and 8 (Figure 1a). The Lower Pico and Repetto onlap older rocks beneath the base of San Pedro Escarpment and cannot be directly correlated into northern San Pedro Basin across a narrow region (Figures 8 and 9). A Miocene–post-Miocene unconformity is interpreted in place of base Repetto in northern San Pedro Basin. Miocene at the seafloor is known from wells and shallow cores on the shelf of northwest PVA (Shelf Projection; Figure 1), and this unconformity is chosen above the known Miocene. Difficult correlations from northern wells 3 and 6 through the bottlenecks of near onlap support our interpretation of approximate top Miocene. These various difficulties likely result in a lack of precision in the ages of horizons in these areas but do not significantly affect our conclusions.
 Fault interpretation criteria included reflections from gently and moderately dipping faults and truncations or offsets of reflections from stratigraphic horizons (Figure 5). The geometries of gently or moderately dipping faults were additionally constrained by overlying folds. Thrust separation faults were interpreted at the transition between flatter, less deformed footwall reflections, and folded reflections. The updip limits of the blind faults were interpreted where vertical separation of stratigraphic reflections across faults could not be resolved.
 Stratigraphic horizons were identified from well logs (“mud logs,” describing rock cuttings), from paleontological reports, published cross sections, and published information on seafloor outcrops (supporting information Appendix 1) [Nardin and Henyey, 1978; Vedder et al., 1986; Wright, 1991; Rigor, 2003]. The seafloor sample locations of Nardin and Henyey  are located on Figure 1a and are color coded for age of rocks. Sonic logs or check shot surveys from wells were used to convert from depth to two-way travel time and, conversely, to convert migrated industry seismic reflection profiles to depth [Rigor, 2003; Sorlien et al., 2006; this study]. Stratigraphy was correlated across the Palos Verdes fault using wells with BOEM paleontology and velocity surveys on either side of the fault, using dated seafloor outcrop [Nardin and Henyey, 1978] (Figure 1), and locally by seismic stratigraphic correlation (Figure 9).
 Velocity control from wells is lacking over the offshore slopes and basins (Figure 1a). Therefore, two-way travel time grids of faults and stratigraphic horizons were converted to depth using 1490 m/s for the water layer, seafloor sedimentary rock velocity of 1723 m/s in the north and 1500 m/s in the south, and a velocity function with a linear increase in velocity with depth (e.g., linear velocity equation [Sheriff, 1973]). The slope of the velocity gradient (constant 0.6) is appropriate for normally compacted terrigenous sediments [Tolmachoff, 1993]. We expect errors in depth below seafloor near 10%, which are comparable to what is expected using stacking velocity analysis of MCS data [Hughes, 1985].
 The offshore digital surface presented in Figure 7 is based on interpretation of a horizon near or just above the top of Miocene on the seismic reflection profiles located on Figure 1d. The interpreted points were gridded using “Projection of Slopes” in the Surface III software (Kansas Geological Survey, 2001). Numerous industry wells were used by Wright  to produce a contour map of onshore base Repetto; over 300 wells are listed in the Wright  appendices. We digitized this contour map and merged it with our offshore data to complete a surface map across the PVA (Figure 7). The ~4.5 Ma base Repetto [Blake, 1991] was deposited about the time of initiation of transpression in and near northern Los Angeles basin [Wright, 1991].
4 Observations: Structural Setting of Palos Verdes Anticlinorium
 Palos Verdes anticlinorium (PVA) is affected by gently-to-moderately-dipping (oblique) thrust reactivated (oblique) normal faults by steeply dipping right-lateral faults and by the folding associated with displacements on both types of faults. Bends bounding restraining segments on less steep right-lateral faults can result in oblique slip and vertical motions. Where the right-lateral faults are closer to vertical, the convergent component can be instead partitioned or partly partitioned onto the more gently dipping faults, creating the broader anticlinorium. Understanding of these motions must also include recognition of regional subsidence resulting from Miocene crustal thinning and sediment and structural loading.
 A pair of clearly imaged north to northeast-dipping Miocene basin-bounding low-angle normal faults underlies part of Santa Monica basin and San Pedro basin. Both faults are known to have Catalina Schist in their footwalls [Smith, 1897; Vedder et al., 1986]. These low-angle normal faults are less obvious where they have been partly overridden by the PVA and disrupted by the high-angle San Pedro Basin fault (Figures 5, 6, S2, and S3). The low-angle faults transition from thrust-reactivated at depth, to shallower preserved Miocene normal faults, to eroded fault scarps onto which postextension strata onlap the footwalls with buttress unconformity. Their traces are mapped near the upper edge of the preserved Miocene faults. Figure 5 displays seismic reflection profiles across the detachment faults in areas updip and to the northwest of where they have been significantly reactivated. Their history as regional-scale large displacement Miocene faults is consistent with them continuing far to the northeast along their downdip projections. Plio-Quaternary regional-scale folding logically reactivates the deep part of these faults as (oblique) thrusts.
