Abstract

Industry seismic reflection data, oil test well data, interpretation of gravity and magnetic data, and seismic refraction deep-crustal profiles provide new perspectives on the subsurface geology of San Fernando Valley, home of two of the most recent damaging earthquakes in southern California. Seismic reflection data provide depths to Miocene–Quaternary horizons; beneath the base of the Late Miocene Modelo Formation are largely nonreflective rocks of the Middle Miocene Topanga and older formations. Gravity and seismic reflection data reveal the North Leadwell fault zone, a set of down-to-the-north faults that does not offset the top of the Modelo Formation; the zone strikes northwest across the valley, and may be part of the Oak Ridge fault system to the west. In the southeast part of the valley, the fault zone bounds a concealed basement high that influenced deposition of the Late Miocene Tarzana fan and may have localized damage from the 1994 Northridge earthquake. Gravity and seismic refraction data indicate that the basin underlying San Fernando Valley is asymmetric, the north part of the basin (Sylmar subbasin) reaching depths of 5–8 km. Magnetic data suggest a major boundary at or near the Verdugo fault, which likely started as a Miocene transtensional fault, and show a change in the dip sense of the fault along strike. The northwest projection of the Verdugo fault separates the Sylmar subbasin from the main San Fernando Valley and coincides with the abrupt change in structural style from the Santa Susana fault to the Sierra Madre fault. The Simi Hills bound the basin on the west and, as defined by gravity data, the boundary is linear and strikes ∼N45°E. That northeast-trending gravity gradient follows both the part of the 1971 San Fernando aftershock distribution called the Chatsworth trend and the aftershock trends of the 1994 Northridge earthquake. These data suggest that the 1971 San Fernando and 1994 Northridge earthquakes reactivated portions of Miocene normal faults.

INTRODUCTION

The Los Angeles metropolitan area of southern California contains millions of people and some of the most expensive real estate in the world. The area is subject to damaging earthquakes, most recently in 1971, 1987, and 1994, which has led to a concerted effort to evaluate the earthquake hazard in terms that decision makers can use for land-use planning and building-code upgrades. Two of the most recent damaging earthquakes, the 1971 San Fernando and 1994 Northridge events, were located in the San Fernando Valley (Fig. 1).

Although the surface and subsurface geology of the San Fernando Valley had been studied extensively since the pioneering study of Oakeshott (1958), the 1971 and 1994 earthquakes were unexpected. In this paper we provide new perspectives on the subsurface geology of this active region from analysis of wells and seismic reflection profiles acquired by the petroleum industry, gravity and magnetic data, and deep-crustal profiling of the Los Angeles Region Seismic Experiment (LARSE) II (Fuis et al., 2003; Thygesen, 2003; Lutter et al., 2004). The combined geological and geophysical data sets permit a new three-dimensional analysis of the basin that provides insight into the relationship between the disparate geologic terranes exposed on its northeast and southwest margins and that constrains the bounding structures of this roughly triangular valley. The subsurface configuration of the basin, including the geometry of the bounding faults, is essential to predicting the ground motion caused by future earthquakes.

Wald and Graves (1998) compared the predicted ground motions from three different three-dimensional velocity models for southern California. The largest differences in the model predictions are in the San Fernando Valley and the deepest part of the Los Angeles Basin. According to their analysis (Wald and Graves, 1998), the structure of the basins, in particular their depths, is a major factor in producing the discrepancies between the velocity models. Other factors such as fault geometry, distribution of Quaternary fill, and uncertainties in shear-wave velocities also are important (e.g., Hartzell et al., 2005). The diversity of predictions in the San Fernando Valley is not surprising because the geometry of the San Fernando Basin is not well known.

Previous efforts at determining the architecture of the basin suggested contrasting geometries for the basin fill. Using gravity data, Corbató (1963) inferred that the deepest part of the San Fernando Basin was south of the Mission Hills thrust fault (MHF in Fig. 1), where he concluded that the thickness of basin fill reaches 4.5 km. Oakeshott (1975), however, drew a cross section in which he showed the deepest part (6 km) of the basin as being north of the Mission Hills thrust fault and thrust beneath the San Gabriel Mountains. In this paper we integrate geophysical and geologic data to gain a better understanding of the structure of the San Fernando Basin and its tectonic setting. We address overall basin geometry by describing the geophysical setting of the area and summarizing results from 2.5-dimensional modeling of gravity and magnetic data (Langenheim et al., 2000, 2001) and deep-crustal profiles from the LARSE II transects (Fuis et al., 2003; Thygesen, 2003; Lutter et al., 2004). Potential-field and seismic refraction methods, however, cannot image details within the basin fill. To elucidate the internal structure of the basin we analyzed ∼105 km of mostly unpublished seismic reflection lines (blue lines in Fig. 2) donated by several oil companies following the Northridge earthquake, and synthesize our findings with data from a number of previous studies of the subsurface geology of the San Fernando Valley (Shields, 1977; Tsutsumi and Yeats, 1999; Wright, 2001a; Yeats, 2001) that utilized wells drilled by the petroleum industry. In combination, these data provide a foundation for determining the architecture of the basin and its relation to seismic hazard and tectonic history.

GEOLOGIC SETTING

The San Fernando Valley is a roughly triangular feature (Figs. 1 and 2) within the Transverse Ranges province, a region noted for its intense and relatively young deformation (Morton and Yerkes, 1987; Donnellan et al., 1993; Wright, 1991; Yeats et al., 1994). The Transverse Ranges province is a region of north-south shortening (gray area in inset, Fig. 1) that extends east-west across the northwest-trending San Andreas fault system. Its topographic features and geologic structure trend east-west and are most strikingly developed in the Ventura Basin. The San Fernando Valley is south of the east end of the present Ventura Basin (Fig. 1). The Santa Monica Mountains, part of the Transverse Ranges, form the southern margin of the San Fernando Valley. The northern margin of the San Fernando Valley is marked by a steep topographic front (Fig. 2) associated with the north-dipping Santa Susana thrust and Sierra Madre faults (Figs. 1 and 2), part of a discontinuous, predominantly north-dipping thrust belt extending from the Santa Barbara Channel on the west to the San Bernardino Valley on the east.

The eastern margin of the San Fernando Valley trends northwest and is bounded by the plutonic and metamorphic rocks of the Verdugo Mountains (Fig. 1), uplifted east of the Verdugo fault. Reflecting the northwest orientation of the San Andreas fault system north and south of the Transverse Ranges (Fig. 1, inset), the Verdugo and northern San Gabriel faults (Fig. 1) are believed to have served as a major locus of dextral slip on the San Andreas system during Middle and Late Miocene time, connected by the Middle Miocene Canton and Devil Canyon faults (e.g., Yeats et al., 1994; Crowell, 2003; Yeats and Stitt, 2003; Yeats, 2004). The western edge of the valley, defined by exposed Paleogene and Cretaceous rocks of the Simi Hills, trends northeast and is bounded by the Chatsworth Reservoir and North Reservoir faults (Fig. 1).

Clockwise rotation of the western Transverse Ranges by 70°–90° since early-Middle Miocene time is now accepted as a major element of the Neogene tectonic evolution of California (Hornafius et al., 1986; Luyendyk, 1991; Nicholson et al., 1994). This plate-boundary deformation began with regional extension and involved volcanism centered at several locales along the southern boundary of the Transverse Ranges. By the start of Pliocene time, the primary locus of San Andreas dextral slip had shifted east to its present position along the northeast edge of the San Gabriel Mountains, and deformation within the western Transverse Ranges became strongly compressive. The western Transverse Ranges block rotated around an axis near its present southeast corner (Nicholson et al., 1994), an area that includes the San Fernando Valley and the easternmost Santa Monica Mountains and is adjacent to the Verdugo–Devil Canyon–Canton-San Gabriel fault zone.

Data from oil test wells record the depositional history that followed Middle Miocene extension. During Late Miocene time, sands derived from within the San Gabriel Mountains swept southward (in present orientation) across the basin floor east of the Simi Hills and, as the Tarzana fan (Fig. 1; Sullwold, 1960; Redin, 1991; Rumelhart and Ingersoll, 1997), across the future site of the Santa Monica Mountains and into the Los Angeles Basin (Wright, 1991; Redin, 1991).

Uplift of the Santa Monica Mountains began in latest Miocene time, at about the time when the locus of San Andreas dextral slip was transferring to its present position along the northern edge of the San Gabriel Mountains. During latest Miocene and early Pliocene time, an east-southeast–trending hinge line separated the deep east Ventura Basin from a shelfal area where marine sediments of the Fernando Formation were deposited over what would become the San Fernando Valley. Overlying strata of the Saugus Formation record the shallowing of this shelfal area and, beginning ca. 0.5 Ma, tectonic inversion of the east Ventura Basin to form the Santa Susana Mountains (Levi and Yeats, 1993). Sediments shed from that active uplift have built a southward-sloping alluvial plain that has pushed the Los Angeles River against the edge of the Santa Monica Mountains (Wright, 2001b). The San Fernando Valley is now an almost fully enclosed basin draining, via the Los Angeles River, through a narrow gap at its southeastern corner.

Basement Terranes

Sedimentary strata in the San Fernando Valley area overlie three main basement terranes (Fig. 3A) exposed in the San Gabriel, Verdugo, and Santa Monica Mountains. The San Gabriel Mountains northeast of the San Gabriel and Vasquez Creek faults are underlain almost entirely by crystalline rocks, including Paleoproterozoic and Mesoproterozoic rocks and Mesozoic intrusive rocks (Ehlig, 1981; Nourse, 2002). Southwest of the San Gabriel and Vasquez Creek faults, rocks of the San Gabriel foothills and adjacent Verdugo Mountains are predominantly quartz diorite and granodiorite with an estimated age of ca. 122 Ma (Crook et al., 1987) intercalated with younger granitic rocks and with older metasedimentary rocks, all of which underwent only one major metamorphic event, probably during the Mesozoic (Ehlig, 1981). North of the San Fernando Valley, gneissic rocks were found in the Bell Bartholomaus well (drill hole 48 in Table 1 and Fig. 2), the Conoco Phillips well (CP in Fig. 1), and the Mobil Macson Mission well in the northern part of the valley (drill hole 52 in Table 1 and Fig. 2; Yeats et al., 1994, p. 1045).

