Studies of the crustal and upper-mantle structure of California along the San Andreas fault system have been underway for more than half a century, beginning with the early studies by Byerly and Wilson (1935) and Byerly (1946) in northern California and by Gutenberg (1943) in southern California. Crustal profiling along and near the San Andreas fault was first accomplished in the early 1960's by Eaton (1963), Healy (1963), and Roller and Healy (1963). Research accelerated after the 1966 M=6.0 Parkfield, Calif., earthquake to include both detailed crustal profiling and installation of dense seismic networks for the study of earthquakes (see chap. 5; Eaton and others, 1970). Since 1970, a wide variety of seismologic methods have been used to investigate crustal and upper-mantle structure in the vicinity of the San Andreas fault system. In this chapter, we summarize the main features of this structure and relate the structure to broad-scale tectonic processes.
Figure 8.1 - Schematic block diagram of Imperial Valley region of the Salton Trough, with unmetamorphosed sedimentary rocks removed and seismic basement cut away along line approximately parallel to the Brawley seismic zone (BZ; see fig. 8.7). Seismic basement consists of rocks with P-wave velocities of 5.5 to 6.5 km/s. In this region, there are two types of seismic basement: one type, on flanks of the Salton Trough, consists of pre-late Miocene igneous and metamorphic rocks; other type, in central part of trough, consists of late Miocene and younger metasedimentary rocks (similar in age and provenance to sedimentary rocks stripped off in this diagram). Pacific and North American plates are separating across the Brawley seismic zone, an inferred onshore spreading center of the East Pacific Rise. North and south of the Brawley seismic zone, these two plates are separated from each other by transform faults, the San Andreas and Imperial faults, respectively. As the plates pull apart,
subsidence occurs within the Brawley seismic zone, sediment is deposited to fill rift from above, and mafic intrusions (basalt, diabase, and gabbro) enter rift from below, metamorphosing sedimentary rocks below a certain depth (generally approx 5 km in central part of rift). This process is repeated until central part of rift consists entirely of young crust. Geographic locations projected downward onto seismic basement for reference.
Seismologic studies of crustal and upper-mantle structure in California make use of three primary data sources: (1) traveltimes of local earthquakes as measured by permanent and temporary seismic arrays, (2) seismic-refraction and reflection profiles, and (3) teleseismic delay times measured by seismic arrays. Traveltimes of local earthquakes, in addition to containing the information needed to locate earthquakes, contain a wealth of information regarding the seismic-velocity structure of the crust and upper mantle. Velocity structure can be determined from these traveltimes by iteratively adjusting an initial velocity model and associated hypo central parameters, using inversion methods (for example, Crosson, 1976; Eberhart-Phillips and Oppenheimer, 1984). The resolution of velocity structure from local earthquake data is a function of the interstation spacing of the network and the abundance and distribution of seismicity.
Seismic refraction and reflection profiles together form a complementary set of seismic measurements. Seismic-refraction profiles provide the highest resolution of seismic P-wave velocities in the crust and upper mantle. The seismic-refraction method, however, generally does not provide the sharpest picture of lithologic interfaces, from which geologic structure is inferred; such a picture is better provided by seismic-reflection profiling.
Teleseismic delay-time studies offer the most effective means of determining the structure of the sub crustal lithosphere. The method is based on interpreting relative arrival times of compressional waves throughout a seismic array in terms of velocity variations at depth beneath the array. The Earth structure in the volume beneath the array generally is described by a series of blocks, and velocity deviations are derived for each block from the observed delay times (Aki and others, 1977; Thurber and Aki, 1987). The California seismic array is ideally suited for such investigations because of its large areal extent and the length of time it has been in operation (see chap. 5).
The primary product of the seismologic methods described above is a model of the seismic P-wave-velocity distribution in the crust and upper mantle. However, the interpretation of seismic P-wave velocities in terms of rock type is highly nonunique because laboratory velocity data indicate that numerous rock types can have similar velocities (for example, Birch, 1960). This interpretation is further complicated by the fact that, in rocks at pressures of less than 2 kbars (depths above 8 km), seismic velocities are strongly affected by the presence of cracks (on all scales) and porosity. In addition, rock velocities are affected by temperature and the presence of water. Thus, the interpretation of P-wave velocities in terms of rock types must involve other data sets, including laboratory velocity measurements on rocks at different pressures, temperatures, and water saturations, surface geologic data, well data, and other geo-physical data, including gravity and
magnetic data. Fortunately, abundant laboratory velocity data (for example, Stewart and Peselnick, 1977; Lin and Wang, 1980), geologic data (see chaps. 1, 3), and geophysical data (see chap. 9) are available for California, making the lithosphere of this region one of the best studied in the world.
In this chapter, we summarize the lithospheric structure and tectonics along the San Andreas fault system of California (fig. 8.2) with maps of crustal thickness and upper-mantle seismic-velocity anomalies, and with crustal cross sections for central and southern California. Structure changes more rapidly parallel to the San Andreas fault in southern California than in central California, and so we supplement the cross section for southern California with a map showing crustal-block motions and a diagram illustrating the different motion of the lithospheric mantle below. Seismic and other data currently are still not dense enough to construct a cross section along the San Andreas fault system itself.
Figure 8.2 - California, showing place names, geologic provinces, selected geologic units, and locations of crustal transects shown in figures 8.4 and 8.6. The San Andreas fault extends from the Salton Trough to triple junction at Cape Mendocino. CPF, Cerro Prieto fault; IF, Imperial fault; SCI, Santa Catalina Island. Fault with crosslining is trench, offshore northern California.
Construction of the crustal cross section for central California led us to a new interpretation of upper-crustal tectonic wedging, the mechanism whereby the Franciscan assemblage was emplaced in the Coast Ranges during the late Mesozoic(?) and Cenozoic. This interpretation extends that of Wentworth and others (1984) to include a two-part history whereby the observed structures atop the wedge, which include both extensional and compressional faults, were created. We further speculate that similar tectonic wedging occurred in southern California from the Mojave Desert to the Chocolate Mountains to emplace the Rand schist and the Pelona-Orocopia schist of Haxel and Dillon (1978) into rocks east of the San Andreas fault.
CRUSTAL-THICKNESS MAP OF CALIFORNIA
A contour map of crustal thicknesses in California (fig. 8.3A) provides an overview of the geophysical setting of the San Andreas fault system. The seismic and gravity data used in compiling this map (fig. 8.3B) were discussed by Mooney and Weaver (1989).
Figure 8.3 - Crustal thickness (A) for California and adjacent regions, modified from Mooney and Weaver (1989), with data sources (B). Contour interval, 2 km. Northeast of the San Andreas fault in central California, thin crust (within enclosed 28-km contour) corresponds to Mesozoic/early Cenozoic forearc basin (Great Valley; see fig. 8.2 for place names), and thick crust (within enclosed 40-km contour) corresponds to magmatic arc of same age (Sierra Nevada). Southwest of the San Andreas fault in central California, this Andean-marginal sequence is repeated but shortened; crust is relatively thin there. In southern California, crustal thickness is relatively uniform (about 30 km), despite considerable tectonic activity throughout most of geologic time, including present subduction of lithospheric mantle (see below). Estimated error in figure 8.3A is 10 percent, or 1 to 1½ contour intervals. Heavy lines, faults-dashed where the Mendocino Fracture Zone extends onshore
(and beneath North American plate); arrows indicate direction of relative movement; crosslined along trench, offshore northern California. Triangles, volcanoes of Cascade Range continental arc. Dot pattern, area of contours on Gorda-plate Moho. IF, Imperial fault. Data sources in 8.3B include seismic-refraction profiles (dotted lines), earthquake networks (wavy outlines), and gravity (dashed outlines). "W" associated with seismic-refraction profile in the Salton Trough indicates that only wide-angle reflections are available to constrain Moho depth.
Crustal thickness along the San Andreas fault increases from 16-24 km in northern California to 28-32 km in southern California. Thus, the crust along the San Andreas fault system is everywhere thinner than the 36-km average for the conterminous United States (Braile and others, 1989), and in northern California it is substantially thinner than this average. To a first-order approximation, crustal thickness resembles the topography (see Jachens and Griscom, 1983, fig. 13).
Cape Mendocino in northern California marks the change from the strike-slip regime of the San Andreas fault to the subduction regime of the Cascade Range. North of Cape Mendocino, the crust thickens eastward from about 16 km at the coast to about 38 km in the southern Cascade Range (fig. 8.3A). Near the coast, this thickness includes both the North American plate and the subducting Gorda plate. Estimates of crustal thickness in the northern Coast Ranges at Cape Mendocino lack seismic refraction or reflection control, but detailed gravity models, heat-flow observations, and teleseismic data indicate an abrupt decrease in both crustal and lithospheric thickness southward of the landward projection of the Mendocino Fracture Zone (see chaps. 9, 10; Zandt and Furlong, 1982; Jachens and Griscom, 1983).