 The Santa Monica Bay detachment is a prominent gently north dipping fault that juxtaposes known Catalina Schist on Redondo Knoll in its footwall [Vedder et al., 1986] against inferred early and middle Miocene rocks in its northern hanging wall (Figures 5-7) [Sorlien et al., 2006; Broderick, 2006]. Gently-to-moderately NE dipping faults, the San Pedro Escarpment faults, are in the hanging wall of the Santa Monica Bay detachment (Figures 1a, 1c, 5, 6, S3, and S4). These faults project to intersect the Santa Monica Bay detachment in the upper couple of km. The San Pedro Escarpment faults die out into the syncline south of Palos Verdes Hills, beneath San Pedro Shelf (Figures 1 and 7).
 The newly recognized Catalina Island detachment is imaged on J-J', K-K', and L-L' (Figure 5). The true dip adjacent to J-J' and K-K' is respectively 21º and 27º, based on depth conversions using the 1-D velocity model. Blue schist–bearing Catalina Schist on Catalina Island is in its footwall [Smith, 1897]. Reflection character on the profiles on Figure 5 is consistent with Catalina Schist underlying both the Catalina Island detachment and the Santa Monica Bay detachment. The mapped upper edges of these faults are about 10–15 km apart, with the Catalina Island detachment at a lower structural level than the Santa Monica Bay detachment. They do not project to intersect in the upper few kilometers of crust. Both faults floor half graben wedges of pre-Pliocene sedimentary rocks (Figure 5).
 Additional NE dipping faults are imaged above the Catalina Island detachment beneath the lower slope of southern San Pedro Escarpment (Figures 8 and 9). They bound subbasins for the likely Miocene interval between top Catalina Schist and ~4.5 Ma base Repetto and are thus Miocene normal separation faults. Pliocene (and probably late Miocene) strata pinch out by onlap onto acoustic basement in the footwall side of these faults (Figures 8 and 9). Similar thin or missing late Miocene and Pliocene strata are seen in the footwall side of the Santa Monica Bay detachment (Figures 5, 6, and S2).
 Transpression or contraction is reactivating these Miocene normal separation faults, with 2–4 km wavelength folding deforming the seafloor of southern San Pedro Basin (Figures 8-11). Viewed more broadly on Figure 7, the PVA is overriding the regional low angle normal faults, with the 10 km right step over of its southwest limb being related to a north-south inferred tear fault (also Figure 1c). The southwest limit of MCS data coverage between L-L' and Q-Q' prevents complete imaging of the upper Catalina Island detachment (Figures 1d and 8). However, LARSE profile 1R in Figure 10 does image the shallow geometry of the Catalina Island detachment and NE dipping faults beneath the lower San Pedro Escarpment, as well as the intersecting San Pedro Basin fault.
 Two steeply dipping, NW-striking faults, the San Pedro Basin fault and Palos Verdes fault, affect parts of the PVA. Local changes in strike of the San Pedro Basin fault create double bends that cause deformation at and below the seafloor consistent with right-lateral slip (Figures 1, 5, and 6). Mixed reverse and normal separation are imaged on single seismic reflection profiles (Figures 6 and S2) with east side-up separation of the seafloor, Quaternary, and Pliocene horizons a mere 2 km northwest of where the separation is down to the east [Broderick, 2006]. An offset submarine channel east of Catalina Island gives a right-lateral slip rate of 1.5 ± 0.3 mm/yr since 12,270 BP [Ryan et al., 2012].
 The Palos Verdes fault is well defined across San Pedro Shelf and in the subsurface of the mainland northeast of Palos Verdes Hills (Figures 7-9, 11, and S4) [Nardin and Henyey, 1978; Wright, 1991; Fisher et al., 2004; Brankman and Shaw, 2009; Francis and Legg, 2007]. Offset of an ~8 ka channel in Los Angeles harbor near Long Beach demonstrated 2.7–3.0 mm/yr right-lateral slip on the Palos Verdes fault [McNeilan et al., 1996]. It continues northwest into Santa Monica Bay, where it splits into several strands with minor vertical separation of top Repetto [Broderick, 2006]. Strands of both the Palos Verdes fault and the San Pedro Basin fault largely die out before intersecting the W-striking Santa Monica–Dume fault (Figure 1) [Sorlien et al., 2006].
 The axial trace of the PVA is continuous for 70 km (Figures 1 and 7). Its southwest limb is linear and continuous for 45 km, right-stepping 10 km across a syncline beneath San Pedro Sea Valley and continuing southeast along San Pedro Shelf. The SW-facing, 700 m-high San Pedro Escarpment is the bathymetric expression of the southwest limb of the PVA (Figures 1, 6-11, S1, S3, and S4). We apply the name Palos Verdes anticlinorium (PVA) to the entire structure, in contrast to the four anticlinoria proposed by others [Nardin and Henyey, 1978; Legg et al., 2004a]. The anticlinorium is made up of many shorter-wavelength (0.5–4 km) anticlines and synclines on the northwest part of PVA and to the southeast on San Pedro Shelf, of two or three anticlines separated by one or two synclines. (Figures 1c, 8–11, S1, and S4) [Fisher et al., 2004]. Short-wavelength secondary folds, as well as the southwest limb and northwest and southeast plunges of the PVA, all seafloor expression (Figure S1). Focusing on local anticlines and synclines seen in the bathymetry or on 2D seismic reflection profiles presents a very different scale impression than images of the entire structure such as in Figure 7.