In the Santa Monica Mountains south of the San Fernando Valley, the Santa Monica Formation, a weakly metamorphosed clastic sequence of Late Jurassic age, is intruded by granodiorite and quartz diorite dated as 102 ± 10 Ma (L.T. Silver, inDibblee, 1982). The age of the plutonic rocks suggests that these rocks are part of the Peninsular Ranges batholith. In the south-central San Fernando Valley, the Chevron Leadwell and Chevron Hazeltine wells (drill holes 29 and 28 in Table 1 and Fig. 2) encountered granitoid basement at unexpectedly shallow depths. These granitic rocks are similar to those in the Santa Monica Mountains but in lithologic and aeromagnetic contrast to rocks in the Verdugo Mountains and San Gabriel foothills (Langenheim et al., 2000, 2001, 2006).

Sedimentary Strata

The three basement terranes are overlain by a sedimentary section that ranges in age from Cretaceous to Quaternary (Fig. 3A). The Santa Monica Formation is overlain with angular unconformity by a sedimentary sequence of Late Cretaceous to Miocene age; only in the Santa Monica Mountains is the unconformity at the base of the Late Cretaceous exposed (Hoots, 1931). Where it overlies Jurassic metasedimentary rocks, the Cretaceous sequence is relatively thin. Farther northwest, in the Simi Hills, Cretaceous marine strata are >1800 m thick (Colburn et al., 1981). Unlike the Cretaceous sequence in the Santa Monica Mountains, the base of Cretaceous strata in the Simi Hills is not exposed, nor is it reached by wells in the Santa Susana Mountains (Yeats, 1987a).

In the Simi Hills, Cretaceous strata are crossed obliquely by the Paleogene Burro Flats fault (BFF in Fig. 1; Yeats, 1987b). West of the Burro Flats fault, Cretaceous strata are overlain by a Paleogene sequence that includes the basal nonmarine Simi Conglomerate, the Paleocene shallow-marine Las Virgenes Sandstone, the marine Paleocene–Eocene Santa Susana Formation, the marine Eocene Llajas Formation, and the nonmarine Eocene–Oligocene Sespe Formation (Fig. 3). East of the fault, the Santa Susana Formation overlaps older Paleocene strata and directly overlies Late Cretaceous beds, documenting the presence of a Paleocene fold-thrust belt (Yeats, 1987b). These rocks form a north-dipping to west-dipping homocline. To the west and south, the homocline is overlain with an angular unconformity by late-Middle and Late Miocene strata, demonstrating that homoclinal tilting predated the late-Middle Miocene (Luisian stage of Kleinpell, 1938). Beneath the San Fernando Valley, Cretaceous–Oligocene rocks are found in drill holes primarily along the western part of the valley (Yerkes and Showalter, 1990; Table 1). In the northern San Fernando Valley, Cretaceous strata are found in the Pacoima oil field (Fig. 2), where the Chevron Pacoima 9 drill hole found the Topanga Formation to be underlain by Cretaceous variegated sandstone and gray siltstone bearing late Coniacian to early Campanian coccoliths. The pre-Late Miocene geology and structural problems were described in more detail in Yeats (2001).

The overlying Late Oligocene to Middle Miocene sequences are notable in two respects: first, they contain volcanic rocks, and second, they are dominated by extensional structures. The oldest strata deposited in this environment are parts of the Sespe Formation and its equivalents, including anorthosite-bearing sedimentary breccia in the east Ventura Basin derived from the San Gabriel Mountains prior to displacement on the Canton and San Gabriel strike-slip faults (Crowell, 2003; Yeats and Stitt, 2003). In the San Fernando Valley, mafic volcanics are exposed in the Pacoima Hills and on the north slopes of the Verdugo Mountains. Mafic volcanics are also found in wells beneath Late Miocene strata in the Topanga Formation (drill holes 11, 12, 24, and 30 in Table 1 and Fig. 2; Aliso Canyon oil field in the Santa Susana Mountains; Yeats et al., 1994). Topanga volcanic rocks exposed near Big Tujunga Canyon (BTC in Fig. 2) are dated as 16.1 Ma (Beyer et al., 2009).

Middle Miocene and older rocks commonly are overlain with angular unconformity by Late Miocene (Mohnian) strata (Modelo Formation), with unconformable relationships well displayed at the crest of the Santa Monica Mountains (Fig. 1). These strata represent the beginning of the sequence forming the San Fernando basin. The most prominent Late Miocene feature is the Tarzana submarine fan of Sullwold (1960), in which deep-water turbidite sandstone is interbedded with fine-grained clastic and organic strata. The Tarzana fan is part of the Modelo Formation, which is 900 m thick southwest of the Verdugo fault and ∼1500 m thick northeast of the fault (Wright, 2001a, fig. 5 therein). From a now-displaced basement source to the northeast, the Tarzana submarine fan spread southward across the Santa Monica Mountains in two distinct lobes (Fig. 1; Sullwold, 1960; Rumelhart and Ingersoll, 1997). These deposits form the reservoir in the Pacoima Oil Field (Fig. 2). The Tarzana fan deposits wedge out toward the Chatsworth Reservoir and North Reservoir faults, indicating that the Simi Hills were a positive topographic feature during Tarzana deposition.

The latest Miocene to early Pliocene Towsley Formation (Delmontian stage as modified by Blake [1991] from Kleinpell [1938]) consists of gray to dark brown siltstone and shale with lenses of sandstone. It overlies the Modelo Formation with angular unconformity in the Horse Meadows oil field (Tsutsumi and Yeats, 1999). The Towsley Formation has been mapped in parts of the San Fernando Valley, but elsewhere is included in the overlying Pliocene Fernando Formation. The Fernando Formation is composed of fine- to coarse-grained marine sandstone interbedded with siltstone. Because of facies changes across the region, it is difficult to date these strata more precisely than post-Mohnian (post–6 Ma) and older than the base of the Saugus Formation, dated paleomagnetically as 2.3 Ma (Levi and Yeats, 1993).

The Saugus Formation, consisting of a lower, shallow-marine to brackish-water Sunshine Ranch Member and an upper nonmarine member, is dated paleomagnetically to be as young as 0.5 Ma (Levi and Yeats, 1993). The Saugus is as thick as 1200 m south of the Mission Hills, but is >3600 m thick in the Sylmar subbasin north of the Mission Hills in the Sun Exploration Stetson-Sombrero well (drill hole 35 in Table 1 and Fig. 2).

GEOPHYSICAL DATA AND ANALYSIS

In this section we describe geophysical data that provide insights into the stratigraphic and structural relationships in the San Fernando Valley. We first present seismic reflection data and analysis that result in depths to Quaternary to Miocene horizons beneath the valley, followed by analysis of the LARSE seismic refraction and potential-field data that define some of the basin structure beyond the depth resolution and spatial coverage of the industry seismic data.

Seismic Reflection Profiles

We obtained 13 industry Vibroseis seismic profiles distributed across the west-central San Fernando Basin. These profiles, totaling ∼105 km in length (Fig. 2), were originally collected for hydrocarbon exploration, but were released to the Southern California Earthquake Center by the petroleum industry after the Northridge earthquake for use in earthquake investigations. One of these profiles (85–12) was previously published by Tsutsumi and Yeats (1999) and 85–7 was presented in Wright (2001a). Five profiles (2893 lines; Fig. 2) were collected in 1978 using 8–48 Hz sources at 220 m intervals, recording with 48 channels at 220 m group intervals. Two profiles (84–202 and 84–205) were collected in 1984 using 14–48 Hz sources at 220 m spacing, recorded by 96 channels at 110 m group intervals. The remaining profiles (85–7, 85–8, 85–9, 85–10, 85–11, and 85–12) were collected in 1985 using 10–48 Hz sources at 110 m spacing, recorded by 96 channels, at 110 m group intervals. Standard seismic processing was applied, including elevation statics, deconvolution, bandpass filtering, velocity analysis, and normal moveout stacking, and, for the 1984 and 1985 data sets, a post-stack migration. All of the seismic lines (both unannotated and annotated) are available as figures in the Appendix.

Reprocessing of the digital field data revealed that significant subbasin reflection energy is not present. The raw field gathers required extensive editing and data merging. Extended Vibroseis correlation was applied (e.g., Okaya and Jarchow, 1989). Strong-amplitude reflections, which might be associated with low-angle detachment surfaces, are not present; neither are lower crustal reflection fabrics or Moho reflections. Reasons for the lack of subbasement and/or crustal reflectivity are: (1) any subbasin low-angle fault zones (e.g., blind thrust ramps) are nonreflective; (2) the materials above and below low-angle fault zones are internally not all that reflective and/or there is not a major acoustic impedance contrast across the two; (3) few geological features are present in the upper to lower crust or Moho to act as reflectors; and (4) in terms of seismic acquisition, insufficient Vibroseis energy has propagated beneath the basin into the middle and lower crust. For the last category, this lack of energy means that any potentially reflecting horizons are not imaged. The lack of propagating energy is possibly due to either insufficient energy at the vibrating source or significant attenuation by the basin sediments. A detailed examination of the seismic field notes indicates that the Vibroseis source levels were suppressed due to their proximity to houses within the densely populated San Fernando Valley. Examination of seismic reflection line 85–12 highlights the absence of strong, continuous reflections below ∼1.8 s two-way traveltime predicted from the synthetic seismogram developed from the sonic log of the nearby Frieda Clark drill hole (Fig. 3; well 25 in Table 1 and Fig. 2).