In central California, the crust thickens eastward from about 25 km near the coast to as much as 55 km in the Sierra Nevada, but this general landward thickening is interrupted by thin crust (25 km) beneath the Great Valley (fig. 8.3A; compare Oppenheimer and Eaton, 1984). The crust of central California represents a Mesozoic and early Cenozoic Andean-type continental margin (see chap. 3; Hamilton, 1969) that has been modified by late Cenozoic strike-slip faulting along the San Andreas fault system and by uplift of the Sierra Nevada. Andean features include a subduction complex (eastern Coast Ranges), a forearc basin (Great Valley), and a magmatic arc (Sierra Nevada). Cenozoic strike-slip faulting along the San Andreas fault system has moved a shortened Andean-marginal sequence outboard of this sequence. Southwest of the San Andreas fault, the batholithic Salinian block (western Coast Ranges) is juxtaposed, across other right/oblique-slip faults of the San Andreas fault
system, against an inactive accretionary prism, or subduction complex (western Coast Ranges and offshore California).
In southern California, the crust thickens eastward from about 20 km at the western margin of the California Continental Borderland to about 32 km in the eastern Transverse Ranges (fig. 8.3A). Over most of onshore southern California, crustal thickness is 30±2 km. Considering the complex tectonic history of this region, including the present subduction of lithospheric mantle (see below), this uniformity in crustal thickness is remarkable.
Crustal structure in central California is grossly two dimensional, as can be readily inferred from the crustal-thickness map (fig. 8.3A). There are five blocks or provinces with subparallel fault boundaries: an accretionary prism, which is partly off shore; the Salinian block, which underlies the western Coast Ranges, including the Santa Cruz Mountains and Gabilan Range; a complex block between the San Andreas and Calaveras faults, underlying the Santa Clara Valley; the Diablo block, beneath the Diablo Range; and the Great Valley/Sierran block (fig. 8.2). To illustrate the crustal structure of central California, we have modified and reinterpreted the part of Centennial Continent-Ocean Transect C2 (Saleeby, 1986) that extends from offshore California at Monterey Bay to the Sierran foothills near Modesto (figs. 8.2, 8.4). Seismic control, which is exceptionally good along this transect, has been augmented since Saleeby's (1986) study primarily by analysis of
seismic-refraction profiles in the Great Valley (fig. 8.4A). The reader is referred to Hill (1978) for an earlier treatment of deep structure along approximately this same transect.
- Crustal structure of central California. A
, Surface geology, depth-converted seismic-reflection data, and models of seismic-refraction, gravity, and magnetic data along western part of Centennial Continent-Ocean Transect C2 (see Saleeby, 1986). B
, Reinterpretation of Transect C2. Major features in figure 8.4B
include, from west to east, (1) offshore, inactive, early Tertiary accretionary wedge; (2) batholithic Salinian block of the Santa Cruz Mountains, positioned between the active oblique-slip San Gregorio-Hosgri and San Andreas faults; (3) Franciscan terranes of the Santa Clara Valley and Diablo Range, interpreted to compose a tectonic wedge; (4) Mesozoic and Cenozoic sedimentary rocks of the Great Valley; and (5) rocks of the Sierran foothills, including Jurassic and older volcanic, plutonic, and related sedimentary rocks accumulated or emplaced in an island-arc setting and Cretaceous plutonic rocks. Tectonic wedge in feature 3 is interpreted to have moved
during the late Mesozoic(?) and Cenozoic, possibly in several episodes, largely along contact between Mesozoic crystalline rocks and overlying Mesozoic sedimentary rocks. In the eastern Great Valley, these sedimentary rocks are still rooted to (or depositionally overlie) this basement. Movement of wedge during present San Andreas transform-faulting regime may be along one or more thrust faults that merge with postulated
8.4 and 8.6
decollement in brittle-ductile transition zone in the crust. This reinterpretation differs from Saleeby's (1986) in eliminating inferred east-dipping subduction zone or thrust fault beneath the western Great Valley. Off shore, interpretation of unmigrated reflection section by D.S. McCulloch (in Saleeby, 1986) has been converted to depth section, using assumed velocities for each inferred geologic unit. In the Santa Cruz Mountains, velocity model consisting of layers 1 through 4 is shown (fig. 8.4A
; see fig. 8.5A
), along with alternative model in which layer 4 is subdivided into layers 4a and 4b. First model gives rise to interpretation a, and alternative model to interpretation b (fig. 8.4B
). a-j, reflectors in the eastern Diablo Range, Great Valley, and Sierran foothills; 1-4, seismic-velocity layers in the Great Valley (fig. 8.4A
) discussed in text. See figures 8.2 and 8.3 for location of Transect C2. No vertical exaggeration.
The offshore region of transect C2 is underlain by an inactive, early Tertiary accretionary prism overlapped by uppermost Oligocene to Holocene sedimentary rocks (see Saleeby, 1986). The San Simeon terrane, consisting of Late Cretaceous Franciscan rocks (disrupted marine sedimentary rocks; see chap. 3), is imbricated in this prism along with poorly known, lower Tertiary sedimentary rocks. The prism is underlain by oceanic crust with an inferred age of about 26 to 20 Ma (Atwater and Menard, 1970; Atwater, 1989). The accretionary prism is juxtaposed against granitic rocks of the Salinian block across the (inactive) Nacimiento fault, which is overlapped by upper Tertiary sedimentary rocks. This fault, in turn, is offset by the (active) right/oblique-slip San Gregorio- Hosgri fault. The Moho is 10 km below sea level near the west end of the transect (Shor and others, 1971) and deepens to 24- to 26-km depth beneath the Gabilan Range and Santa Cruz Mountains (Walter and Mooney, 1982).
We follow D.S. McCulloch (in Saleeby, 1986) in showing steep northeastward dips on both the Nacimiento and San Gregorio-Hosgri faults (fig. 8.4A) that are based on marine reflection data. Focal mechanisms for earthquakes on the San Gregorio-Hosgri fault at this latitude indicate nearly pure strike slip on vertical planes; however, farther south, they indicate chiefly reverse faulting on northeast- or southwest-dipping planes (see chap. 5).
The area between the Nacimiento and San Andreas faults is underlain by a batholithic terrane that has been transported northwestward by the San Andreas (and other?) fault(s) by amounts estimated to range from 550 km (see Ross, 1978) to 2,500 km (Champion and others, 1984). Plutonic rocks include tonalite, granodiorite, and quartz monzonite of mostly Late Cretaceous age (Ross, 1978; Mattinson, 1982). Metamorphic pendants and screens include mostly quartz-rich clastic rocks of amphibolite facies. Ross and McCulloch (1979) postulated that these upper-crustal plutonic and metamorphic rocks are not rooted to the lower crust but are in fault contact with a buried terrane, possibly consisting of Franciscan rocks.
The velocity structure derived by Walter and Mooney (1982) from Stewart's (1968) seismic-refraction measurements in the Gabilan Range and Santa Cruz Mountains can be subdivided into four separate crustal layers with velocities of 2.1-4.6 km/s (layer 1), 5.3-5.6 km/s (layer 2), 6.0-6.15 km/s (layer 3), and 6.35-6.55 km/s (layer 4). Layer 4, middle and lower crust, can alternatively be modeled as two layers of velocities 6.3 km/s (layer 4a) and 6.6-6.8 km/s (layer 4b). These layer velocities can be correlated to rock type using surface geologic data and laboratory velocity data. Layer 1 corresponds to outcrops of Cenozoic sedimentary rocks along the transect. Basement outcrops along or near the transect include abundant quartz monzonite (Ross, 1972). Lin and Wang (1980) studied the velocity behavior of a sample of quartz monzonite from this region as a function of pressure and temperature, and constructed a velocity-depth curve for this rock appropriate for the Coast Ranges.
On their curve (fig. 8.5A), the rock is slightly faster than layers 2 and 3 and slower than layer 4. Walter and Mooney (1982) interpreted layers 2 and 3 as granitic rocks similar to this quartz monzonite. The somewhat lower velocity of these two layers in comparison with the laboratory sample may be interpreted to result from (1) megascopic fractures in the Earth, not present in the laboratory sample; (2) a slightly lower content of mafic minerals (which have high seismic velocity) in the granitic rocks beneath the transect in comparison with the laboratory sample; or (3) both. Layer 4 is intermediate in velocity between the quartz monzonite sample and gabbro samples (horn-blende gabbro and olivine gabbro) from the Coast Ranges. Walter and Mooney (1982) interpreted this layer to correspond to gneiss of intermediate composition, on the basis of a comparison of layer 4 with other laboratory data. In an alternative model, however, where middle and lower crust are separated as layers 4a and 4b,
layer 4b may be reasonably interpreted as gabbro (fig. 8.5A).