4.3 Lateral and Forward Propagation of Palos Verdes Anticlinorium
 Stratigraphy carefully correlated to age control allows the timing of its deformation to be interpreted. This deformation can be faulting, folding, or both. The regional correlation of multiple stratigraphic horizons allows discussion of initiation of phases of deformation and changes in rate by location and time.
 Turbidites deposited onto the flank of an anticline with seafloor relief will thin and sometimes onlap onto the fold. Fanning of dips of turbidites is a measure of the growth of a progressively tilting anticline limb [Wickham, 1995]. Lack of strong tilting of the shallow section onto the southwest limb of PVA reflects the young age of strata above the Hueneme unconformity (Figures 5, 6, S2, and S3). Interpreting older tilting for timing of initial contraction is made complicated because tilting with the same sense occurred during Miocene extension. The position and sense of relief of secondary folding near the upper tip of faults can be used to distinguish extension from contraction. Thinning between top Repetto and top Miocene is by tilting on Figure S2 but is also by reverse separation on faults on Figures 6 and S3 and on F-F' on Figure 5. Both tilting and fault separation suggest that initiation of contractional folding of the PVA is Pliocene in the northwest. The section between the Miocene–post-Miocene unconformity and top Repetto is progressively tilted on Figures S2, S3, and S4, and both the northeast and southwest limbs are progressively tilted above top Miocene on E-E' (Figure 5). However, we do not know what part of the Repetto and older parts of the Pliocene might be missing and thus cannot be certain when during Pliocene time contraction initiated. If deposition was more-or-less continuous through Pliocene time, then contraction across northwest PVA initiated during the earliest Pliocene.
 The Miocene–post-Miocene unconformity is fully involved in contractional folding on J-J' (Figure 5). This post-Miocene folding is unrelated to PVA and is instead related to Redondo Knoll and the San Pedro Basin fault. Furthermore, the interval between the Miocene–post-Miocene unconformity and top Lower Pico is so thin that much of the Pliocene could be missing and the folding could be late Pliocene. Lack of precision in the interpretation in this area, discussed above in section 3.2, exacerbates this problem. Lack of contraction on K-K' supports the interpretation that the folding on J-J' is local (Figure 5).
 Slight thinning of the Repetto (4.5–2.5 Ma) [Blake, 1991] onto an anticline limb adjacent to the Palos Verdes fault on Figure 8 is permissive of minor late Pliocene shortening near the Palos Verdes fault. However, Figure 9 shows no thinning of that interval on the southwest upthrown side of that fault compared to its downthrown northeast side. Thinning on Figure 8 is much more pronounced above top Repetto. There is no systematic thinning of strata below the ~1.95 Ma [Blake, 1991] top Lower Pico onto the southwest limb of the PVA (Figures 8-11). Early(?) Pliocene initiation of shortening across the northwestern PVA and Quaternary initiation of shortening in the southeast suggests lateral propagation of anticlinal folding. Such lateral propagation of folding above the Santa Monica Bay detachment was coupled with a propagation toward the basin in the southeast. The PVA broke forward to fold into the basin above the Catalina Island detachment (Figure 7). A sharp north-south boundary of basin folding is seen in the subbottom maps; this boundary is likely a tear fault (Figures 7 and 12). The deep penetration digital data are somewhat sparse across this fault but an east dipping blind fault can be interpreted, with its upper tip below 3 s two-way travel time. It is not well imaged because little of it extends above acoustic basement. Strong down-to-the-west progressive tilt affects the forelimb above this blind fault between the Miocene–post-Miocene unconformity and the Pico unconformity.
5 Vertical Motions
 Regional subsidence is common in the California Continental Borderland and is expected due to Miocene crustal thinning, subsequent cooling, and loading from sediment and nearby thrust loading [e.g., Pinter et al., 2003]. Recognition of this regional subsidence has important implications to earthquake hazard from blind thrust faulting, because uplift above a subsiding base level can be far faster than uplift from sea level. Indeed, in Santa Barbara Channel, the offshore Oak Ridge fault has 1–2 mm/yr active vertical component of slip by faulting and associated folding, yet most of its relatively uplifted hanging wall has subsided with respect to sea level through Quaternary time [Marshall, 2012]. The structural relief of PVA is due to Plio-Quaternary continued subsidence coupled with (oblique) thrust and oblique right-reverse fault slip.
 In the following sections, we distinguish rock subsidence or rock uplift from water deepening or shallowing or surface uplift. England and Molnar  pointed out that surface uplift equals uplift of rock minus exhumation. Here we extend this terminology to subsidence and submarine settings. It is possible to have rock subsidence but a decrease of the water depth because of sediment deposition (“shallowing”). For example, rock layers within Los Angeles basin have experienced kilometers of Plio-Quaternary subsidence while the basin filled at a faster rate [Redin, 1991; Mayer, 1991], hence implying both rock subsidence and shallowing.
5.1 Early Miocene Denudation and Later Miocene Rock Subsidence
 Before early Miocene extension and the subsequent crustal cooling, the land-sea distribution was very different than it currently is [e.g., Nicholson et al., 2011]. The part of the onshore study area between the Newport-Inglewood fault and PVA was uplifted and deeply eroded during part of early and middle Miocene time [e.g., Wright, 1991]. Subsequent subsidence resulted in the deposition of middle Miocene or younger sedimentary rocks immediately above Catalina Schist (Figures 4, 8, 10, and S4). East of the Newport-Inglewood fault, in Los Angeles basin, basement was already buried before early Miocene time but deposition was near sea level at 18 Ma [Mayer, 1991].