Information from Well Sonic Logs

Sonic logs of four key wells located within the San Fernando Basin near the seismic profiles were obtained from industry and commercially digitized at 0.15 m intervals (Brocher et al., 1998). The logs (Fig. 4) indicate that the upper 2.5 km of strata are seismically slow; the two wells that intersected basement (drill holes 28 and 29 in Table 1 and Fig. 2) show marked increases in velocity at the basement interface. The time-depth curves are used to convert the seismic profile two-way traveltimes to depth and allow for assignment of geologic units to reflecting horizons. Because density logs were also available, the construction of synthetic seismograms was possible. Figure 3 illustrates the synthetic seismogram for the Frieda Clark drill hole using the sonic and density logs and a seismic source representative of the Vibroseis source used for these profiles. These synthetic data were used to identify reflective horizons within the stratigraphy and to tie the wells to the seismic reflection data.

Well-Seismic Horizon Correlations

Stratigraphic horizons in the wells were converted from depth in meters to two-way traveltime in seconds using the Burnet well time-depth curve (Fig. 4), adjusted to match the elevation datum of the seismic reflection line, and then projected onto the nearest seismic reflection lines. Reflections then were traced along the line. Intersections of seismic lines allowed correlation of reflections onto lines that are not immediately adjacent to the four key drill holes.

The time-depth curves (Fig. 4) indicate that seismic velocities are slightly faster in the Burnet well (drill hole 40 in Table 1 and Fig. 2) than in the Frieda Clark well (drill hole 25 in Table 1 and Fig. 2). The differences in the time-depth curves suggest that any structure contour maps may be as much as 250 m too deep for a two-way traveltime of 1.8 s in those areas where the Frieda Clark time-depth curve is more appropriate. Although the depth to the horizon may be in error, the shape of the horizon surface should be accurate even if miscorrelation of one reflection below or above the appropriate reflector is made. Additional error may result from slight differences in elevation datum between sets of seismic reflection profiles.

LARSE II Transects

The LARSE phase II included 4 refraction and/or wide-angle reflection profiles that extended through or were confined within the San Fernando Valley (Figs. 1 and 2). The main profile extended approximately north-south through the valley with shotpoints nominally spaced 1 km apart in order to produce both refraction and low-fold reflection data. Two profiles extended approximately east-west in the south and north parts of the valley, lines 3000 and 4000, respectively, with the goal of obtaining refraction and/or wide-angle reflection data on depth to basement in these parts of the valley. A fourth profile, line 5000, simply recorded refraction data between the shotpoint at the east end of line 4000 and a shotpoint in the Santa Susana Mountains on the main profile. The position of line 4000 was also chosen to be parallel to and near (within 500 m) an industry reflection line (85–9; see following).

Low-fold reflection data from the main profile were modeled in Fuis et al. (2003). Refraction data were modeled in Lutter et al. (2004) using velocity tomography, and in Fuis et al. (2005) using forward modeling. The forward model is constrained not only by refractions, as is the tomography model of Lutter et al. (2004), but also by wide-angle reflections from layer boundaries. The refraction data from lines 3000, 4000, and 5000 are similarly modeled by velocity tomography as well as forward modeling that takes into account wide-angle reflections (Thygesen, 2003).

Gravity and Magnetic Data

The isostatic residual gravity map of the study area (Fig. 5) is based on more than 5600 gravity measurements (Langenheim et al., 2000; Wooley et al., 2003). Station spacing varies widely in the region. Mountainous areas (e.g., the San Gabriel Mountains) tend to have sparse gravity station coverage and may have only 1 station per 10 km2; basin areas generally have denser gravity station coverage, as many as 23 stations per km2 and averaging ∼1 station per km2. The data have been processed to generate isostatic residual gravity anomalies, which reflect density variations in the upper and middle crust; details on reduction and processing of the data were presented in Langenheim et al. (2000) and Wooley et al. (2003).

The aeromagnetic map (Fig. 6) consists of detailed aeromagnetic data collected over the region in 1995 (U.S. Geological Survey, 1996) and prior data over the Ventura Basin (U.S. Geological Survey, 1980). Most of the new data were flown along north-south–trending flight lines spaced 800 m apart with an average terrain clearance of 550 m. A map of these data superposed on geology is available at a scale of 1:100,000 for most of the study area (Langenheim et al., 2006). Aeromagnetic data were gridded at an interval of 250 m and have not been reduced to pole.

To delineate subtle anomalies in the gravity field, the gridded gravity data were filtered by subtracting the upward continuation of the gravity data from the actual data. Upward continuation is the analytical transformation of gravity data measured on one surface to a higher surface; this operation smoothes the data by attenuating short-wavelength anomalies (Dobrin and Savit, 1988). Subtraction of the upward-continued field from the original data results in short-wavelength residual anomalies, which arise from near-surface density sources. The gravity data were upward continued to 1 km above ground level and then subtracted from the original data to produce the residual gravity map (Fig. 7).

To help delineate trends and gradients in the gravity and magnetic fields, we used a computer algorithm to locate the maximum horizontal gradient (Cordell and Grauch, 1983; Blakely and Simpson, 1986). Density boundaries (dark gray circles in Fig. 5) were calculated from the isostatic residual anomalies. Magnetization boundaries (white dots and crosses in Fig. 6) were calculated from data filtered to emphasize sources below ∼2 km (thus damping anomalies resulting from shallow, anthropogenic sources), and transformed to magnetic potential anomalies (Baranov, 1957), which also shifted the anomalies over the sources. Gradient maxima are located approximately over vertical or nearly vertical contacts that separate rocks of contrasting densities or magnetization. For moderate to steep dips (45° to vertical), the horizontal displacement of a gradient maximum from the top edge of an offset horizontal layer is always less than or equal to the depth to the top of the source (Grauch and Cordell, 1987).

Analysis

The well data, in particular those with sonic and density logs, provide the basis for identifying geologic horizons in the industry seismic reflection profiles. Figure 8 shows five lines of particular interest, including seismic reflection line 85–7, which shows a pair of concealed normal faults beneath the southeastern San Fernando Valley. Four horizons were identified on the profiles: (1) the top of Sunshine Ranch member of the Saugus Formation (blue line), (2) the base of the Saugus Formation (purple line), (3) the top of the Modelo Formation (red line), and (4) the base of Modelo Formation (orange line). The top of the Towsley Formation could only be identified and correlated in a few wells and was not deemed to be reliably identified on the industry profiles. The four horizons were tied into other profiles. Figure 9 shows structure contour maps of the base of the Saugus and Modelo Formations, based on seismic traversing and available well control.

The sonic logs and identified horizons from the industry seismic reflection profiles guide assignment of modeled velocity layers from the LARSE II transects to geologic units (Fig. 10). All of the models are forward models of the velocities. Layers 1 and 2 of the refraction models (Fig. 10), having velocities ranging from 1.3 to 1.9 km/s, are interpreted to be unconsolidated or weathered sedimentary rocks. Layer 3 of the refraction model (Fig. 10A), which is represented by velocities of 2.3–2.7 km/s, corresponds closely to the north-dipping reflective rocks of industry line 2893-O (Fig. 8A) that are identified to include parts of the Saugus Formation (Sunshine Ranch Member), Fernando (sometimes referred to as Pico) Formation, and Modelo Formation. The sonic log from the Chevron Frieda Clark well (Fig. 4) in the central part of the valley (Fig. 2; Table 1, drill hole 25) indicates velocities for the Sunshine Ranch Member, Fernando Formation, and Modelo Formation of ∼2.3–2.8 km/s, in agreement with the refraction velocities for layer 3. Rocks penetrated below the Modelo Formation in this well, at a two-way traveltime of 1.8 s, are the Topanga Formation, with velocities of 3–4 km/s, in agreement with refraction velocities for the upper part of layer 4. Boreholes nearer the LARSE II main line (boreholes 23 and 24; Fig. 2; Table 1) penetrate the Topanga Formation at ∼2 km depth, in agreement with the refraction model.

Layer 5 of the refraction model, having velocities generally close to or >5 km/s (Fig. 10A), is interpreted nominally as basement, which includes crystalline rocks and the Santa Monica Formation. At the crest of the Santa Monica Mountains, this layer approaches the surface in cross section (Fig. 10A) and appears to correspond to the closest approach (250 m) in map view of outcrops of Santa Monica Formation to the LARSE II main line (Fig. 1). Velocities approaching 5 km/s in the Leadwell and Hazeltine wells in granitic basement rocks support the assignment of layer 5 as basement (drill holes 28 and 29, Figs. 2 and 4; Table 1).

Forward seismic refraction models of lines 3000 and 4000 (Figs. 10B, 10C) have layers similar to the main LARSE II line. Layer boundaries along these lines agree well with those along the main line at the line junctions (Fig. 10A). The largest disagreement is the depth to the top of layer 5 at the junction of line 4000 and the main line (0.5 km), and structural complexity at this junction (a step) leads to ambiguity in interpreting the depth to the top of layer 4. On line 3000, modeled layer 3 coincides with the reflective rocks on industry line 2893Q, in a projection of this line 1.2 km downdip onto line 3000 (Fig. 10B). This correspondence matches that between layer 3 on the main LARSE II line and reflective rocks of industry line 2893O (Fig. 10A).

The correspondence between refraction layer 3 and reflective rocks breaks down on line 4000, onto which industry line 85–9, 500 m south, is superposed (Fig. 10C). Line 4000 crosses the Northridge Hills fault and anticline at an acute angle (Fig. 2), and clearly the modeled refraction layer geometry is too simple to capture the structure here. Wells along and near line 4000 and the reflection data on line 85–9 provide the best detail of the structure, although the first-order structure shown by the LARSE transects is supported by the general agreement between the observed gravity variations along these transects and those predicted by converting the velocities to density using the relationship of Gardner et al. (1974).