Figure 8.5 - Velocity-depth curves. A, Santa Cruz Mountains and Gabilan Range (Salinian block). B, Diablo Range. Heavy curves from seismic results (Walter and Mooney, 1982); light curves from laboratory velocity measurements and heat-flow modeling (two different geotherms assumed below about 10-km depth; Lin and Wang, 1980). See text for discussion of layers shown in figure 8.5A.
Alternative interpretations of these several crustal layers are also possible, given the fact that different rock types may have similar velocities. Stewart and Peselnick (1977) and Lin and Wang (1980) studied the velocity behavior of Franciscan rocks, also common in the Coast Ranges (Jennings and Strand, 1958). Two lithologic components of the Franciscan assemblage, unmetamorphosed and metamorphosed graywacke, produce velocity-depth curves (fig. 8.5B) that bracket those for most other components of the Franciscan assemblage (including basalt). On the basis of velocity data alone, layers 2 and 3 might be interpreted as Franciscan rocks, but surface geologic data lead us to reject this interpretation. On the basis of velocity data alone, however, layer 4 is most likely not Franciscan rocks. Thus, if the middle and lower crust of the Salinian block represents a different terrane from the upper crust, as postulated by Ross and McCulloch (1979), that terrane is most likely not
SANTA CLARA VALLEY-SAN ANDREAS TO CALAVERAS FAULTS
In our cross section (fig. 8.4B), we show alternative interpretations of layer 4, given the alternative velocity models discussed above. In one interpretation (a, fig. 8.4B), layer 4 is entirely gneiss of intermediate composition. In a second interpretation (b, fig. 8.4B), layer 4a is intermediate gneiss, and layer 4b is gabbro. In interpretation a, no buried terranes are present; in interpretation b, the lower-crustal gabbro may be a buried terrane (oceanic crust) or magmatic ally underplated gabbro.
In the Santa Clara Valley, between the San Andreas and Calaveras faults, Franciscan assemblage (Permanente terrane; Blake and others, 1982) is overlain by outliers of Late Jurassic Coast Range ophiolite and Upper Cretaceous Great Valley sequence (McLaughlin and others, 1988a). The Franciscan assemblage includes melange, volcanogenic sandstone, pillow basalt, serpentine, chert, and limestone. The Franciscan sedimentary rocks were deposited in equatorial waters and presumably transported thousands(?) of kilometers northward before accretion to the North American Continent (Blake and others, 1982).
Along transect C2, the San Andreas fault juxtaposes a thick section of Tertiary marine sedimentary rocks on the southwest against slivers of Coast Range ophiolite, Great Valley sequence, and other Tertiary marine rocks on the northeast that have been imbricated along the southwest-dipping Sargent fault and related thrust faults (McLaughlin and others, 1988a). Presumably, the granitic rocks of the Salinian block and Franciscan assemblage are juxtaposed at depth. In contrast, similar rocks are juxtaposed on either side of the Calaveras fault, including Coast Range ophiolite, Great Valley sequence, and, at depth, presumably the Franciscan assemblage.
Aftershocks of the M=7.1 Lorna Prieta earthquake of 1989 indicate a steep southwestward dip (approx 70°) on the San Andreas/Sargent fault zone, and the main shock produced subequal components of strike- and reverse-slip motion (Platker and Galloway, 1989). Relatively low elevations in this region, however, indicated that the motion along this fault zone in the past has been chiefly strike slip, and seismicity before the Lorna Prieta earthquake (Olsen and Lindh, 1985; Olsen, 1986) indicates a complex fault zone that may include both vertical and southwest-dipping fault strands (see fig. 8.4B).
Although a steep (80°-85°) eastward dip on the Calaveras fault is indicated by earthquakes (see cross sec. D-D', fig. 5.7B; Reasenberg and Ellsworth, 1982; Oppenheimer and others, 1988), such an attitude is not resolvably different from a vertical dip (shown in fig. 8.4B), given errors in earthquake locations.
A seismic-refraction profile across the Santa Cruz Mountains and Santa Clara Valley reveals a heterogeneous upper crust (Mooney and Colburn, 1985). Layer offsets and velocity changes are visible in the model for this profile at the Zayante-Vergeles, Sargent, and Calaveras faults but, surprisingly, not at the San Andreas fault. An additional discontinuity is visible at an inferred buried fault in the central Santa Clara Valley (fig. 8.4A). Vertical zones of low velocity, 1 to 2 km wide, extending to a depth of as much as 3 to 5 km, are visible at a few of these faults (Mayer-Rosa, 1973; Bltimling and others, 1985; Mooney and Colburn, 1985). The surficial layer (2.1-4.5 km/s) corresponds to different rocks in different places (fig. 8.4B). The "basement" layer has a velocity (5.4-6.0 km/s) appropriate for either granitic or Franciscan rocks at shallow crustal levels (Mooney and Colburn, 1985; see discussion above and fig. 8.5); presumably, it represents Franciscan rocks
except west of the San Andreas fault. The higher-velocity basement (6.0 km/s) in the eastern Santa Clara Valley may represent metamorphosed Franciscan rocks. A strong reflector is visible at 8- to 9-km depth beneath the Santa Clara Valley, but the seismic velocity below it is unknown. By analogy with the strong midcrustal reflector in the Diablo Range (see below), we infer this reflector to be the top of accreted island-arc and (or) oceanic crust.
Moho depth beneath the Santa Clara Valley is not known accurately enough to resolve whether the Moho steps downward to the east at the San Andreas and Calaveras faults or dips smoothly eastward between control points in the Santa Cruz Mountains/Gabilan Range (24- to 26-km depth) and the Diablo Range (29- to 30-km depth). McEvilly and Clymer (1975) conducted a seismic-reflection survey across the San Andreas fault south of its junction with the Calaveras and found a crustal thickness of 24 km with no change in thickness across the fault. Peake and Healy (1977), however, indicated a change in crustal thickness at the fault in this area.
The Diablo Range, the east-central Coast Ranges between the Calaveras fault and the Great Valley, is underlain chiefly by Franciscan assemblage. These rocks constitute at least three thrust sheets or nappes that are folded into an antiform (fig. 8.4; Blake, 1981; Saleeby, 1986). The youngest thrust sheet, the Burnt Hills terrane (Blake and others, 1982; Saleeby, 1986), consists of mid-Cretaceous blueschist-facies graywacke, arkose, conglomerate, argillite, and chert, approximately equivalent in age and provenance to mid-Cretaceous forearc sedimentary rocks of the Great Valley sequence. The Burnt Hills terrane is exposed in the core of the antiform. The oldest thrust sheets are the Upper Jurassic (informal) Garzas tectonic melange (Cowan, 1974) and the Yolla Bolly terrane (Blake and others, 1982). The Garzas tectonic melange consists of mafic blueschist-amphibolite, greenstone, serpentinized peridotite, and metagray-wacke; it contains fragments of Upper Jurassic rocks (Coleman and
Lanphere, 1971; Suppe and Armstrong, 1972) similar to those accreted in the Sierran foothills during the Nevadan orogeny (see below). The Yolla Bolly terrane lithologically resembles the Burnt Hills terrane, although there are some important differences. Both the Garzas tectonic melange and Yolla Bolly terrane crop out on the flanks of the antiform. The Coast Range ophiolite and Great Valley sequence lie above the Franciscan rocks on the low-angle Coast Range fault, which is complexly offset by the steeply dipping Cenozoic Tesla-Ortigalita fault on the northeast flank of the Diablo Range (fig. 8.4). (We follow Jayko and others, 1987, in referring to the "Coast Range thrust fault" as simply the "Coast Range fault" - see below.)
GREAT VALLEY AND SIERRAN FOOTHILLS
The velocity structure of the Diablo Range derived by Walter and Mooney (1982) from seismic-refraction data collected by Stewart (1968) includes, beneath a 3.5- to 5.3-km/s near-surface layer, a 5.7- to 5.9-km/s layer beginning at 3-km depth, a 6.7- to 7.1-km/s layer beginning at 15-km depth, and the Moho at 29-km depth (fig. 8.4A). Importantly, a strong reflection is observed from the layer boundary at 15 km, indicating a strong velocity contrast between upper and lower crust. Blümling and Prodehl (1983) reanalyzed the same data and derived a similar velocity structure, except that they interpreted more phases in the data and added a lower-crustal low-velocity layer (5.3? km/s), with its base at about 26-km depth.
Seismic-reflection data have been collected in the eastern Diablo Range (Zoback and Wentworth, 1986) and compiled with other seismic data (Wentworth and others, 1987). These reflection data include a band of strong reflectors in the upper crust that dips shallowly east (reflectors a, fig. 8.4A), a weak reflector in the middle crust that dips shallowly west (reflector b), and a weak reflector at about 30-km depth (reflector c). The shallowly west-dipping reflector b, appears to link the top of the Great Valley basement with the top of the 6.7- to 7.1-km/s layer (Wentworth and Zoback, 1989).