 The Pliocene bathyal to abyssal depths of southern PVA and central Los Angeles basin result from middle and late Miocene deposition not keeping up with rock subsidence [Mayer, 1991]. Northern PVA was not as deep as areas to the south and east during earliest Pliocene time [Redin, 1991], as shown by thin Miocene strata above Catalina Schist basement there and at Playa del Rey (Figure 4) [Wright, 1991]. Rock and surface subsidence dominated the future sites of Los Angeles basin and PVA during middle Miocene time into part of Pliocene time, resulting in a deep marine basin, and the current structural relief did not yet exist.
5.2 Pliocene-Quaternary Rock Subsidence and Structural Growth of Palos Verdes Anticlinorium
 The PVA is partly the product of the late Pliocene-Quaternary rock subsidence of its flanks. Lack of deep erosion of the PVA on the outer shelf means that Quaternary rock uplift could not have been much more than the change in water depth from lower bathyal to the present shelf. Benthic foraminifera indicate deposition of Pliocene “Repetto” at lower bathyal or abyssal depths in areas that would become the crest of the PVA as well as the bottom of Los Angeles basin (supporting information Appendix 1). Redin  mapped paleobathymetry of much of Los Angeles basin and southern PVA to exceed 1700 m near the base of Pliocene strata. The current, greater than 5.5 km, depth of ~4.5 Ma base Repettian strata in central Los Angeles basin mapped by Wright  is consistent with 3–4 km of rock subsidence since then, assuming about 2 km paleowater depth (Figure 7) [Mayer, 1991]. Sediment loading can explain most of the rock subsidence in the last 3 Myr [Mayer, 1991], although tectonic loading from mountains above south verging thrust faults north of Los Angeles basin also contributes to rock subsidence of Los Angeles basin [Yeats and Beall, 1991]. BOEM paleontology from wells on southern PVA on both sides of the Palos Verdes fault (supporting information Appendix 1) indicates no evidence of early Pliocene water shallowing despite aggradation of Repettian Stage sediments. Therefore, rock subsidence was ongoing near the crest of southern PVA through at least early Pliocene time. In summary, rock subsidence of Los Angeles basin and rock uplift of PVA created the modern structural relief between them (Figure 2c) [Redin, 1991; Mayer, 1991; Wright, 1991; Ward and Valensise, 1994]. This relief growth must have occurred through Quaternary time (post-2.5 Ma) but may have initiated during the latter part of Pliocene time.
5.2.1 Direct Evidence of Post ~1.95 Ma Subsidence San Pedro Slope
 There is evidence for Quaternary rock subsidence of the lower southwest limb and southeast plunge of offshore PVA. Figure 11 images an angular unconformity (U3) that truncates the upper half of the Lower Pico. The unconformity coincides with or is barely above top Lower Pico so presumably is just slightly younger than 1.95 Ma. Its geometry is suggestive of a wave-cut platform whose inner edge (=shoreline angle, = back edge) is now just below 600 m depth. This eroded platform is imaged on a dozen profiles in this area and implies at least 500 m subsidence from a presumed −100 m glacial sea level lowstand level. Using <1.95 Myr and >500 m subsidence, subsidence rate exceeds 0.26 mm/yr. On Figure 11 the platform is on the upthrown side of the Palos Verdes fault, high on the NE limb of an anticline. Total subsidence and subsidence rate would be greater in structurally lower positions including San Pedro basin. Subsided wavecut platforms capping knolls suggests that the offshore flanks of PVA continue to subside [Fisher et al., 2004].
 Additional evidence for Quaternary rock subsidence is imaged 25 km to the southeast (Figure 12). An angular unconformity (U3) also suggestive of a wave-cut platform truncates top Lower Pico and is approximately the same age as U3 on Figure 11. The depth of its inner edge at 800 m implies 700 m of subsidence since sometime after 1.95 Ma. The subsidence rate would be a minimum 0.36 mm/yr. On this same profile, the top of Catalina Schist forms a planar surface (U4) at 1100 m depth that we also interpret as wave-eroded. The sedimentary rock above this surface is the 2.5–1.95 Ma Lower Pico [Blake, 1991]. Assuming deposition resumed soon after the rock platform subsided below lowstand sea level, the implied rock subsidence rate from last erosion of U4 to present is a minimum 0.40 mm/yr.
 Structural, bathymetric, and facies/paleodepth data discussed above consistently indicate that the Palos Verdes anticlinorium, including the Shelf Projection, Palos Verdes Hills, and San Pedro shelf, is a broad structural high that has been actively growing during late Pliocene through Quaternary time. This growth is just enough near its crest to counteract regional rock subsidence. It thus exhibits little physiographic change relative to sea level except for the surface uplift of Palos Verdes Hills (Figure 1B) [Woodring et al., 1946; Lajoie et al., 1991].