Gravity and magnetic models of the main basin-bounding structures (Figs. 11–13) were constructed using the surface geology (Yerkes and Campbell, 2005), the oil test well data (Shields, 1977; California Division of Oil, Gas and Geothermal Resources, 1982; Tsutsumi and Yeats, 1999; Yeats, 2001), and the seismic reflection profiles. Physical-property measurements from sonic and density logs and from core and rock samples guided the choice of density and magnetization for the various geologic units (see Langenheim et al., 2000, for more details). This information provided an initial estimate of model parameters into a 2.5-dimensional gravity and magnetic modeling program based on generalized inverse theory. The amplitude of the anomaly is not the only attribute to match; matching its gradients and inflections are critical parameters that provide important information on the depth to the top of the source and its shape. Application of the gravity and magnetic data to our interpretation of the structure of the San Fernando Valley region is discussed in the following.

These various data sets (surface and well geology, seismic reflection profiles, LARSE lines, and gravity and magnetic data) were used to construct cross sections illustrating the sedimentary architecture and tectonic framework of the valley (Fig. 14).

STRUCTURE OF THE SAN FERNANDO VALLEY

The seismic reflection, gravity, magnetic, and LARSE II data provide insights into Miocene and younger deformation within the San Fernando Valley. Surrounded by structural complexities produced by transtension, transrotation, and subsequent shortening (e.g., Ingersoll and Rumelhart, 1999), the main San Fernando Valley, Simi Hills, and adjacent Santa Monica Mountains represent a relative island of calm that preserves, with minimal late Neogene disruption, its geologic history from Late Cretaceous to the present.

Miocene Rifting

Geophysical data reported herein reveal a pattern of Middle Miocene extensional faulting that produced the Simi Hills, the Leadwell high, and similar smaller features. The North Leadwell fault zone (Figs. 7 and 8C) is a significant structure identified by this study. Although the Leadwell high was found by the Leadwell and Hazeltine core holes drilled in 1963, its nature and boundaries were unknown at that time. Seismic reflection profile 85–7 (Figs. 2 and 8C), made available by industry in the 1990s, images the granitic basement of the Leadwell high offset by a pair of down-to-the-north normal faults with 1000 m of structural relief at the base of the Modelo Formation (Wright, 2001a; Fig. 9B), but no offset of the top of the Modelo Formation. That seismic profile and corehole data show the Topanga Formation thinning and wedging out northward onto the crest of the Leadwell high, and the Modelo Formation thinning northward and overlapping the Topanga to directly overlie granitic basement (Fig. 14B). Because the Modelo Formation consists here of low-density hemipelagic and biogenic sediments, the velocity contrast between these organic, often diatomaceous, strata and granitic basement (Fig. 4A; Leadwell) produces a characteristic strong reflection over the Leadwell high (e.g., Figs. 8C and A13). Profile 85–7 also shows thinning and gentle folding of the Saugus and uppermost Fernando Formations across the fault edge, perhaps caused by the onset of compressional shortening in late Pliocene time.

The orientation of the North Leadwell fault zone and its relationship to other structures are revealed by the filtered isostatic residual gravity map (Fig. 7). That map clearly shows the west-northwest strike of the North Leadwell fault zone. We speculate, based on the filtered gravity, that another basement high east of the Leadwell high and north of the North Leadwell fault zone may be offset along that fault in a right-lateral sense from the Leadwell high (Fig. 7).

Once clearly seen on profile 85–7 and guided by the filtered gravity, the North Leadwell fault zone and the characteristic Topanga-basement contact reflection were recognized on profiles 84–202 and 85–12 (Figs. 8B, A11, and A10). The Topanga-basement contact reflection, well imaged at the south end of profiles 84–202 and 85–12 at ∼2 s traveltime, is also seen on profile 2893P (Fig. A3) near profiles 84–202 and 85–12, defining the northwest corner of the Leadwell high. Profile 85–10 (Fig. A5) shows an apparent down-to-the-east fault affecting the base of Modelo Formation that may be part of this fault zone; it is seen, though less clearly, to the southeast at ∼3 s traveltime on profiles 85–9 (Fig. A4) and 85–12 (Fig. A10), and to the northwest on profile 85–11 (Fig. A8), and is mapped in Figure 9B.

On the northwest margin of the Leadwell high, a northeast-trending gradient on the filtered gravity map (Fig. 7) may represent the Van Nuys fault, a proposed Miocene normal fault (Yeats, 2001) invoked to explain the absence in the Leadwell and Hazeltine coreholes of Paleogene and Cretaceous strata that are at least 2300 m thick (Yeats, 2001; his fig. 3) west of the Leadwell high. Alternatively, the eastward termination of those strata may be due to truncation beneath the overlying Modelo Formation. A pronounced angular unconformity at the base of the Modelo Formation is imaged on seismic reflection profile 2893Q (Figs. 10B and A1), and less clearly farther north on seismic profile 2893AC (Fig. A2). Approximately 4 km to the southwest of 2893Q on the northern flank of the Santa Monica Mountains, the Modelo Formation also overlies in angular unconformity folded Topanga and older rocks (Fig. 1; Yerkes and Campbell, 2005). The basal Modelo unconformity imaged on profile 2893AC (Fig. A2) extends eastward into the zone of the proposed Van Nuys fault, and this horizon has been mapped without offset onto the Leadwell high (Fig. 9B). However, there may be a pre-Modelo Van Nuys fault, imaged in the filtered gravity map, parallel to and coeval with the Chatsworth faults described in the following.

The western margin of the San Fernando Basin involves the northeast-striking Chatsworth Reservoir and North Reservoir faults (Fig. 2), which are Miocene normal faults that separate the Simi Hills from the San Fernando Valley. Those faults have been mapped in outcrop and in the subsurface (Shields, 1977) and coincide with a steep gravity gradient (gray circles in Fig. 5), where the faulting juxtaposes southeast-dipping Paleocene to Miocene rocks on the southeast against northwest-dipping Cretaceous rocks on the northwest (Fig. 12). The filtered gravity map (Fig. 7) clearly shows the correspondence of the Chatsworth faults with the gravity gradient along the edge of the Simi Hills, both extending southwest as far as the area where the Simi Hills block is overlapped unconformably by rocks of the Modelo and Topanga Formations. In that area the gravity gradient deviates to the south, the mapped trace of the Chatsworth Reservoir fault begins to strike more westward, and a significant gravity high between the two suggests a buried geologic feature of unknown nature. To the northeast, the Chatsworth faults do not extend beyond the area of the Northridge Hills fault (Shields, 1977; Yeats, 2001).

The boundary between the northwest-dipping homocline of Cretaceous and Paleogene rocks of the Simi high and Pliocene and younger sediments of the east Ventura Basin is shown variously as the Brugher fault and the subsurface Frew fault (Yeats, 1987a) (Fig. 9B). At its eastern end, adjacent to the San Fernando Valley, this homocline and associated faults are overlapped with profound unconformity by the Saugus Formation, further obscuring the already complex structure. The Brugher and related faults intersect the Chatsworth faults at an acute angle that is especially well shown on the filtered gravity map (Fig. 7). Structure contours on the Chatsworth Reservoir fault (Shields, 1977; Fig. 9B) suggest that it should be imaged on reflection seismic profile 85–9 (Figs. 10C and A4). Although this portion of the profile is complicated by its acute angle of intersection with the axis of the Northridge Hills anticline, the profile, west of the intersection with line 85–11, may image the two Chatsworth faults beneath the horizon mapped as the top of the Modelo Formation.

Another seismic reflection profile (85–10; Fig. A5) intersects the eastern corner of the Simi Hills block and, based on well control directly on the eastern portion of the seismic line and ties via other seismic lines to wells farther east, clearly shows the stratigraphic relationships. Eocene strata (possibly bounded on the east by the Devonshire fault) are overlain by a thin interval of Topanga Formation. The Modelo Formation thins westward onto the Simi Hills block. Overlying the Modelo, the Fernando, Sunshine Ranch, and upper Saugus units thin and wedge out against the Simi Hills block.

The east corner of the Simi Hills block forms the western termination of the Northridge Hills anticline and blind thrust. Those relationships are discussed herein.

Southern Margin of the San Fernando Valley

Basement along the southern margin of the valley slopes to the north, as indicated by decrease in gravity values north of the Santa Monica Mountains (Fig. 5) and by deepening of layers 3 and 4 along the LARSE II main line (Fig. 10A). Seismic reflection profiles in the southwestern part of the valley (2893O, Fig. 8A; 2893AD, Fig. A6) illustrate onlap of Modelo and younger strata onto older rocks exposed in the axis of the Santa Monica Mountains (Figs. 14A, 14B). The gentle northward slope extends as far north as profile 2893AC. A similar gentle northward slope in Modelo and younger strata is seen along seismic reflection profile 85–7 (Fig. 8C) in the southeastern part of the valley.

The uniform decrease in gravity values from the Santa Monica Mountains northward across the San Fernando Basin cannot be modeled with a gently north dipping Miocene section unless one assumes that the Miocene sedimentary strata have densities of 2000 kg/m3 or less or that lower density basement is present beneath the southern part of the basin. Based on physical property data, lower density granitic basement is the preferred solution, and is shown in Figure 11 (see Langenheim et al., 2000, for more detail). Some support for lower density basement beneath the southern part of the valley is from Pujol et al. (2006), who found lower basement tomographic velocities beneath the southern part of the San Fernando Valley.