Between 3- and 15-km depth, seismic velocities in the Diablo Range are well bracketed by velocity-depth curves predicted for end-member rocks of the Franciscan assemblage (fig. 8.5B; Steward and Peselnick, 1977; Lin and Wang, 1980). Within this depth range, the observed velocities also are slightly lower than those predicted for most granitic rocks (see fig. 8.5A). Between 15- and 20-km depth, the observed velocities agree well with those predicted for gabbro (fig. 8.5B; Lin and Wang, 1980) or, possibly, high-grade metamorphic rocks (Birch, 1960; Christensen and Fountain, 1975). The 6.7- to 7.l-km/s layer in the Diablo Range may represent the middle and lower crust of an island arc or several imbricated island arcs. If so, this layer might include mixed intermediate and mafic plutonic rocks, including compositions from granodiorite to gabbro, as well as metamorphic rocks (see description of the Coast Range ophiolite by Evarts, 1977). Its relatively high velocity
indicates that rocks of mafic composition must dominate or that the rocks are of amphibolite to granulite facies. This "island arc" interpretation is consistent with linking this layer to rocks beneath the Great Valley and thence to rocks of the Sierran foothills, which represent the middle and upper crust of island arc(s) (Saleeby, 1986). The 6.7- to 7.l-km/s layer, however, may also represent middle and lower oceanic crust, or diabase and gabbro, similar to the lowest layer of oceanic crust at the west end of transect C2 (fig. 8.4A). The 6.7- to 7.1-km/s layer is too thick, however, to represent a single layer of oceanic crust. It consists of either several slices of tectonically underplated oceanic crust or of oceanic crust that has been augmented by mafic intrusions after underplating. If the low-velocity zone of Bltimling and Prodehl (1983) is present, the 6.7- to 7.1-km/s layer may include oceanic sedimentary rocks tectonically underplated along with the oceanic crust.
We show a fault contact between the Franciscan assemblage and the 6.7- to 7.1- km/s layer in the Diablo Range to reflect the eastward transport of a wedge of Franciscan rocks (fig. 8.4B), similar to that discussed by Wentworth and others (1984). This interpretation departs from that of Saleeby (1986), who linked the shallowly east-dipping reflectors in the eastern Diablo Range (a, fig. 8.4A) with a hypothetical subduction zone or thrust fault beneath the Great Valley and Sierran foothills (see section below entitled "Discussion - Tectonic Wedging").
Rocks of the Great Valley are known from exposures in an upturned section on the east side of the Diablo Range and from wells. The upturned section rests structurally above the Franciscan assemblage on the low-angle Coast Range fault, although in many places this relation is obscured by younger high-angle faults. This upturned section includes, from oldest to youngest, Middle and Late Jurassic Coast Range ophiolite and a related tuffaceous unit; Upper Jurassic and Cretaceous Great Valley sequence, which is chiefly forearc flysch; lower Cenozoic postarc marine and terrestrial sedimentary rocks; and upper Cenozoic continental-arc sedimentary rocks (Maddock, 1964; Evarts, 1977; Bartow and others, 1985).
DISCUSSION - TECTONIC WEDGING
At the latitude of transect C2, the Coast Range ophiolite is interpreted to be a (rifted) island-arc assemblage because it contains abundant silicic volcanic and intrusive rocks (Bailey and Blake, 1974; Evarts, 1977; Hopson and others, 1981; McLaughlin and others, 1988b). Its contact with the overlying sedimentary rocks, though faulted in most places, is believed to be fundamentally depositional (Bailey and others, 1970); on transect C2 it is demonstrably depositional (Evarts, 1977).
The Great Valley sequence and younger rocks exposed in the upturned section in the eastern Diablo Range appear to be nearly twice as thick as the section of sedimentary rocks penetrated in wells farther east in the Great Valley (fig. 8.4). Some of this apparent westward thickening results from the stratigraphic addition of older rocks to the basal part of the section in the west; some may be caused by imbrication along thrust faults (Wentworth and others, 1984). Similar apparent thickening west of the synclinal axis of the Great Valley has been documented in other localities as well. In the southern Great Valley, Wentworth and others (1984) indicated an apparent doubling of thickness west of the axis, and in the northern Great Valley, an apparent trebling of thickness.
Most of the basement rocks that have been penetrated by wells in the Great Valley have been identified as granitic rocks (Saleeby, 1986). Rocks exposed in the Sierran foothills, east of the Great Valley, may be related to basement rocks beneath the Great Valley, but they are not so dense or magnetic (see below; fig. 8.4A).
Deep structure along transect C2 in the Great Valley has been elaborated in some detail by Colburn and Mooney (1986), Holbrook and Mooney (1987), and Dean Whitman and others (unpub. data, 1985) from seismic-refraction data, and by Wentworth and others (1987) primarily from seismic-reflection data. Seismic velocities in the sedimentary section range from 1.6 to 4.1 km/s where these velocities can be clearly ascribed to sedimentary rocks, such as near well No. 1 (fig. 8.4A). In the eastern Diablo Range, velocities as high as 4.7 km/s may also be due to sedimentary rocks (fig. 8.4A). East of the synclinal axis in the Great Valley, reflections within the sedimentary section are subparallel to the top of basement, which is marked by the disappearance of reflections (f, fig. 8.4A). West of the synclinal axis, these reflections (d, fig. 8.4A) diverge slightly from the inferred top of basement (e, fig. 8.4A).
Beneath the sedimentary rocks of the Great Valley are several layers of increasing seismic velocity: a 5.5- to 5.7-km/s layer, 1.5 to 2 km thick (layer 1), a 6.0- to 6.3-km/s layer, 2.5 to 6 km thick (layer 2); a 6.6- to 6.75-km/s layer, 4 to 7 km thick (layer 3); and a 6.9- to 7.2- km/s layer, about 7 km thick (layer 4) (fig. 8.4A). In addition, there is a thin, laterally discontinuous 7.0-km/s layer embedded in the top of layer 3. Well No. 1 indicates that layer 1 is granitic rocks. Farther west, however, this layer may be interpretable either as granitic rocks or as Franciscan assemblage, which have similar velocities at this depth (fig. 8.5). In the original data of Colburn and Mooney (1986) and Holbrook and Mooney (1987), there is no perceptible reflection from an interface between layers 1 and 2 (as there is, for example, between layer 1 and the overlying sedimentary rocks), and so these two layers may, in fact, grade into one another. Layer 2 could then also be
granitic rocks, and layers 1 and 2 together would constitute a velocity-depth section similar, for example, to upper crust of the batholithic Salinian block (figs. 8.4A, 8.5). Layers 3 and 4 (6.6-7.2 km/s), which are analogous to the lower-crustal layer in the Diablo Range (6.7-7.1 km/s), may represent the middle and lower crust of accreted island arc(s) and (or) oceanic crust. The Moho is well documented at about 27-km depth. Deep reflection data beneath the Great Valley (Wentworth and others, 1987) indicate a conspicuous east-dipping band of reflections (g, fig. 8.4A) and less conspicuous subhorizontal and west-dipping reflectors.
Rocks of the Sierran foothills consist of Lower to Upper Jurassic mafic to felsic volcanic and plutonic rocks and related sedimentary rocks (argillite, chert, and flysch) that were accumulated or emplaced in an island-arc setting (Clark, 1964; Schweickert and Cowan, 1975; Saleeby, 1982; Schweickert and Bogen, 1983). The basement and metamorphic wallrocks for the intrusive rocks are tectonically disrupted and polymetamorphosed Paleozoic ophiolitic rocks (approx 300 Ma; Saleeby, 1982).
The island arc(s) in which the Jurassic rocks of the Sierran foothills were formed collapsed against the margin of the North American Continent during the Late Jurassic Nevadan orogeny (Jones and others, 1976). How this collapse occurred is problematic. Steeply east-dipping faults and upright antiforms are seen in the Sierran foothills, but a study by Moores and Day (1984) of surface relations 300 km north of transect C2 indicates obduction of the arc(s) on west-dipping thrust faults. These rocks were intruded during the Early Cretaceous by mafic to intermediate plutons belonging to the western phase of Sierra Nevada plutonism (Evernden and Kistler, 1970).
The deep structure of the Sierran foothills is known from the reconnaissance seismic-refraction experiment of Spieth and others (1981), the reflection profiling of Zoback and Wentworth (1986), and the compilation of reflection, refraction, and potential-field results by Wentworth and others (1987). The refraction data can be modeled with a 6.2-km/s basement from near the surface to about 30-km depth, a 6.6-km/s lower crust, and a Moho at 39-km depth. Other models are possible, however, and the Moho may be as shallow as 30 km (Spieth and others, 1981). We have projected the seismic-reflection results and gravity/magnetic boundary of Wentworth and others (1987) from 45 to 60 km southward onto transect C2. Two conspicuous west-dipping sets of reflections are visible, as well as a few subhorizontal reflectors. The gravity/magnetic boundary, however, has a moderate eastward dip.