5.2.2 Decreased Late Quaternary Subsidence Shallow Subsurface Los Angeles Basin
 Detailed late Quaternary stratigraphy across central Los Angeles basin indicates that ongoing rock subsidence has slowed down for at least the last quarter million years. The lower part of the Harbor Sequence is coastal and beach, deposited during the Oxygen Isotopic Stage 8 lowstand [Ponti et al., 2007; Edwards et al., 2009]. The peak of this lowstand is at 252 ka [Lisiecki and Raymo, 2005], at a depth of 85 m [Miller et al., 2005]. The deepest part of the base of the Harbor Sequence is ~150 m in part of Long Beach downthrown to the Palos Verdes fault and ~120 m in central Los Angeles Basin [Ponti et al., 2007; Edwards et al., 2009]. Combining these depths, age, and eustatic sea level produces ongoing subsidence rates in the shallow subsurface of less than 0.27 mm/yr for Long Beach and less than 0.14 mm/yr for central Los Angeles basin. One possible explanation for much decreased subsidence rates of Central Los Angeles basin toward the end of Quaternary time is increased thrust faulting and structural thickening beneath the basin. If this decrease in subsidence rate is not related to thrust-thickening beneath Los Angeles basin, then the rate of structural relief growth of the PVA slowed and thrust earthquake hazard would be less than the late Pliocene-Quaternary average rate.
6.1 (Oblique) Thrust Reactivation of Low-Angle Regional (Oblique) Normal Faults
 Major Miocene basin-bounding low-angle normal faults are clearly imaged dipping northeast beneath PVA and Los Angeles basin (Figure 5). Their positions are consistent with the regional Miocene rock subsidence and surface subsidence in their hanging walls. The Catalina Island detachment and subsidiary faults above it bounded a deep bathymetric basin that is today part of San Pedro Shelf and slope. Crustal thinning and subsequent cooling occurred throughout the future locations of PVA and Los Angeles basin. The PVA formed within the hanging walls of these faults during Plio-Quaternary time in the north and Quaternary time in the south.
6.2 Slope Basin Inversion Contractional Rejuvenation of Extensional Paleohighs and Inversion of Minor Basins
 Basin inversion along the PVA is not as simple as is the case for the Santa Monica Mountains and Northern Channel Islands, where Miocene is thin to missing beneath Plio-Quaternary basins and is kilometers thick across the emerged islands and mountains [Seeber and Sorlien, 2000; Sorlien et al., 2006]. On PVA, a Quaternary anticlinal hinge developed along a previous depocenter adjacent to the Palos Verdes fault as a basin inversion (Figures 8 and S4) [e.g., Brankman and Shaw, 2009]. In contrast, the depocenter location continues relatively unchanged beneath San Pedro Escarpment. Miocene extensional half grabens above the NE dipping faults became post-Miocene synclines, and the tips of tilted blocks or extensional rollover anticlines continue to develop in contraction, as the border faults reactivate as blind thrust faults (Figures 5, 8, and S4).
6.3 The Palos Verdes Anticlinorium: An Inverted Extensional Orogen
 We propose a model for the evolution of the southern PVA that reproduces its first-order features (Figure 13). It takes into account reactivated faults and data such as paleodepth that have not been considered in some of the published models. Other inputs include stratigraphic age, sediment thickness, and geometries of shallow faults and folds. It neglects strike-slip motion. However, post-Miocene right-lateral motion on the Palos Verdes fault is ~5.5 km [Brankman and Shaw, 2009], far less than the along-trend continuity of the southern PVA (Figures 1 and 7). Despite this simplification and other nonuniqueness (such as the vertical motion of the footwall), the reconstruction points to some important constraints. The model explains the vertical motion history, spatial and temporal distribution of sediment thicknesses, the Catalina Schist-Miocene unconformity, and the folding.
 Acoustic basement is imaged throughout the anticlinorium. This basement is known from wells near and northeast of the crest of PVA to be the unconformity at the top of Catalina schist (Figure 10). We hypothesize a paleohorizontal surface near sea level at the end of early Miocene uplift and denudation, as supported by data discussed above in section 5.1 (Figure 13a). Such deep early Miocene denudation followed by subsidence is also seen in much of the outer California Borderland, with rock subsidence beneath the modern basins reaching 4 km [Pinter et al., 2003; Schindler, 2010; De Hoogh, 2012; Nicholson et al., 2011]. Crustal thinning and subsequent cooling result in rock subsidence except where later thrusting locally thickened the crust.
 The imaged set of several SW-dipping basement surfaces and NE-dipping faults above the basal fault suggest parallel tilted “bookshelf” extensional blocks typical of the shallow part of extensional orogens (Figures 8 and 13) [e.g., Wernicke and Axen, 1988]. Stratigraphic evidence for these faults is discussed above in section 4.1. This bookshelf geometry is consistent with an underlying low-angle normal fault with Miocene NE-verging extension. It corresponds to the clearly imaged Catalina Island detachment (Figure 5). It forms the top of basement, likely Catalina schist. Models for tectonic denudation of the Catalina Schist include large magnitude extension [e.g., Crouch and Suppe, 1993]. Such a fault would logically continue beneath Los Angeles basin because of kinematic problems if truncated by the ancestors of high-angle faults such as the Palos Verdes and Newport-Inglewood. The detachments could be offset by younger high-angle faults but would still exist on both sides as discontinuities available for thrust reactivation. The basal fault is just barely imaged at the southwest edge of N-N' in Figures 8 and 13c), because its shallow part is southwest of the profile and because it is within the difficult-to-image Catalina Schist basement.