The gravity data, LARSE line 3000, and east-west profile 2893Q (Figs. 10B and A1) reflect pre-Modelo structure (Yeats, 2001) along the southern margin of the valley that is concealed beneath the gently tilted section of the Modelo and younger rocks. Density boundaries (gray circles, Fig. 5) show that the edge of the gravity high associated with exposures of the Santa Monica Formation protrudes several kilometers northward into the San Fernando Valley just east of the LARSE II transect. The LARSE 3000 transect obliquely crosses this gravity high, and in general indicates westward deepening of the basement (layer 5 in Fig. 10B) as the pre-Modelo package thickens below an unconformity. West-dipping reflections beneath the unconformity on seismic reflection profile 2893Q (Figs. 10B and A1) correspond to layer 4 of the LARSE refraction model and are likely correlative to Cretaceous (and possibly Topanga) strata that are exposed a few kilometers to the south and that have west dips of 24°–25° (Yerkes and Campbell, 2005). The relatively flat lying package of reflections west of the LARSE II main line between 1 and 2.5 km depth (Fig. 10B) may be related to the stack of low-angle detachment faults mapped south of the profile that involves Middle Miocene sedimentary and volcanic rocks; alternatively, they may be multiples.

Schnurr and Koch (1979) proposed a major east-west–striking fault separating the Santa Monica and San Gabriel basement terranes along the southern margin of San Fernando Valley. Magnetic data shed light on this proposed fault because of the prominent contrast in magnetic pattern between these two terranes, i.e., broad, low-amplitude magnetic anomalies caused by weakly magnetic rocks exposed in the eastern Santa Monica Mountains (an anomaly pattern that is indistinguishable from that of the San Fernando Valley and the Simi Hills) versus the prominent magnetic anomalies produced by magnetic rocks in the San Gabriel and Verdugo Mountains (Fig. 6). The absence of a prominent east-west magnetic boundary argues for no major east-west–striking fault separating the Santa Monica and San Gabriel terranes along the southern edge of San Fernando Valley.

Additional support for geologic continuity between the San Fernando Valley and the Santa Monica Mountains directly to the south, at least since the end of Middle Miocene time, comes from continuity of features mapped in outcrop to those imaged on the seismic reflection profiles. The western lobe of the Tarzana fan (Sullwold, 1960; Fig. 1) can be projected from outcrop to the south end of reflection seismic profile 2893O (Fig. 8A), across a gap of only 1 km and at a uniform angle of dip. Profiles 2893Q (Figs. 10B and A1) and 2893AC (Fig. A2) show patterns of turbidite sand bodies very similar to those shown on Sullwold's (1960) east-west restored section across the western lobe of the fan. A gentle, north-plunging anticlinal fold at Chalk Hill (Fig. 9B) involves the Modelo western lobe outcrop, and this fold is seen on profiles 2893AC and 2893Q, providing yet another tie across the southern boundary of the valley.

A pre-Modelo basement high (Fig. 1) exposing the Santa Monica Formation separates the western and eastern lobes of the Tarzana fan (Sullwold, 1960) and is clearly imaged on the filtered gravity map (Fig. 7), southwest of the Leadwell high. Turbidites would have swept southeast through the bathymetric low between the Santa Monica and Leadwell basement highs, and paleocurrent directions in the eastern lobe are to the east and southeast (Fig. 9B).

Granitoid rocks of the Nichols pluton (Fig. 1) are dated as 102 ± 10 Ma (L.T. Silver, inDibblee, 1982). The Leadwell basement was dated as 120 Ma by the Sr/Sr method (R. Kistler, oral commun. to T. McCulloh, inWright, 2001b), with no error range given. Thus, it is possible that these two granitoid bodies could be the same age and limit any significant faulting across the southern boundary of the San Fernando Valley to be older than these rocks.

The only significant fault along the southern margin of San Fernando Valley is the Benedict Canyon fault (Fig. 1), a northeast-striking fault with ∼2 km of left separation of the Cretaceous-basement contact. This fault may extend east beneath the San Fernando Valley, based on a fault penetrated in a Los Angeles Department of Water and Power corehole (CS-01; Fig. 2) that found quartz diorite faulted over “dark brown limey marine shale” (C. Plumb, 2001, personal commun.), presumably Modelo Formation.

A broad magnetic high beneath the eastern Santa Monica Mountains and southwestern San Fernando Valley is possibly caused by the northward thrusting of magnetic basement rocks associated with the La Cienega block in northwestern Los Angeles Basin (LC in Fig. 6; Langenheim et al., 1994) and is modeled at depth (6 km and deeper) beneath the southwestern San Fernando Valley in the southeastern part of profile E-E′ (Fig. 12).

Eastern Margin of the San Fernando Valley

The Verdugo fault forms the eastern margin of the basin (Figs. 1 and 14C). Near the city of Burbank (Fig. 2), the fault forms a series of southwest-facing scarps (s in Fig. 2), which offset deposits of probable Holocene age (Weber et al., 1980). To the northwest, the fault is concealed, except for an exposure in a gravel pit southeast of the Pacoima Hills (Fig. 2; Tsutsumi and Yeats, 1999). The fault coincides with a groundwater-level change (Weber et al., 1980). These features are aligned consistently along a northwest trend, which also coincides with steep gravity and aeromagnetic gradients (Figs. 5 and 6). We argue that the Verdugo fault thus apparently forms the southwestern margin of a highly magnetic block that extends from the Verdugo Mountains northeast to the Vasquez Creek fault of Powell (1993).

Exploratory drilling leading to the 1974 discovery of the Pacoima oil field (Fig. 2) provided the first evidence of the position and geometry of the Verdugo fault in the subsurface (Schnurr and Koch, 1979; Wright, 2001a). Two wells (Tsutsumi and Yeats, 1999, Fig. 4g therein) penetrated the Verdugo fault and defined a fault dipping 55° to the northeast, with a tightly folded northwest-trending anticline of Modelo Formation in its hanging wall. In the footwall, the Pacoima field is trapped in a gentle west-trending anticline at the east end of the Northridge Hills anticlinal trend (Fig. 9).

None of the industry seismic reflection profiles crosses the Verdugo fault, but LARSE II line 4000 shows an eastward shallowing of the basement toward the fault, which is consistent with the gravity data (Fig. 10C). Tomography from the LARSE line 4000 shows slower velocities beneath Pacoima Hills (Thygesen, 2003), in agreement with modeling of the gravity and magnetic data (see following).

Because of the clear expression of the Verdugo fault in the gravity and magnetic data, we use these data to examine the geometry of the fault. Two models (Fig. 13) across the eastern margin (F-F′ and G-G′, Figs. 5 and 6) show that the Verdugo fault changes character along strike. The northwest-trending gravity gradient associated with the fault is interrupted by a north-south–striking gradient near point X (Fig. 5). Northwest of this point, the inferred fault trace is near the bottom of the northwest-trending gradient, implying a northeast dip. Southeast of X, the fault trace is near the top of the gradient, suggesting a southwest dip. Aeromagnetic data southeast of X indicate a 5-km-wide gradient southwest of the fault; this gradient narrows northwest of point X (Fig. 6).

Southeast of point X, the model along profile F-F′ (Fig. 13A) addresses the geometry of the Verdugo fault, where it is located by scarps. Modeling across the Verdugo fault suggests that the fault dips steeply (>60°) to the southwest (Fig. 13A), and thus may not be a thrust fault as shown on a schematic cross section by Weber et al. (1980), whose northeast-dipping attitude for the Verdugo fault is based on foliation in the crystalline basement rocks exposed in the Verdugo Mountains. Weber et al. (1980) showed a fault with alluvium on crystalline bedrock with a southwest dip, but they argued that it is a minor fault and not indicative of the geometry of the major basin-bounding structure. An alternative explanation for the structure shown in F-F′ is that the apparent normal geometry is the result of a stack of several thrust faults stepping southwest away from the mountain front (inset in Fig. 13A). The magnetic terrane boundary, however, dips steeply to the southwest and extends to depths of 10 km according to our model. This scenario would arise from superposition and reactivation of faults, with the shallow faults being the result of Pliocene and younger transpression and the deeper structure being the result of Miocene transtension.

Northwest of point X, along profile G-G′ (Fig. 13B), modeling of the gravity and magnetic data indicate a thrust or reverse geometry for the fault bounding the Pacoima Hills; only a sliver of basement rocks in the upper 3 km is present in the hanging wall of the northeast-dipping fault, consistent with the low-velocity zone beneath the Pacoima Hills imaged on the tomographic model for LARSE line 4000 (not shown; Thygesen, 2003).

Model G-G′ is nearly coincident with a cross section by Tsutsumi and Yeats (1999) that showed the Verdugo fault as a northeast-dipping reverse fault. They suggested that the reverse fault is a reactivation of a Miocene normal fault, based on a thicker Miocene section on the upthrown block. Their cross section (Tsutsumi and Yeats, 1999, Fig. 4g therein) is not a straight line, but jogs to the northwest of the Pacoima Hills and does not cross the gravity and magnetic signature associated with the Verdugo fault. Both gravity and aeromagnetic data indicate abrupt deepening of the basement floor along the projection of the Verdugo fault northwest of the Pacoima Hills, consistent with its northwesternmost penetration by well data, which indicate basal Modelo Formation at a subsurface depth of ∼1.5 km. The abrupt deepening of the Neogene section across what has been called the Pacoima segment boundary (Tsutsumi and Yeats, 1999) is difficult to explain. It is not likely a Miocene cross fault because the Modelo Formation east of it (Fig. 14C) is 1500–1800 m thick, nearly the maximum recorded thickness in the region, and includes Tarzana fan sands deposited on the basin floor, It thus is post-Modelo in age and formed after the Verdugo fault ceased to be a major element of the San Andreas right-slip system.