Our projection of the results of Wentworth and others (1987) is uncertain not only because of the distances involved but also because their profile terminates on the east in an area that is anomalous both geologically and geophysically. In this area, batholithic rocks (trondhjemite) engulf most accreted rocks of the Sierran foothills (Jennings, 1977) and are associated with a gravity low (Oliver and others, 1980). Our projection, however, may be defensible as follows. (1) The batholithic rocks responsible for the gravity low probably do not extend below 10-km depth (R.C. Jachens, oral commun., 1988); most of the reflectors that we have projected are largely below that depth. (2) The modeled gravity/magnetic boundary is approximately similar in shape throughout the length of the Great Valley (Andrew Griscom, oral commun., 1988); in our projection, we have attempted to correct for the difference in azimuth between transect C2 and the profile of Wentworth and others (1987) by
assuming a strike parallel to the Great Valley.
Given the geologic and seismic constraints discussed above, we have interpreted the cross section through the Great Valley and Sierran foothills (fig. 8.4B), using some of the ideas of Wentworth and others (1984, 1987) for the configuration of an inferred tectonic wedge of Franciscan rocks, and some of the ideas of Saleeby (1986) for structure within crystalline rocks. The uppermost part of our cross section (to approx 2-km depth) on the east flank of the Diablo Range (fig. 8.4A) was supplied by R. C. Evarts (written commun., 1989). Below this area, we have added a hypothetical west-dipping thrust fault to bring the Great Valley sequence beneath the easternmost block of the Coast Range ophiolite and to grossly satisfy the velocity constraints of Dean Whitman and others (un pub. data, 1985; fig. 8.4A). East of the Coast Range ophiolite, we postulate thrust faults that largely follow bedding planes in the upturned section of the Great Valley sequence, similar to
those postulated by Wentworth and others (1984) for the northern Great Valley. These "backthrust" faults are required for emplacement of the wedge and help explain the thickening of the Great Valley sequence in the western limb of the syncline (see section below entitled "Discussion - Tectonic Wedging"). From the easternmost backthrust fault in the Great Valley to the San Andreas fault, we have modeled the discontinuity between variably reflective rocks of lower velocity (Franciscan assemblage, Coast Range ophiolite, and Great Valley sequence; 1.7-5.8 km/s) and poorly reflective rocks of higher velocity (mafic rocks of the Diablo Range and crystalline basement of the Great Valley; 5.5-6.8 km/s) as the floor thrust fault of the wedge. Wentworth (1987) presented a similar interpretation.
The details of composition and structure in the crystalline rocks beneath the Great Valley and Sierran foothills are speculative. Saleeby (1986) interpreted these rocks to consist fundamentally of slabs or nappes of island-arc and oceanic rocks obducted along west-dipping Nevadan thrust faults intruded by chiefly Early Cretaceous Sierran granitic plutons. We have adopted this basic scheme and added some details, interpreting layers 1 and 2 in the basement beneath the Great Valley (5.5-6.3 km/s; see above) as post-Nevadan felsic plutonic rocks, although, as noted above, the western part of layer 1 (5.5 km/s) may be Franciscan assemblage. We interpret the east-dipping gravity/magnetic boundary of Wentworth and others (1987) as the average top of mafic crust (pre-Nevadan gabbro, diabase, or basalt) in the inferred obducted sequence. Alternatively, this boundary may be the average top of mafic, magnetic intrusions in the crust (post-Nevadan gabbro) or the average base of felsic,
nonmagnetic intrusions (post-Nevadan granitic rocks). At the location where this boundary was actually modeled, it may be the average base of a large trondhjemite intrusion. We associate the east-dipping reflections beneath the central Great Valley (g, fig. 8.4A) with the thin, discontinuous 7.0-km/s layer of Holbrook and Mooney (1987), although the depth correspondence is imperfect, and we interpret this feature as a gabbroic dike. Alternatively, these east-dipping reflections may represent an east-dipping fault zone. Following Saleeby (1986), we correlate the upper and lower west-dipping bands of reflections in the eastern Great Valley and Sierran foothills (h, j, fig. 8AA) with the Bear Mountain and Melones fault zones, which may represent Cenozoic reactivations of inferred west-dipping Nevadan thrust faults.
Wentworth and others (1984) interpreted the juxtaposition of Franciscan assemblage and a coeval section consisting of Coast Range ophiolite and Great Valley sequence as having occurred during landward movement of the Franciscan assemblage as a tectonic wedge. They reinterpreted the "Coast Range thrust fault" of Bailey and others (1970), a subduction megathrust between the Coast Range ophiolite and the Franciscan assemblage, as the roof thrust of the wedge. More recently, the thrust nature of the "Coast Range thrust fault" has been reevaluated. Jayko and others (1987), testing an hypothesis by Platt (1986), produced abundant evidence that the contact between Franciscan assemblage and Coast Range ophiolite is a detachment surface along which the upper plate was extended during uplift of the Franciscan assemblage. Their evidence is the consistent attenuation, as opposed to repetition, of geologic section across this discontinuity and associated faults above it.
They proposed the term "Coast Range fault" for this discontinuity, which we adopt here. Evidence of attenuation is present even on transect C2, in that the two outcrops of the Coast Range ophiolite in the eastern Diablo Range (fig. 8.4A) represent an abridged section of ophiolite: The western outcrop is partially serpentinized ultramafic rock of the basal part of an ophiolite, whereas the eastern outcrop is the sill complex and volcanic flows of the upper part of an ophiolite. These two parts of the ophiolite are now juxtaposed across the crooked, steeply dipping Tesla-Ortigalita fault. Although this fault now offsets the Coast Range fault, it may represent reactivation of a normal fault that originally soled into the Coast Range fault (compare Raymond, 1973).
PAST AND PRESENT TECTONIC REGIMES
The extensional nature of the Coast Range fault poses several problems for emplacement of the Franciscan assemblage as a tectonic wedge. Where is the roof thrust fault of the wedge? How did the Franciscan assemblage reach its current position with an extended overlying section of the Coast Range ophiolite and Great Valley sequence? Was the Franciscan assemblage uplifted from beneath the western Great Valley? The apparent continuity between the Great Valley basement and the 6.7- to 7.l-km/s layer in the Diablo Range indicates a negative answer to the last question.
These problems can be solved if the extensional event was separated in time and space from the compressional event, or tectonic wedging. Jayko and others (1987) reviewed the published evidence regarding the geologic history of extensional faulting. In one place, the Coast Range fault and associated faults are overlapped by sedimentary rocks of Oligocene and younger age, and in another place by sedimentary rocks of Paleocene and younger age. The occurrence of detritus derived from the Franciscan assemblage in Paleocene and Eocene strata of the Coast Ranges (Dickinson, 1966; Berkland, 1973) indicates that the lower plate was exposed by the early Tertiary. Jayko and others (1987) inferred that uplift of the Franciscan assemblage and associated extensional faulting in the upper plate occurred during the Late Cretaceous and ( or) early Tertiary.
The history of compressional tectonics in the Coast Ranges is sparse and varies from place to place. In the northern Coast Ranges, thrust faulting and folding began during the early Tertiary (Blake and others, 1987; M. C. Blake, Jr., oral commun., 1989), and compressional deformation is continuing today in rocks of the northern Great Valley (Harwood and Helley, 1987). In the southern Coast Ranges, at least four Cenozoic deformations or uplifts, indicated by unconformities or eastward-migrating depocenters, have ages of late Paleocene, late Eocene to early Miocene, late Miocene, and late Pliocene (Namson and Davis, 1988; Namson and others, 1990; Rentschler and Bloch, 1988). Modern thrust faulting and folding still is occurring, as indicated by the 1983 Coalinga earthquake (see chap. 5; Eaton, 1990).
Landward movement of the Franciscan assemblage as a wedge may have even begun in the Mesozoic. In the northern Coast Ranges, several northwest-striking faults (Paskenta, Elder Creek, and Cold Fork faults) offset rocks structurally above the Franciscan assemblage (but not the Franciscan assemblage itself) and represent major discontinuities in the depositional environment of the Great Valley sequence (Jones and Irwin, 1971). These faults, which have displacements of tens of kilometers to as much as 100 km, are interpreted to have moved primarily during the Cretaceous (Jones and Irwin, 1971), although the latest limit on the time of movement is about 3.4 Ma (Hardwood and Helley, 1987; M.C. Blake, Jr., oral commun., 1989). Wentworth and others (1984) and Jayko and others (1987) interpreted these faults as tear faults in the plate structurally above a wedge of Franciscan assemblage.