 After a late Miocene phase of continuing extension, the structure highlights the regionally recognized end Miocene/early Pliocene transition from extension to shortening (Figure 13c) [Yeats and Beall, 1991; Wright, 1991; Schneider et al., 1996; Seeber and Sorlien, 2000; Brankman and Shaw, 2009]. Thrust slip beneath Los Angeles basin eventually slowed the rock subsidence there and near Long Beach, as discussed in sections 5.2.2. Forward (basinward) propagation of thrusting represents crustal thickening that is expected to counteract the base level subsidence in Figure 13c. The reverse component of slip on the onshore and nearshore restraining segment of the Palos Verdes fault is superimposed on crustal thickening due to blind thrust slip on NE dipping faults and contributes to rock uplift in the hanging wall of that restraining segment, with surface uplift onshore.
6.4 Comparison of Fault Models
 Our correlations of dated horizons and digital 3-D representations of faults and these horizons allow testing of the major existing models, in the context of evidence for vertical motions.
Bohannon et al.  interpret the basement relief between Palos Verdes Hills and the San Pedro Basin to be due to normal slip on a southwest dipping San Pedro Escarpment fault that outcrops near the base of San Pedro Escarpment. However, this interpretation does not explain the broad and regional southwest dip of Pliocene and Miocene strata beneath the escarpment (Figures 8, 9, S3, and S4). It is this dip, not vertical separation across faults, which explains most of the relief of Miocene rocks near the Shelf Projection and San Pedro Shelf. Furthermore, we interpret acoustic basement, likely Catalina Schist, to be about the same depth or shallower in the southwest footwalls of NE dipping faults than in their hanging walls. This shallow basement would be in the hanging wall SW of the Bohannon et al, 2004 normal fault (Figures 8, 10, and S4). We interpret SW dipping faults along the southern San Pedro Escarpment to die out laterally over relatively short distances (Figures 1a and 1c), and there is no direct evidence that these faults have large slips.
 The oblique right-reverse slip model for a restraining segment of the SW dipping part of the Palos Verdes fault produces surface uplift exceeding 0.3 mm/yr only over onshore Palos Verdes Hills (approximately the area of terrace 5 on Figure 1b), with uplift dying out to less than 0.1 mm/yr on the Shelf Projection and San Pedro Shelf (Figure 2a) [Ward and Valensise, 1994]. Furthermore, Ward and Valensise  use a 142° rake angle or 56% right-lateral and 44% reverse, in their model for Quaternary uplift related to Palos Verdes fault slip. McNeilan et al.  document that offset of a 7.8–8.0 ka channel by the Palos Verdes fault in Los Angeles harbor is 7:1 to 8:1 right-lateral to vertical ratio. Thus, either the Palos Verdes fault had a larger reverse component before 8 ka, or even less of the rock and surface uplift of PVA can be attributed to Palos Verdes fault slip. Structural relief growth of a minimum 0.26 mm year and possibly a minimum of 0.36 mm/yr is needed to keep these shelves from sinking along with San Pedro Basin (section 5.2.1). These minimum relief growth rates would be even higher if top Lower Pico, and the associated angular unconformity, is younger than 1.95 Ma, as is possible and discussed in the supporting information Appendix 1. Crustal thickening by (oblique) thrust slip on reactivated regional-scale low-angle normal faults can account for the growth of the PVA above adjoining subsiding basins.
 The two categories of thrust models differ on whether the southwest limb of the PVA is a forelimb or a backlimb (Figures 2b, 2c, and 3). In Figure 2c, the PVA is a SW verging anticline developing above a single NE dipping fault. Alternatively, the PVA verges to NE and develops above a SW dipping roof thrust that branches off a deeper NE dipping thrust and forms a tectonic wedge (Figures 2b, 3b, and 3c). One critical difference is the way the thrust faults relate to the preexisting normal faults. While in our preferred model the NE dipping thrust reactivates a preexisting normal fault, in the wedge model as published by Shaw and Suppe  (Figure 3b) the thrust faults and the wedge cut across steeply dipping Miocene normal faults rather than reactivating them.
 We prefer the simpler SW verging PVA model because it is the only one that accounts for all imaged structures and for the observed growth rate of the PVA. Specifically: (1) A regional SW-dipping shallow thrust fault distinct from the Palos Verdes fault as required in the wedge model is not imaged. (2) The tip of the thrust wedge of Shaw and Suppe  (Figure 3b) is near the coast and thus does not explain the San Pedro escarpment. (3) the NNE-dipping Santa Monica Bay detachment and the Catalina Island detachment faults are each well-imaged and the PVA folds above them in their southeast parts (Figures 5-7). (4) The NE-dipping thrust-reactivated “bookshelf” faults above the Catalina Island detachment are clearly imaged in MCS profiles, including LARSE-IR (Figure 10) [Ten Brink et al., 2000; Fisher et al., 2004; Baher et al., 2005]. These faults are much longer along strike and more numerous than SW dipping blind thrusts (Figures 1a and 1c), consistent with southwest vergence of the deeper system.