The Verdugo fault is best viewed as a relict Miocene transrotational or transtensional fault that was reactivated as a transpressional fault during the Pliocene and Quaternary. At the northern end of the Verdugo Mountains, the fault was down to the northeast during Modelo deposition (Tsutsumi and Yeats, 1999; Fig. 14C). Projected to the northwest, the Verdugo fault is colinear with a similar abrupt change in Modelo thickness in the east Ventura Basin across a northwest-trending hinge line (Yeats et al., 1994, Fig. 8 therein). This lends further support to the postulated connection between the Verdugo fault and precursors to the northern San Gabriel fault. Yeats and Stitt (2003) continued the Canton fault across the Santa Susana fault into the San Fernando Valley, and suggested that it followed the Verdugo fault at the range front of the Verdugo Mountains, where it separates weakly magnetic rocks to the southwest from magnetic rocks to the northeast. Along-strike changes in dip direction can be characteristic of faults with significant horizontal slip. The change in dip direction along strike of the Verdugo fault as imaged in the magnetic data can be explained if the fault is a continuation of the Canton fault and thus had significant lateral displacement in addition to vertical offset. In the east Ventura Basin, the Canton fault may have accommodated from 15 to 17 km (Powell, 1993) to as much as 35 km (Crowell, 2003) of right-lateral displacement between ca. 16 and 10 Ma. Thermochronologic (U-Th/He) data (Arkle, 2008) from the Verdugo Mountains indicate rapid cooling starting ca. 17–13 Ma; this may reflect the initiation of slip on the Verdugo-Canton fault. These data do not rule out a second, more recent phase of exhumation during the Quaternary.

It is difficult to follow the Verdugo-Canton fault northwest of the Pacoima Hills, as its Miocene and early Pliocene trace is buried beneath several kilometers of Saugus and Quaternary sediments, and any surface traces would have been quickly erased by active alluvial-fan deposition. On both sides of the Verdugo-Canton alignment, localized blocks containing Saugus Formation younger than 2.3 Ma (Levi and Yeats, 1993, Fig. 1 therein; Tsutsumi and Yeats, 1999) have been rotated clockwise by as much as 34°. This may reflect Quaternary activity on the Verdugo-Canton alignment in response to shortening. Geodetic data indicate a continuation of shortening and right-lateral shear (2.1 ± 1.3 mm/yr) along the Verdugo fault (Walls et al., 1998). Based on groundwater-table data, Weber et al. (1980) suggested that the Holocene Verdugo fault may curve westward and join or merge with the Mission Hills fault, a suggestion adopted by Tsutsumi and Yeats (1999). A subtle, low-amplitude magnetic high that follows the Verdugo fault as it curves to the west to the Mission Hills fault (Fig. 6) is consistent with this interpretation.

The structure of the hanging wall of the Mission Hills fault (Fig. 14B) at its postulated juncture with the Verdugo fault features a tight, faulted anticline (Wright, 2001a) with a rootless wedge of Modelo Formation in its core. The Modelo in outcrop here and in the nearby Chevron Rinaldi 1 (well 38, Table 1) includes both sands of the Tarzana fan facies and organic silts and shales, similar to the Modelo strata in continuous subcrop 600–1000 m below the Mission Hills wedge, which was dragged up vertically or horizontally along one or both of the intersecting faults.

Santa Susana Fault

The Santa Susana Mountains form the northern boundary of the San Fernando Basin west of the Verdugo fault, and have been elevated on the Santa Susana fault since ca. 0.5 Ma (Levi and Yeats, 1993; Yeats et al., 1994). The Santa Susana fault zone extends along the southern and eastern edges of the Santa Susana Mountains to its eastern end at San Fernando Pass (SFP in Fig. 1), where it links with the Sierra Madre fault in a zone of very complicated structure. The two faults are differentiated by their dissimilar structural styles: the Santa Susana is a thrust fault that flattens near the surface and at depth (Fig. 14A), with sedimentary rocks in its hanging wall to a depth of ∼5 km; the Sierra Madre fault is a reverse fault (also labeled as Hospital fault in Fig. 14B) with basement rocks in its hanging wall. At the surface near San Fernando Pass (SFP in Fig. 1) a northeast-striking cross fault at the western end of the basement outcrop marks the boundary between the two faults. At depth, that boundary may be marked by a northwest-trending gravity gradient that coincides with the Sierra Madre fault to the east. Gravity data suggest that basement found in the Mobil Macson Mission well (well 52, Fig. 2) may be a sliver caught up in this complex area of faulting or along the deep Verdugo-Canton fault. None of the seismic reflection profiles crosses the Santa Susana fault, although the LARSE II main line (Figs. 1 and 10A) indicates a basin depth of ∼5 km (Fuis et al., 2003; Lutter et al., 2004).

Tectonic evolution of the Santa Susana and related faults is relevant to new interpretations of the San Fernando Valley presented in this study. The east Ventura Basin opened as a southeast-trending rift during Middle and Late Miocene time (Yeats et al., 1994) and accumulated a thick sequence of Topanga and Modelo strata. The southwest margin of the Miocene trough is marked by an old north-dipping normal fault along the zone of the present Santa Susana fault. South of that hinge line the Modelo has a maximum thickness of 1070 m (Yeats et al., 1994), compared to as much as 2000 m in the axis of the trough. The latest Miocene Towsley Formation is restricted to the north side of that proto–Santa Susana normal fault (Huftile and Yeats, 1996). During deposition of the Fernando Formation (Pliocene), that hinge line shifted south a short distance to the ∼24-km-long Frew fault (Fig. 9B) (Yeats, 1987a). The Frew is similar to an ancestral Oak Ridge fault (inset in Fig. 1) farther to the west in that both separate thick sequences on the north from thin sequences on the south.

A change to contractile tectonics occurred during the Pliocene, and progressive shortening across the east Ventura Basin inverted the Miocene basin and began to form the Santa Susana Mountains. Beginning after 2.3 Ma, the Santa Susana fault was initiated, and rocks within the east Ventura Basin were thrust southward across the Miocene and Pliocene hinge lines. The first occurrence of Modelo clasts in the Saugus Formation at 600–700 ka (locally called the Pacoima Formation) indicates that uplift of the Santa Susana Mountains had begun by that time (Huftile and Yeats, 1996), but probably not before.

The Frew fault has been mapped (Yeats, 1987a, Fig. 9.5 therein) to within 2–3 km of the north edge of the San Fernando Valley (Fig. 9B) and to within ∼7 or 8 km of the mapped traces of the North Leadwell fault system. At the depth of the base of the Modelo depth (∼2 km), the northernmost strand of the North Leadwell faults projects readily into the 2 km contour on the Frew fault. The more southern faults of the North Leadwell system can be projected in similar fashion to the Brugher and Devonshire faults. A northwest-striking fault was mapped beneath the northwest part of the Northridge Hills blind fault (Wright, 2001a) on lines 85–12 (Fig. 8B) and 85–11 (Fig. 8B) on the basis of reflections between 1 and 2 s traveltime that appear to step down to the north, as do the strands of the North Leadwell fault zone, and appears to extend northwest toward the Brugher fault. In this area the linearity of older faults might be affected by deformation against the corner of the Simi Hills block.

These faults might form a continuous Miocene–Pliocene fault system, all initially down-to-the-north normal faults, as preserved at its eastern end in the North Leadwell fault system (and far to the west on the Oak Ridge fault (Yeats, 1987b, 1988). In the intervening area, however, this fault system could have been rotated into reverse faults, even as the Frew fault, during compressive shortening across the deep trough of the east Ventura Basin. The Oak Ridge fault was also described (Yeats, 1988; Yeats and Huftile, 1995) as a reactivated normal fault, and in Yeats and Huftile (1995), it was suggested that the 1994 Northridge earthquake was on the Oak Ridge fault system, which includes the Frew and Brugher faults (at least its older part). Along its full length, this zone would have acted as a depositional hinge line with sediments thinly deposited on its south side, especially at sea knolls, such as the Leadwell high, Simi Hills, and South Mountain (SM in inset of Fig. 1; Yeats, 1965), and thick sedimentation to the north. The Frew fault has been inactive since some time during Saugus deposition, and the North Leadwell system has been inactive since Late Miocene time.

The curvature of this Miocene fault system, trending west-northwest in the San Fernando Valley and adjacent to the Santa Susana fault, to west-southwest at South Mountain (SM in inset of Fig. 1) and beyond, is a feature that calls for further study (see Fig. 4 in Yeats, 2001). The more important question is whether the blind fault source for the 1994 Northridge earthquake will have counterparts westward along this fault system.

Sylmar Subbasin and Sierra Madre Fault

The Sierra Madre fault bounds the north side of the deepest part of the San Fernando Valley (Sylmar subbasin), as indicated by the Sun Exploration Stetson-Sombrero drill hole (well 35, Table 1 and Fig. 2), which bottomed in Saugus gravel at a depth of 3666 m. A thrust geometry for the faults bounding the northern margin of the basin is clearly expressed by the extension of low gravity values (–20 mGal) northward of these faults onto basement rocks (Fig. 5), compared to typical values of –10 mGal or greater over basement rocks throughout the area. This observation suggests that low-density material is thrust beneath the crystalline basement rocks and is consistent with geologic mapping (Oakeshott, 1975). Our gravity model of the northern basin-bounding fault (Fig. 11) shows that the fault has an apparent dip of ∼15° in the upper 2 km, steepening to ∼70° as it intersects the San Gabriel fault (Fig. 1). The difference between the fault geometry from the gravity model versus that of cross-section B-B′ to the east (Fig. 14B) highlights how basement becomes less involved in the thrusting to the west. The deep basin projects beneath the surface trace of the San Gabriel fault at this location; Fuis et al. (2003) showed that the structure changes 5 km to the west, as the Santa Susana thrust system steps to the southwest, away from the Sierra Madre and San Gabriel faults.

Gravity modeling along D-D′ (Fig. 11) demonstrates that basin strata exceed 5 km in thickness in the northern part of the basin and may be as great as 8 km. The inferred basin is significantly deeper than that estimated by Corbató (1963) or Oakeshott (1975), but is consistent with P-wave seismic tomographic results of Pujol et al. (2006). The gravity low here is more pronounced than at the LARSE II main line, 5 km to the west, which indicated a basin depth of 5 km. The greater depth extent along D-D′ results in part from using higher densities for the Pliocene strata as constrained by core density measurements (Corbató, 1963). Velocity modeling of the cross lines of the LARSE II study supports the use of higher densities north of the Mission Hills fault (Thygesen, 2003), with higher velocities in the upper sedimentary layer along the northernmost line (LARSE II line 5000; Figs. 1 and 2) than modeled along LARSE II lines 3000 and 4000 to the south.