In light of the above data and interpretations, we postulate (1) that uplift of the Franciscan assemblage and extension of the upper plate, consisting of Coast Range ophiolite and Great Valley sequence, occurred during the Cretaceous (or, at the latest, during the early Tertiary, if Cretaceous movement on the Paskenta-Cold Fork fault system is not linked to landward wedge transport) well west of the present Diablo Range; and (2) that a tectonic wedge of Franciscan assemblage was subsequently driven landward, with the extended upper plate riding passively atop it. This wedge is interpreted to have moved along a floor thrust fault aligned with the contact between the Great Valley sequence and its crystalline basement. To the west of the present Diablo Range, where movement initiated, the basement was an outboard part of the Coast Range ophiolite. Beneath the Great Valley, where the movement is presently occurring, the basement is similar to the Coast Range ophiolite but contains
numerous younger plutons. A roof thrust fault apparently developed only near the east tip of the wedge (fig. 8.4B); presumably, erosion kept pace with uplift near the tip. Differential vertical or horizontal movements of the wedge may have produced tear faults, such as the Paskenta, Elder Creek, and Cold Fork faults, and may have reactivated extensional faults to produce complex faults, such as the Tesla-Ortigalita fault.
The Mendocino triple junction has moved northward through offshore central California during approximately the past 20 Ma, and subduction of the Farallon plate (or its derivative) was replaced by transform motion of the Pacific plate past North America (see chap. 3; Atwater, 1970, 1989). If tectonic wedging occurred during the late Mesozoic and Cenozoic, in association with all of the episodes of tear faulting or compression outlined above, then clearly it was driven during both subduction and transform regimes. At present, it is being driven by a transform regime. At least two additional arguments can be made that wedge motion-indeed, probably a major fraction of wedge motion-occurred during the subduction regime. The first argument is simply based on geometry: The east boundary of the Coast Ranges, inferred to coincide approximately with the buried tip of the wedge, largely parallels Mesozoic structures in the Sierran foothills and the Great Valley rather than the late Cenozoic
San Andreas fault (Wentworth and Zoback, 1989; C.M. Wentworth, oral commun., 1990). The second argument, developed below, is based on the total apparent displacement of the wedge.
If the inferred tectonic wedge of Franciscan assemblage extends to the San Andreas fault, as we have shown (fig. 8.4B), then a minimum shortening of about 70 km has occurred along faults at the top and bottom of the wedge in the Diablo Range. Likewise, in the northern Coast Ranges, the inferred tear faults in the plate above the wedge have a total displacement-and, thus, shortening-of many tens of kilometers (Wentworth and others, 1984), possibly as much as 100 km (Jones and Irwin, 1971).
Although a transform regime has replaced a subduction regime in central California over approximately the past 20 Ma, plate-margin compression, necessary to drive the wedge, has persisted for only approximately the past 5 Ma (Page and Engebretson, 1984). At about 5.5-4.5 Ma, transform motion was also transferred from offshore faults to the modern San Andreas fault system (see chap. 3; Atwater, 1989; Humphreys and Weldon, in press). Present plate-margin compression is understandable from (1) the slight misalignment of the direction of relative plate motion (N. 35° W.; Minster and Jordan, 1978) and the strike of the San Andreas fault (N. 40° W.), and (2) the opening of the Basin and Range province. Crouch and others (1984) calculated from these two effects a rate of shortening across the Coast Ranges that, integrated over the past 5.5 Ma, predicted a total shortening of 28 to 72 km. Most of this shortening could be accounted for in small fault displacements and folds
distributed throughout the Coast Ranges (Crouch and others, 1984). Thus, the minimum shortening of 70 to 100 km represented by the tectonic wedge, as discussed above, would appear to equal or exceed the maximum shortening calculated for the transform regime, a result suggesting that some, if not most, of the wedge motion occurred during the subduction regime.
Shear coupling between the subducting plate and overlying accretionary prism (Franciscan assemblage) could conceivably drive the wedge during the subduction regime. Such a mechanism has been postulated for southern Alaska by Fuis and Platker (in press). To drive the wedge during a transform regime appears to require a less obvious mechanism, such as plate-margin compression combined with differing deformation in the upper and lower crust. Such a mechanism is developed below.
Sibson (1982) pointed out, on the basis of strength considerations, that ductile flow could be expected in the middle crust, below the maximum depth of earthquake hypocenters. Several workers (Crouch and others, 1984; Namson and Davis, 1988; Eaton and Rymer, in press) have postulated a decollement near the base of the seismicity in the Coast Ranges (avg 15-km depth; see chap. 5; Wesson and others, 1973) into which thrust and oblique-slip faults on both sides of the Coast Ranges sole. They envision differential movement between upper and lower crust caused by differing alignment of the transform faults in these two layers, or by shortening of the lower crust by ductile thickening.
We have incorporated the idea of a Coast Range-wide detachment in our cross section (fig. 8.4B). In the Diablo Range, we show a young thrust fault at the base of the inferred tectonic wedge soling into the brittle-ductile transition zone, which in this area is, coincidentally, near the interface between Franciscan rocks and mafic crust. Although we also indicate soling of the San Gregorio-Hosgri fault into such a zone and underthrusting of the Salinian block by the early Tertiary accretionary prism, focal mechanisms in this region indicate pure strike slip on the San Gregorio- Hosgri fault (see chap. 5) and argue against this interpretation. Such an interpretation of a Coast Ranges-wide midcrustal detachment requires that the deformational style and (or) location of the San Andreas fault system change from the upper to the lower crust.
If we have correctly inferred the geologic history of wedge movement, it is remarkable that such movement has apparently occurred in two quite different tectonic regimes, a subduction regime and a transform regime.
The crustal structure of southern California is complicated by the Big Bend in the San Andreas fault, situated between the Coast Ranges and Transverse Ranges, and by onshore spreading centers of the East Pacific Rise, situated in the Salton Trough (figs. 8.2, 8.3). The Big Bend is thought to result from westward movement of the Sierra Nevada relative to the Mojave Desert, along the Garlock fault (Hill and Dibblee, 1953). The San Andreas fault crosses the Transverse Ranges, between the Big Bend and Salton Trough, at an angle oblique to relative plate motion, while somehow remaining a largely vertical, strike-slip fault.
The onshore spreading centers in the Salton Trough are situated at echelon offsets between the San Andreas, Imperial, and Cerro Prieto faults (see fig. 3.8; Lomnitz and others, 1970). These three faults are interpreted as transform faults; the San Andreas links the northernmost spreading center in the Salton Trough with the Mendocino triple junction. A progressive decrease in spreading rate northward along the East Pacific Rise is inferred to give rise to movement on the San Jacinto, Elsinore, San Miguel/Newport- Inglewood, and other faults in southern California and Mexico (Lomnitz and others, 1970; Elders and others, 1972).
First, we discuss a transect across southern California. Centennial Continental-Ocean Transect C3 (Howell and others, 1985). Second, because of the three-dimensionality of the geology and tectonics in southern California, we include a discussion of block motions, largely from Weldon and Humphreys (1986).
We modify and reinterpret the section of Centennial Continent-Ocean Transect C3 (Howell and others, 1985) that extends from Santa Catalina Island to the Colorado Desert (fig. 8.2). This section of the transect crosses four blocks or provinces, the California Continental Borderland (hereafter referred to simply as the "borderland"), Peninsular Ranges, Salton Trough, and Chocolate Mountains (fig. 8.6). The transect crosses the Newport-Inglewood, Elsinore, San Jacinto, and Imperial strike-slip faults. Constraints for the transect include surface geology, isotopic studies, seismic-refraction profiling (which is sparse, except in the Salton Trough), tomographic studies, and potential-field studies.
- Crustal structure of southern California. A
, Surface geology, isotope data, and models of seismic-refraction, gravity, and magnetic data for part of Centennial Continent-Ocean Transect C3 (see Howell and others, 1985). B
, Reinterpretation of Transect C3. Major features in figure 8.6B
include, from west to east, (1) Franciscan assemblage overlying mafic crust in the borderland; (2) Peninsular Ranges batholithic block, consisting of west half inferred to be underlain at depth by mafic (island arc or oceanic) crust and east half inferred to be underlain at depth by intermediate continental Precambrian(?) rocks;
8.4 and 8.6
(3) late Cenozoic rift, the Salton Trough, whose central part is inferred to be underlain by entirely new crust that includes, from top to bottom, sedimentary rocks, thermally metamorphosed sedimentary rocks, and gabbro generated at onshore spreading center; and (4) Pelona-Orocopia schist of Haxel and Dillon (1978) (similar to the Franciscan assemblage), interpreted to compose tectonic wedge. Tectonic wedge in feature 4 is postulated to have been obducted onto continental crust (see text); its tip would lie well east of east end of cross section. This reinterpretation differs from Howell and others' (1985) primarily in interpreting mafic crust at shallower depths beneath the borderland and western Peninsular Ranges (5-8 km versus 11-15 km) to better match seismic and potential-field results. See figures 8.2 and 8.3 for location of Transect C3; see figure 8.4 for explanation. No vertical exaggeration.