 In our preferred model for regional NE dipping faults without major SW dipping roof thrusts, the PVA has a broad forelimb. Such a broad forelimb is at odds with many widely accepted fold models but is predicted in some. The fanning of dips characteristic of progressive tilting of the southwest limb of PVA is featured in insets to Figures 11 and S3. Unfortunately, syn-thrust strata are not preserved on the upper part of the PVA and a record of possible tilting is unavailable there. Progressive tilting is not seen in the backlimbs for the kink models in Figures 14a and 14b [Seeber and Sorlien, 2000; Suppe et al., 1992] but is present in the Figure 14c, which includes a wide forelimb [Wickham, 1995]. Broad progressively tilting displacement-gradient folds are most likely to form where preexisting normal faults have been reactivated as thrusts. Normal faults tend to be listric and generally more appropriately oriented for thrust motion in their lower than in their upper parts. During thrust reactivation, therefore, the deep part of such a fault tends to slip more than the shallow part and an anticline with progressively tilting forelimb will form [e.g., Sibson, 1995].
6.5 Fault-Slip Rates
 This work argues that the PVA is rooted on two linked regional NE-dipping faults that reach under the Los Angeles basin and that these faults are being reactivated as thrusts in Plio-Quaternary time. The basin inversion model in Figure 13c requires 3 km of postextension shortening. This shortening started after 3.0 Ma and more likely as late as 1.95 Ma in the southeastern part of the PVA (section 4.3; Figure 8). This age range implies an average slip rate on the basal thrust fault between 1.0 and 1.7 mm/yr. A more direct estimate is obtained from a planar ramp model that accounts specifically for the post-1.95 Ma 1 km structural relief on the southeastern PVA. In this simple model, displacement is 2.9 or 2.0 km and slip rate is 1.6 or 1.1 mm/yr for a fault dipping 20º or 30º, respectively. These dips are similar to those imaged on the upper part of the Catalina Island extensional detachment in Figures 5 and 7. The deformation of the Hueneme unconformity by one of the San Pedro Escarpment faults (F-F' in Figures 5 and 6) suggests high young deformation rates at least locally on northwest PVA.
6.6 Steady Uplift Versus Structural Growth
 The above fault displacements are derived from the growth of the PVA. This growth contributes to crustal thickening and may or may not result in surface uplift depending on vertical motion at the base of the structure. PVA growth loads the lithosphere and contributes to subsidence. Additional loads from active contractional structures around the PVA as well as sedimentation and thermal cooling are likely to also contribute by flexure to regional base level subsidence. This subsidence can therefore account for the small or even lack of uplift along the PVA, despite PVA growth and crustal thickening. Conversely, uplift is not a proxy of structural growth and if used as such can lead to drastic underestimate of thrust-fault slip, as has been the case for the PVA (Figure 1B).
 The least astonishing growth model for the PVA is a steady rate into the present. Geologic constraints allow for more complex evolutions but include strong evidence that the PVA has been active during the last glacial cycle. The youngest (Holocene) strata at the base of San Pedro Escarpment are folded (Figures 8, 9, and 11) [Fisher et al., 2003, 2004; Bohannon et al., 2004]. Post-125 ka submarine fan strata are offset and folded by strands of the San Pedro Escarpment faults (Figure 6, F-F' on Figure 5). The same Hueneme unconformity and overlying strata are folded by the “bookshelf” faults above the Catalina detachment (U1, orange, in Figure 9), and possibly correlative U1 is strongly tilted farther south (Figure 11). The PVA is clearly an active Plio-Quaternary structure.
6.7 Geologic Versus GPS Rates
 The 1.0–1.7 mm/yr Plio-Quaternary slip rate we estimate for the basal thrust responsible for the PVA is consistent with the current shortening rate across the Los Angeles Basin. A 4.5 ± 1 mm/yr contraction modeled from GPS is concentrated in the northern part of the basin [Argus et al., 2005] and is correlated with an intense belt of earthquakes [Hauksson, 2000, 2011]. This relatively high rate may stem in part from local faulting such as the Puente Hills fault [Shaw and Shearer, 1999; Argus et al., 2005]. Interpretations of this proposed root zone include structurally deeper candidates thrust faults that are possible candidates for the basal thrust responsible for the PVA [Davis et al., 1989; Shaw and Suppe, 1996; Shaw et al., 1999, Plesch et al., 2007]. In general, the pattern of current deformation and seismicity and the long-term growth of the PVA are reminiscent of subduction zones or contractional mountain belts such as the Himalayas, where GPS shortening and hypocenters are concentrated near the transition between creeping deep parts of the faults and the locked upper part. This interseismic activity is typically concentrated at the deep end of a future rupture reaching as far as the deformation front, 100 km or more updip on the megathrust [Avouac, 2003; Ide et al., 2011]. At least qualitatively, therefore, current activity seems consistent with the deeply rooted large thrust fault we propose and with this fault being locked and accumulating strain during an interseismic period.