The isostatic gravity map (Fig. 5) shows a large (amplitude of ∼70 mGal), east-west–trending gravity low over the Ventura Basin. That low curves without interruption into the San Fernando Valley gravity low, leading us to suggest that the main Middle Miocene basin-bounding structures that formed the east Ventura Basin (Yeats et al., 1994) extend into the northern San Fernando Valley. The position of the 50 mGal gravity low over the northern San Fernando Valley indicates the asymmetric nature of the valley, with the deepest portion of the basin in its northern part (Figs. 11 and 14B).

More than half of the ∼8 km section in the Sylmar subbasin (Fig. 14B) is Saugus gravel and sand; the Modelo Formation is ∼2000 m thick based on thicknesses in the Merrick syncline to the east and the east Ventura Basin to the northwest; and the remaining 700–800 m are Fernando and Towsley Formations. Thus, maximum subsidence in the Sylmar trough was during the Quaternary, at the same time as maximum uplift of the Santa Susana Mountains directly to the west. Downfolding of the Sylmar subbasin appears to have taken up some of the horizontal shortening that occurred on the Santa Susana fault to the west, exhibiting a remarkable contrast in response to regional compressive forces in adjacent areas. Perhaps the deeply buried Miocene Verdugo-Canton fault plays a role in separating these two very different modes of deformation not only during the Miocene but also during subsequent contraction. West of this alignment, extensive Miocene rifting of the east Ventura Basin produced a fundamentally different geometry in the brittle crust than found east of the Verdugo-Canton fault.

At the east end of the Sylmar subbasin, the Pacoima segment boundary separates the Sylmar subbasin from the Merrick syncline, where the depth to basement is ∼2.5 km (Figs. 13B and 14C). This implies ∼5.5 km of vertical separation across the segment boundary, nearly all of it during Saugus time. The Pacoima Wash, on the segment boundary, separates a southward- sloping alluvial plain on the west from a dissected terrace on the east, uplifted ∼30 m or more above the Pacoima Wash, on which the formations within the Merrick syncline are well exposed.

Other Structures within the San Fernando Valley

Five seismic reflection profiles cross the Northridge Hills anticline and the north-dipping Northridge Hills fault; four of them cross the Mission Hills fault. The seismic reflection profiles do not image deep enough to address whether the Northridge Hills and Mission Hills faults merge at depth into a décollement, as proposed by Tsutsumi and Yeats (1999) and Fuis et al. (2003), but they image the upper part of the Northridge Hills and Mission Hills faults. The Mission Hills fault is ∼2.5 km south of the Santa Susana fault, extends east-west for a distance of nearly 10 km, and forms the boundary between the floor of the San Fernando Valley and the foothills to the north (Dibblee, 1991, 1992). It is imaged at the northern ends of profile 85–11, 85–8, 85–12, and 84–202 (Figs. 8B and A8–A11) and separates a package of horizontal to gently north dipping reflections to the south from more steeply dipping, less coherent reflections to the north. A hanging-wall anticline is evident on profiles 85–8 (Fig. A9) and 85–12 (Fig. 8B). The fault appears to dip steeply to the north, which is consistent with well data (Tsutsumi and Yeats, 1999). Well data also indicate a relatively constant thickness of the Fernando Formation across the Mission Hills fault, suggesting that slip initiated after the end of deposition of those strata (Tsutsumi and Yeats, 1999). The lowest gravity values in San Fernando Valley occur on the hanging wall of the Mission Hills fault (Fig. 5), suggesting that this fault may have had an earlier history with the north side down.

The Northridge Hills anticline, and the Northridge Hills fault on its south flank, follow an arcuate trend ∼16 km long, extending from the Verdugo fault to the corner of the Simi Hills block (Fig. 9A). These structures were described in Tsutsumi and Yeats (1999) and imaged by the same five seismic profiles cited above, and by profile 84–205 (Figs. 8C and A12). Profile 84–205, just west of the Pacoima oil field, shows a gently folded anticline within a sedimentary sequence ∼4000 m thick. Beneath the south flank of the fold, a reflection dipping north at a low angle is believed to represent the Northridge Hills blind fault. In this area, the surface expression of the fold is blanketed by alluvial fan deposits.

To the west, the anticlinal axis first is expressed at the surface at seismic profile 84–202 (Fig. A11). The seismic profile shows an asymmetric fold with a steepened south flank above the Northridge Hills fault, which passes into bedding just above a strong reflection attributed to basement within the North Leadwell fault zone. The Modelo Formation wedges out onto the Leadwell high, and apparent growth triangles in the uppermost Fernando Formation on both flanks suggest that folding began in late Pliocene time. The fault is shown with a small displacement at the top of the Modelo Formation, but with none at the base of the Saugus Formation. Here the stratigraphic section is ∼3500 m thick.

The Northridge Hills fault was penetrated by the Chevron Woo 1 well (Fig. 2, well 30) 1.3 km west of line 84–202, within the Modelo Formation at a depth of 1870 m. Another 1.1 km farther west, profile 85–12 (Figs. 8B and A10) shows a higher amplitude anticline than that on profile 84–202. The fault slightly displaces the base of the Saugus Formation, and a subsidiary thrust above the Towsley Formation has created a fault wedge and a small anticline above the south flank of the fold. That minor fold may be reflected in the topography, where older alluvium is brought to the surface (Dibblee, 1992) to form a hill that rises 20–25 m above the surrounding plain. Here the seismic profile also shows the North Leadwell fault zone and the northwest corner of the Leadwell high (Fig. 9B). The sedimentary section north of the North Leadwell faults is slightly thinner compared to profile 84–202. On profile 85–12, the fold appears to have been pushed up and southward out of the deep, narrow low between the Mission Hills fault and the Leadwell high.

Profile 85–8 (Fig. A9) crosses the anticlinal axis in a saddle between two surface expressions of the fold. Approximately 800 m west of the profile, a slope with ∼10 m relief forms the south side of the fold, and here the base of the scarp was trenched and boreholes drilled for information on the Holocene history of the Northridge Hills blind fault (Baldwin et al., 2000). Baldwin et al. (2000) found evidence that an unconformity on the Pliocene–Pleistocene surface is warped into a monocline that has 13 ± 2 m of vertical separation across the fault. Profile 85–8 shows the fold to be tighter than on 85–12, the fault cuts somewhat higher in the section, displacement of the top of the Modelo is significantly greater, and the fault appears to be steeper at depth. In the footwall, two probable down-to-the-north faults cut the Miocene section; the northern one aligns well with the North Leadwell fault zone (Fig. 9B).

The westernmost of the north-south seismic profiles that cross the Northridge Hills structure, 85–11 (Figs. 8A and A8) crosses the surface fold, where dips to 20° on the south flank and 22° on the north flank have been mapped (Dibblee, 1992). The profile shows a much broader fold than in 85–8, but with similar complications across the shallow axis. Here the post-Topanga sedimentary section has thinned to ∼2000 m. The Northridge Hills fault is drawn significantly steeper than to the east, based in large part on nearby subsurface data (Shields, 1977), and a late Pliocene fold is imaged in its footwall. In the syncline south of the fault, two questionable normal faults within the Miocene section may be part of the North Leadwell fault system.

The five seismic reflection profiles show the changing geometry of the Northridge Hills fold and fault. The degree of folding increases westward, at least in part, as a function of the thickness of the sedimentary section. As this fold-fault structure encroaches westward against the Simi Hills block, the fault is much steeper; a cross section drawn 700 m west of profile 85–11 (Tsutsumi and Yeats, 1999, Fig. 4b therein) shows the fault dipping 77° to the north below the top of the Modelo, and flattening into a thrust at shallower depths. Still farther west, where the Saugus Formation and overlying alluvium are thrust onto the corner of the Simi Hills block, the axes of tight anticlines in the shallow lip of the fault step north, south, and north again (Dibblee, 1992), with south-flank dips as steep as 70°. At a larger scale, the arcuate nature of the anticlinal trend also seems shaped by deeper structure: its western third is parallel to the underlying North Leadwell fault system and its central part is parallel to the Mission Hills fault.

The Northridge fold-fault structure initiated sometime during Saugus deposition as a thin-skinned structure (Tsutsumi and Yeats, 1999) within the Neogene formations in the footwall of the Santa Susana fault; the steeper Mission Hills fault occupies an intermediate position between the other two faults. The Northridge Hills and Mission Hills faults may join at depth (as suggested by Tsutsumi and Yeats, 1999, Fig. 7 therein), and are being underthrust by the deep fault (Fig. 14A), which was the source of the 1994 earthquake.

Other Structures—Pliocene Anticline

Seismic reflection data (2893O, Fig. 8A; 2893AD, 85–8, Figs. A6, A8) in the west-central part of the valley indicate a northeast-plunging fold that affects the top of the Modelo Formation (Fig. 9B), but does not appear to fold the base of the Saugus Formation (Fig. 9A). The age of the fold is therefore early Pliocene. The fold may extend as far southwest as line 2893P (Fig. A3), the westernmost end of which shows the east limb of the anticline. Filtered gravity data (Fig. 7) show a diffuse high that coincides with the anticline, as mapped by the seismic reflection data, but it is difficult to map its extent with confidence beyond seismic reflection control. The crest of the anticline is penetrated by the Atlantic Richfield Mulholland and Northridge CH 1 drill holes (wells 11, 12, Table 1; Fig. 2). According to data from the operator (Thom Davis, 1995, personal commun.), the fold is cut by a pre-Pliocene down-to-the-west normal fault striking N12°W (Fig. 9B), dipping ∼70°, and with ∼200 m of separation on a marker within the Modelo Formation. The operator's interpretation shows the latest Miocene truncated beneath a post-Miocene unconformity. As shown by seismic mapping, the fold strikes northeast, parallel to the western margin of the basin. Based on the anticline's trend, we speculate that the folding may been shaped by the Chatsworth faults.