The borderland is broken up by right-slip faults into several northwest-trending blocks. Our cross section (fig. 8.6) begins on the easternmost block, the "Catalina terrane" (Howell and others, 1985), bounded on the east by the Newport- Inglewood fault. The Catalina terrane is underlain, beneath patches of Tertiary volcanic rocks, by Franciscan assemblage, on the basis of outcrops on Santa Catalina Island (Platt, 1975, 1976; Jones and others, 1976) and submarine dredge and core samples (Vedder and others, 1974). The block west of the Catalina terrane, the "San Nicholas terrane" (Howell and others, 1985), is inferred to be underlain, beneath Cenozoic marine sedimentary rocks, by rocks similar to the Great Valley sequence and Coast Range ophiolite of central California, possibly in fault contact with Franciscan assemblage at depth (Vedder and others, 1974).
A reversed seismic-refraction profile just west of Santa Catalina Island indicates P-wave velocities of 5.8 km/s to 6-km depth and of 6.7 km/s to the Moho at about 24-km depth (fig. 8.6A, Shor and Raitt, 1958). This velocity-depth section is similar to that for the Diablo Range of central California (see above), where Franciscan rocks are equated with the 5.8-km/s interval, and middle and lower crust of island arc(s) and (or) oceanic crust are equated with the 6.7-km/s interval. In this region, there is no clear evidence of landward movement of the Franciscan assemblage as a tectonic wedge, although such evidence may surface during future investigations. As in the Diablo Range, the lower crust must have been brought to its present 18-km thickness by (1) imbrication of slices of island-arc crust, (2) tectonic underplating of several thicknesses of oceanic crust, and (or) (3) magmatic underplating. Subduction continued beneath the borderland until sometime between 30
and 20 Ma (see Atwater, 1970), depending on the latitude to which the borderland is palinspastically restored.
The Peninsular Ranges are underlain in the west by supracrustal rocks, including, from top to bottom, Cenozoic marine sedimentary rocks, Cretaceous forearc sedimentary rocks, Lower(?) Cretaceous and Upper Jurassic andesite (Santiago Peak Volcanics), and Middle Jurassic flysch (Bedford Canyon Formation) that was disrupted and overturned before the Late Jurassic (Larsen, 1948; Jennings, 1977; Criscione and others, 1978). These rocks are intruded by Early Cretaceous plutons of the Peninsular Ranges batholith that include chiefly tonalite and gabbro and show no special age trends (static magmatic arc; Silver and others, 1979). About 80 km east of the coastline, both prebatholithic and batholithic rocks change (fig. 8.6A): To the east, the prebatholithic rocks are dominantly metamorphosed clastic rocks of amphibolite grade, and the batholithic rocks are chiefly tonalite and granodiorite whose ages decrease progressively eastward (from 105 to 80-90 Ma; migrating magmatic arc;
Silver and others, 1979). Major-element chemistry and oxygen isotopes indicate that deep crustal rocks in the west half of the batholith are dominantly primitive and tholeiitic but, in the east, more aluminous and oxidized (fig. 8.6A). Older crust that was once at the Earth's surface is inferred at depth in the east (Silver and others, 1979).
Seismic constraints for the deep structure of the Peninsular Ranges are sparse. Using blasts at the Corona Quarry in the northernmost Peninsular Ranges, Gutenberg (1951) and Shor (1954) obtained an unreversed refraction profile, extending southward to the United States- Mexican border, along with a reflection record at the blast site. Interpretation of these data by Shor and Raitt (1958) indicated velocities of 5.9 km/s to 8-km depth, 6.8 km/s to 26-km depth (with a possible low-velocity zone in this interval), and 7.0 km/s to the Moho at 30- to 32-km depth (fig. 8.6A). In contrast, a study by Nava and Brune (1982) using a blast at the same quarry, reversed by an earthquake in Baja, Mexico, indicated a Moho depth of 42 km. Hearn and Clayton (1986a, b) used as many as 600,000 arrivals from local earthquakes in southern California to map the velocity of the crust and upper mantle, using tomography. Their map indicates that the west half of the Peninsular Ranges has a higher
average upper-crustal velocity and a lower average mantle velocity in comparison with the east half. Their map of Pn delays for the Peninsular Ranges suggests no crustal root and an average crustal thickness of nearly 30 km. Gravity modeling of the Peninsular Ranges (Fuis and others, 1984) and isostatic calculations also indicate a maximum crustal thickness of 30 to 33 km. In our cross section (fig. 8.6B), we adopt a maximum crustal thickness of 33 km.
An additional constraint on crustal structure is the modeling by Jachens and others (1986; R.C. Jachens, written commun., 1988) of strong magnetic and gravity steps (500 nT and 40 mGal, respectively) in the central Peninsular Ranges: A moderately east dipping boundary is modeled between more magnetic, dense rocks on the west and less magnetic, lighter rocks on the east. This boundary is poorly defined at the latitude of our transect; it correlates approximately (within 15 km or so) with the boundary between the east and west halves of the Peninsular Ranges batholith, as discussed above (fig. 8.6A). In the cross section (fig. 8.6B), we interpret an eastward deepening of mafic rocks, including prebatholithic and (or) batholithic mafic rocks (gabbro, diabase, and metamorphic rocks), along this magnetic/gravity boundary. R.C. Jachens (oral commun., 1989) indicated that, in some places, this boundary is so planar as to be interpretable as a fault. As beneath the
borderland, the mafic rocks beneath the Peninsular Ranges may have reached their current thickness by thrust imbrication, tectonic underplating, or magmatic underplating. We speculatively show some tectonic underplating on the west side.
The Salton Trough is the landward extension of a ridge/transform-fault system, the East Pacific Rise, of the Gulf of California (see fig. 3.13). This system became well established during the late Cenozoic (approx 5 Ma) as the plate boundary jumped inland from offshore Baja California (Atwater, 1970, 1989; Humphreys and Weldon, in press).
The Salton Trough is underlain by upper Cenozoic sedimentary rocks and minor amounts of volcanic rocks, which are exposed chiefly around its edge and are penetrated in wells. Onset of rifting and major subsidence in the Salton Trough was followed by marine incursion during the latest Miocene to late(?) Pliocene, as indicated by the Imperial Formation (Dibblee, 1954; Powell, 1984). The thick Cenozoic sedimentary section is offset by Quaternary faults, both exposed and buried, and is intruded by Quaternary volcanic rocks, both silicic rocks that form volcanoes at the two inferred onshore spreading centers (fig. 8.7) and mafic rocks that are penetrated in geothermal wells (Elders and others, 1972; Robinson and others, 1976). Faulting in the Salton Trough occurs primarily on conjugate northwest- and northeast-striking faults and is largely strike slip (Johnson and Hadley, 1976; Johnson, 1979; Fuis and others, 1982). North-south-striking faults, however, such as the north end of the
Imperial fault, the Brawley fault, and north-south-striking seismicity lineaments (that outline inferred spreading centers; figs. 8.1, 8.7), have normal components and lead to the subsidence that ultimately created the Salton Trough. Earthquake hypocentral depths indicate that brittle fault motion extends to about 12-km depth in the Imperial Valley but deeper in the adjacent Peninsular Ranges along the San Jacinto fault (Doser and Kanamori, 1986).
Figure 8.7 - Tectonic block motion in southern California (modified from Weldon and Humphreys, 1986, and Humphreys and Weldon, in press). Various blocks (italicized names near motion vectors) move through region where the San Andreas fault trends obliquely to plate motion, between the Big Bend and the Salton Trough, without major convergence with each other. Through this region they move counterclockwise, following nearly concentric arcs (arcs and radii, thin red lines). New crust, which is forming in wake of the Salton and Perris blocks in the Salton Trough, is created by sedimentary-basin fill and gabbroic intrusions at onshore spreading centers, outlined by seismicity lineaments. High-velocity mantle beneath the Transverse Ranges is interpreted as cold, sinking lithospheric mantle, and low-velocity mantle beneath the Salton Trough as hot upwelling asthenosphere or lithospheric mantle containing partial melt (Humphreys and others, 1984; Humphreys and Clayton, in press). Motion vectors for the
Mojave Desert and Sierra Nevada modified to incorporate results of Sauber and others (1986).
Detachment faulting on the east flank of the Salton Trough, in the Chocolate Mountains and other ranges, preceded the Pliocene and later basin-forming tectonics in the Salton Trough (Dillon, 1975; Berg and others, 1982; Frost and others, 1982). Similar faulting on the west flank of the Salton Trough, however, may have both preceded and overlapped in time the tectonics in the Salton Trough (Wallace and English, 1982; Schultejahn, 1984; Isaac and others, 1986).
Biehler and others (1964) and Fuis and others (1982, 1984) demonstrated from seismic surveys that the sedimentary rocks (1.8-5.5 km/s) in the central Salton Trough are as much as 5 km thick (fig. 8.6A). Below 5-km depth, a low-velocity (5.6 km/s) "basement," which is not separated from the overlying sedimentary rocks by a velocity discontinuity, is inferred to be metamorphosed (greenschist facies) sedimentary rocks (Fuis and others, 1982, 1984); this "basement" layer extends to 12-km depth. High heat flow in the Salton Trough (see Lachenbruch and others, 1985) is inferred to cause the metamorphism of the sedimentary rocks. Thus, the entire section of inferred upper Cenozoic sedimentary rocks, metamorphosed and unmetamorphosed, is as much as 12 km thick.