6.8 Implications for Earthquake Hazard
 Earthquake hazard analysis relies heavily on the geometry and long-term kinematics of active faults obtained from geologic investigations. Such data are particularly important where deformation is partitioned among many moderately active faults, such as our study area, where GPS modeling alone is highly nonunique. Both wedge (Figure 2b) and single thrust (Figure 2c) models for the PVA root into SW-directed deep thrust system(s) beneath Los Angeles basin. However, displacement fields and earthquakes anticipated from these models differ in ways that are important for hazard and can be used to test them. The single fault (C) and shallow wedge (B) models entail similar total fault surface within the seismogenic depth range and thus similar long-term moment release rates. The difference is in the size of the fault that can rupture in a single earthquake. Maximum magnitudes are higher for larger faults. Larger rupture areas and higher magnitudes imply larger displacements and thus longer recurrence times for a given slip rate [e.g., Stein, 2007]. Qualitatively therefore the maximum earthquakes from (C) are expected to be less frequent but larger than from (B). Increasing the maximum magnitude generally increases the long-term cumulative losses because damage is a highly nonlinear function of magnitude [Rowshandel et al., 2005]. The most dramatic implication of the two models in terms of hazard, however, may be in the tsunami potential. A complete rupture of the fault in (C) might reach the seafloor with particularly large displacements, similarly to the 2011 Tohoku earthquake [Ide et al., 2011], and would be likely to generate a tsunami much larger than would a rupture in (B), even if a wedge earthquake involved both faults. Finally, an oblique right-reverse earthquake on the Palos Verdes fault (model A) (Figure 2a) would be unlikely to cause a large tsunami. Since the onshore and inner shelf Palos Verdes fault dip is steeper than the blind thrusts and dips southwest away from Los Angeles basin, maximum magnitude and onshore ground motion should both be lower than for rare ruptures of the blind thrusts.
7 Summary and Conclusions
 We analyzed a large body of data, primarily multichannel seismic reflection profiles and reports from wells and core holes, to identify and map stratigraphic horizons offshore of the Los Angeles region in southern California. By combining our results with a published horizon map on land, we generated a digital representation of a near-base Pliocene horizon across the entire onshore-offshore extent of the Palos Verdes anticlinorium (PVA). This representation shows the PVA to be a regional structure that transcends the complex superposition of numerous folds and faults. The PVA is a broad, 70 km long structural high that grew in Plio-Quaternary time along what is now coastal Los Angeles and controls the position of this portion of the coast. The eroded crest of the PVA forms the broad continental shelf, on which the Palos Verdes Hills are a relatively small, more rapidly uplifting erosional remnant. The PVA verges seaward and its southwest submarine forelimb forms the 700 m high seafloor San Pedro Escarpment. The backlimb northeast of the PVA is also the buried southwest flank of the LA Basin. The most likely structures flooring the PVA are the low angle Santa Monica Bay detachment and Catalina Island detachment dipping northeast below the Los Angeles area that may or may not root into a single décollement. These mostly blind faults are clearly imaged west of the PVA and locally imaged below its forelimb. Basement depth and Miocene strata distribution suggest that these faults existed in the Miocene as large displacement normal faults and were reactivated in the Plio-Quaternary as thrust faults. This reactivation can account for some of the features of the PVA, including the wide forelimb.
 Folding initiated during early Pliocene time in the northern PVA, late Pliocene time in the central PVA, near the Palos Verdes fault, but not until early Quaternary time beneath the San Pedro Escarpment in the southern PVA. Quaternary structural relief growth is used to estimate thrust slip rates for the southern low-angle thrust fault system. Dip-slip rates for the southeast part of the PVA vary between 1.1 and 1.6 mm/yr assuming faults dipping 30º or 20º. This is consistent with our cross section, reconstruction that includes 3 km of shortening during late Pliocene and Quaternary time. Ongoing uplift of the crest of the anticlinorium above its subsiding southwest flank and deformation of young strata indicate that the PVA and its basal thrust faults, which are rooted under Los Angeles Basin, are active. The large size, proximity, and low slip rate characteristics of this fault suggest that it is important for the hazard of the metropolitan area by contributing long recurrence but high impact earthquakes.
 This research was supported by the Southern California Earthquake Center; SCEC contribution number 1192. SCEC was previously funded by NSF Cooperative Agreement EAR-0106924 and USGS Cooperative Agreement 02HQAG0008. The work in Santa Monica Bay was earlier supported by USGS-NEHRP, USDI/USGS 03HQGR0048. Multiple major revisions including additional interpretation were done while funded to work on an adjoining area by USGS-NEHRP G12AP20069. We gratefully acknowledge the support of Paradigm for the discounted use of Gocad and Seismic Micro Technology (now IHS) for the donation of The Kingdom Suite Software. Frank Victor of U.S. Minerals Management Service (now BOEM) provided publicly-released well data and MMS paleontology.
 Stratigraphic data were also provided by the California core repository in Bakersfield, the California State Long Beach University log repository, California State Lands office in Long Beach, Marc Kamerling, Rob Mellors, and Andy Rigor. Robert “Dan” Francis provided detailed stratigraphic correlation examples that lead us to revise the stratigraphic interpretation of the northern part of San Pedro basin. Background geologic discussions with Marc Kamerling, Mark Legg, and Craig Nicholson are appreciated. Doug Wilson provided discussion of the components of basin subsidence. Marie-Helene Cormier and Robert Bauer edited the revised text. Robert Yeats, three anonymous reviewers, and associate editor Isabelle Manighetti provided detailed suggestions for improvement of a previous version of the manuscript. Robert Yeats and two anonymous reviewers provided prompt and helpful reviews for this rewritten version; editor Thorsten Becker provided additional helpful comments. The authors acknowledge WesternGeco and the U.S. Geological Survey for providing the seismic reflection data for the purpose of this research. We regret that Bill Normark and Kris Broderick did not survive to enjoy the publication of this work.