DISCUSSION

The geophysical data presented here define the geometry and stratigraphic relationships of basin-bounding and internal faults in the San Fernando Valley. In the following we discuss some of the implications of this framework for seismic hazard and tectonic evolution of this area, in particular the imprint of preexisting structures.

Basin Shape and Seismic Hazard

The geophysical and well data indicate that, in general, the San Fernando Valley is partitioned into two subbasins: a southern basin having a Pliocene and a younger section that dip to the north and a localized northern basin (Sylmar subbasin) that is filled with a thick Pliocene–Quaternary section. The Mission Hills fault and the Verdugo fault north of the Pacoima Hills form the boundary between these two subbasins.

The influence of basin shape on ground shaking is highlighted by the overlay of buildings damaged by the 1994 Northridge earthquake on the filtered gravity map (Fig. 7). In general, damaged buildings are concentrated in regions of medium to low gravity values along the western, northern, and southern margins of the basin or along gravity gradients. Some of the damage may not be related to basin effects, but rather to triggered slip along the Northridge Hills fault (Johnson et al., 1996). However, the concentration of damage in the southeastern part of the valley suggests that the Leadwell high, a Miocene structure, may have localized damage. Hartzell et al. (1997) inferred, on the basis of aftershock observations, that larger site amplifications in Sherman Oaks and Northridge were caused by structures in the upper 1–2 km, such as folds and buried basins. A shallow (500 m or less) seismic reflection profile in Sherman Oaks (blue line, Fig. 7) delineates a wedge of inferred Pliocene and Pleistocene strata that thins and terminates in the area of high earthquake damage, suggesting possible geometric focusing or basin-edge effects (Stephenson et al., 2000). This focusing does not fully explain the variation in amplification measured in the area (as much as a factor of 4), but can explain a factor of amplification as high as of 2.5 in peak horizontal velocity. The filtered gravity highlights those pockets of deeper or less dense fill in the basin that may serve to amplify shaking in the valley.

Implications for Three-Dimensional Velocity and Density Structure

Comparison of the potential-field models (Figs. 11, 12, and 13) with the Southern California Earthquake Center community velocity model 4.0 (Southern California Earthquake Center, 2010) indicates that the velocity model does an adequate job of delineating the subbasins in San Fernando Valley, especially in the eastern part of the area. The community velocity model does not work as well in the western part of the valley or at imaging the sharpness of the basin margins (Figs. 12 and 13). The model does not image folds mapped by the seismic reflection profiles that may serve to amplify ground motion, as shown by Hartzell et al. (1997) for the Northridge Hills anticline, using the geometry imaged along profile 85–12. Red-tagged buildings are also located along the margins of the northeast-trending Pliocene anticline south of the Northridge Hills area (Figs. 7 and 9), perhaps suggesting that the fold may have served to enhance local ground motions. Structure contour maps for the base of the Saugus and Modelo Formations (Fig. 9) can be used as a starting point for modeling these effects in the San Fernando Valley.

Reactivation of Miocene Structures

This study reaffirms the influence of preexisting structure on active faults. The Santa Susana and Simi faults have Miocene ancestors with a sense of displacement opposite to that in the Quaternary (Hanson, 1983; Yeats, 2001). In Tsutsumi and Yeats (1999), it was noted that the Verdugo fault was down to the north in the Miocene, the opposite of its present separation. The 1994 Northridge earthquake fault is the eastern extension of the Oak Ridge fault (Yeats and Huftile, 1995), which in the Ventura Basin was north side down in the Miocene.

Surface ruptures that formed in the 1971 earthquake within the south flank of the Merrick syncline (Figs. 9 and 14C), and some of those ruptures farther west in the Sylmar subbasin, appear to be flexural slip faults that follow bedding planes within the Modelo Formation (Tsutsumi and Yeats, 1999), rather than direct extensions of the causative fault. Heaton (1982) postulated that the 1971 earthquake was a dual event, with a second event ∼4 s after the mainshock, on another steeply dipping thrust fault coincident with the San Fernando fault, and rupturing upward from a depth of 8 km to the ground surface. This seems fully compatible with flexural slip within the south flank of the Merrick syncline and Sylmar subbasin.

The northeast-trending gravity gradient along the western margin of the valley (Fig. 5) corresponds to a northeast trend in the 1971 San Fernando aftershocks called the Chatsworth trend (Hanks et al., 1971), and coincides approximately with the northwest edge of a magnetic block (Z in Fig. 6). These aftershocks were southwest of the mainshock and were characterized by left-lateral focal mechanisms. Whitcomb (1971) suggested that this seismicity represented motion of the lower thrust block against a block to the northwest along a left-lateral strike-slip fault. The aftershocks of the 1994 Northridge earthquake also appear to follow the northeast-trending gravity gradient as it bends into alignment with the east-west–trending gravity gradients associated with the Ventura Basin margins. Hauksson et al. (1995), however, did not find evidence of reactivation of the Chatsworth trend in the 1994 Northridge aftershocks. They instead suggested that the San Fernando lateral ramp in the Santa Susana thrust fault influences the spatial distribution of aftershocks in the region. This lateral ramp is on trend with the Chatsworth Reservoir fault and coincides with a change in trend from east-west to northeast-southwest along the Santa Susana thrust fault (Fig. 2). The base of the aftershock distribution west of the ramp is 3–5 km higher than that east of the ramp (Hauksson et al., 1995), the same sense of offset along the Chatsworth Reservoir fault with east-side down. This may hint at an influence (or speculatively reactivation) of the western basin margin on the active hanging-wall structures of the 1971 and 1994 earthquakes.

The northeast-dipping fault plane for the 1971 San Fernando earthquake and the southwest-dipping fault plane for the 1994 Northridge earthquake (Mori et al., 1995) show minimal overlap, and the area of overlap is bisected by the deeply buried Miocene Verdugo-Canton fault. This may suggest a relationship between these opposing deep active faults and the contrasting modes of deformation on either side of the Verdugo-Canton alignment: uplift of the Santa Susana Mountains to the west and subsidence of the Sylmar subbasin to the east.

Implications for Palinspastic Reconstruction

Large-scale rotation of the western Transverse Ranges has been invoked in many palinspastic reconstructions of southern California, including those by Hornafius et al. (1986), Luyendyk (1991), Yeats et al. (1994), Ingersoll and Rumelhart (1999), Fritsche et al. (2001), and Yeats (2004). Another such effort is beyond the scope of this paper. Nevertheless, future iterations need to consider several elements revealed or confirmed by our study.

  1. There is no significant post–Middle Miocene faulting along the boundary between the San Fernando Valley and the Santa Monica Mountains.

  2. The southeastern (Verdugo) part of the Miocene Verdugo-Canton fault system still acts as a major feature within the crust.

  3. The Miocene Oak Ridge normal-fault system probably extends through the San Fernando Valley.

  4. The North Leadwell fault system exemplifies Middle to Late Miocene extensional rifting in the western Transverse Ranges.

  5. The east Ventura and Sylmar subbasin were a continuous depositional trough during Late Miocene and early Pliocene time, prior to the onset of latest Pliocene and Quaternary contraction.

We also add that reconstructions involving the Tarzana fan must take into account post-Mohnian rotation of the western Transverse Ranges.

CONCLUSIONS

Integration of geophysical, geologic, and oil test well data supplies new constraints on the basin geometry of the San Fernando Valley. These data rule out any significant structure along the southern margin of the valley, with seismic reflection data showing a relatively undisturbed northward-tilted Modelo and younger section. This section is not significantly deformed away from the western and eastern margins of the basin or south of the Northridge Hills thrust fault. In the southeastern part of the valley, seismic reflection and gravity data reveal a pre-Modelo basement high, bounded on the north by normal faults that extend across the valley and may connect to the Oak Ridge fault system. The Leadwell basement high appears to have influenced paleocurrent directions in the Tarzana fan and may also have influenced damage patterns from the 1994 Northridge earthquake.

North of the Mission Hills fault, the Sylmar subbasin is significantly deeper (as much as 8 km) and characterized by a thick Pliocene–Quaternary section that is thrust beneath the basement rocks of the San Gabriel Mountains. Magnetic data suggest a major boundary at or near the Verdugo fault, which strikes northwest along the east margin of the basin, and shows a change in dip sense of the fault along strike. This fault is likely the continuation of the Canton fault, a Miocene strike-slip fault, and may also have had an earlier transrotational faulting history. The western margin of the basin, as defined by gravity data, is linear and strikes ∼N45°E. The northeast-trending gravity gradient follows part of the 1971 San Fernando aftershock distribution called the Chatsworth trend, and the aftershock trends of the 1994 Northridge earthquake. These data suggest that the 1971 San Fernando and 1994 Northridge earthquakes may have been localized along portions of Miocene normal faults.

APPENDIX

For all Appendix figures, the top panel shows uninterpreted seismic reflection profile, the middle panel shows seismic reflection profile with interpretations, and the bottom panel shows gravity and magnetic variations along the profile.

This research was supported by the U.S. Geological Survey (USGS) and by the Southern California Earthquake Center (SCEC). SCEC is funded by National Science Foundation Cooperative Agreement EAR-0106924 and USGS Cooperative Agreement 02HQAG0008. This is SCEC contribution 1447. We thank Chevron Corporation for donation of seismic reflection and well data to the SCEC, the California Division of Oil and Gas, and other oil companies for providing well logs, core descriptions, paleontological interpretations, directional surveys, and dipmeter analyses. We also thank Bob Powell, Dan Scheirer, Ray Ingersoll, and an anonymous reviewer for thorough and helpful reviews.