Below 12- to 14-km depth in the Salton Trough, a high-velocity (7.1-7.2 km/s) "subbasement" that is indicated by seismic-refraction data (fig. 8.6A) is inferred to be gabbro generated at one of the nearby spreading centers (Fuis and others, 1982, 1984). Modeling of seismic-refraction and gravity data indicate that the Moho in the central Salton Trough is 23 to 28 km deep (Fuis and others, 1982, 1984). The central Salton Trough is interpreted to be underlain entirely by late Cenozoic crust (fig. 8.6B).
Buried scarps separating old crust (plutonic and metamorphic rocks; 5.9-6.0 km/s) from new crust (sedimentary and basaltic rocks; 1.8-7.2 km/s) are visible by seismic methods on both sides of the Salton Trough (Fuis and others, 1982; Fuis and Kohler, 1984). On the west side of the rift, where the new-crust/old-crust boundary is ragged in outline (fig. 8.7), we interpret normal faults (fig. 8.6B); on the east side, where this boundary is linear, we interpret a strike-slip fault. In our cross section, faults on the west side of the Salton Trough are inferred to have originated by pullaway from the Cerro Prieto spreading center to the southeast; the fault on the east side is inferred to be a largely passive suture (figs. 8.6B, 8.7; Fuis and others, 1982). A similar rift configuration is seen, for example, in the Gulf of Elat (Gulf of Aqaba, Red Sea; Ben-Avraham, 1985).
Rocks on the east flank of the Salton Trough are igneous and metamorphic rocks that compose two or more fault-bounded packages, or tectonostratigraphic terranes (see Howell and others, 1985). A complex of metasedimentary and mafic metaigneous rocks described by Dillon (1975) may include two Precambrian terranes, the Joshua Tree and San Gabriel terranes, described farther north by Powell (1981). This complex is intruded by intermediate to felsic Mesozoic plutons and rests on the low-angle Chocolate Mountains thrust fault above the (informal) Pelona-Orocopia schist of Haxel and Dillon (1978; see also Haxel, 1977). The Pelona-Orocopia schist consists chiefly of metagraywacke and lesser metapelite, metabasite, metachert, marble, and serpentinite (albite-epidote-amphibolite facies) of uncertain but probable late Mesozoic or early Tertiary age (Conrad and Davis, 1977; Miller and Morton 1977, 1980). It resembled the Franciscan assemblage but lacks melange.
TECTONICS- THE THREE-DIMENSIONAL PICTURE
Many workers have speculated on the depositional environment and origin of the Pelona-Orocopia schist. Haxel and Dillon (1978) postulated formation in an ensimatic rift basin with continent on both sides - not unlike the current Salton Trough. Powell (1981) favored an origin as a parautochthonous continental-marginal deposit. In any case, from its quartz content, the Pelona-Orocopia schist clearly originated near a continent and incorporated continental detritus. It was thrust beneath the continental metasedimentary-metaigneous complex some time after Mesozoic plutonism (80 Ma; Powell, 1981) and before Oligocene volcanism (35 Ma; Crowe 1978; Crowe and others, 1979). The thrust fault may have been reactivated one or more times as a low-angle normal, or detachment, fault (Frost and others, 1982).
Evidence from refraction profiling in the western Mojave Desert across the Rand schist, which has been correlated with the Pelona-Orocopia schist (Ehlig, 1968), indicates relatively low-velocity crust beneath this body (max 6.4 km/s; Fuis and others, 1986) that we infer to be continental crust. We speculate that the Pelona-Orocopia schist also rests on continental crust and that the Rand and Pelona-Orocopia schists were emplaced as a tectonic wedge into continental crust in a manner similar to the Franciscan assemblage of central and northern California. We hypothesize that the metasedimentary-metaigneous complex structurally above the schist is analogous to either (1) rocks of the Coast Range ophiolite/Great Valley sequence which rode passively atop the wedge in central and northern California after being extended during uplift of the Franciscan assemblage, or (2) rocks of the Great Valley sequence which were peeled up along backthrust faults during landward movement of the wedge.
In southern California, tectonic wedging clearly occurred before the present transform regime, presumably during subduction of the Farallon plate (or its derivative). The geologic data discussed above indicate that the Salton Trough has undergone extension, rather than compression, for approximately the past 5 Ma (probably even longer; see Humphreys and Weldon, in press).
Crustal thickness is unknown in the Chocolate Mountains; however, the Colorado Desert, to the east and north, has a generally thin (26-28 km) crust (fig. 8.3) and a local root (32 km deep) under the Whipple Mountains metamorphic-core complex (Fuis, 1981; Jill McCarthy, written commun., 1988).
The geology and, presumably, the deep structure of southern California illustrated along transect C3 (fig. 8.6) is grossly two dimensional as far north as the Transverse Ranges. In the Transverse Ranges, the rocks on the southwest side of the San Andreas fault are similar to those in the Chocolate Mountains. These rocks are bounded on the south and west by older, deformed strands of the San Andreas fault system (fig. 8.7; Powell, 1981). The tectonics also changes in the Transverse Ranges: Crustal-block motion swings to the west to follow the trend of the San Andreas fault, as discussed below.
Using Quaternary geologic and geodetic evidence, Weldon and Humphreys (1986) documented complex motion of crustal blocks in southern California that is not simply predictable from the motion vectors of the Pacific and North American plates. These motion vectors predict a large component of convergence across the San Andreas fault in the Transverse Ranges between the Big Bend and the Salton Trough (fig. 8.7). For a total offset on the San Andreas fault system of about 300 km (Hill and Dibblee, 1953; Crowell, 1962, 1981; Powell, 1981), a maximum of 45 km of uplift in the Transverse Ranges would be expected (Weldon and Humphreys, 1986). However, the preservation in the Transverse Ranges of upper Cenozoic sedimentary rocks and of offset bedrock features on either side of the San Andreas fault argues against such major uplift and associated consumption of crust, as does the relatively minor crustal root in the Transverse Ranges (fig. 8.3). Weldon and Humphreys (1986) constructed a
kinematic model in which crustal blocks between the San Andreas fault and a system of borderland and other offshore faults rotate counterclockwise, parallel to the San Andreas fault, between the Salton Trough and the Big Bend (fig. 8.7). Approximately two-thirds of the relative northwestward motion of the Pacific plate past the North American plate is taken up by the San Andreas fault system, including the San Jacinto fault; approximately one-third of it is taken up by the Elsinore fault, a system of borderland faults, and offshore faults in central California, including the San Gregorio-Hosgri fault (fig. 8.7); and only a minor fraction of it is taken up within the blocks (see Humphreys and Weldon, in press).
A marked advance in the P-wave traveltimes of teleseismic arrivals in southern California is associated with the Transverse Ranges and extends across the San Andreas fault (Hadley and Kanamori, 1977; Raikes, 1980). Tomographic analysis of this anomaly indicates that it results from a vertical slablike region of relatively high velocity in the mantle which extends downward as far as 250 km (Humphreys and others, 1984; Humphreys, 1985; Humphreys and Clayton, in press). The amount of velocity increase, a maximum of 3 percent, is most reasonably explained by a thermal difference in the mantle. This velocity increase, coupled with a velocity decrease in the upper 90 km or so of mantle beneath the Salton Trough, led Humphreys and Hager (1984 and in press) to infer small-scale mantle convection between the Salton Trough and the Transverse Ranges. This convection involves passive rising of asthenosphere beneath the Salton Trough and cooling and sinking of lithosphere beneath the
Transverse Ranges. The vertical extent of the inferred lithospheric slab beneath the Transverse Ranges, 250 km, is similar to the 300-km estimate of total offset along the San Andreas fault system. However, because the cooled mantle slab extends across the San Andreas fault, most of the mantle seems to be moving independently of the crust (fig. 8.8; Hadley and Kanamori, 1977; Humphreys and others, 1984; Humphreys, 1985; Humphreys and Hager, in press). The horizon of decoupling is apparently at or below the Moho because crustal material is not entrained in the slablike feature. Additional decoupling may be occurring in the crust, similar to that postulated for central California (Yeats, 1981; Webb and Kanamori, 1985). Decoupling at the Moho requires that the deformational style and (or) location of the San Andreas fault system change from the crust to the mantle (fig 8.8). We note that mantle drag on the crust is required to maintain the Big Bend in the San Andreas fault because plate-edge forces alone
would tend to "short-circuit" the San Andreas fault south of the Big Bend and cause most plate motion to be taken up on the San Jacinto, Elsinore, or more westerly faults (Kosloff, 1978; Humphreys, 1985).
Figure 8.8 - Motion of crustal blocks in southern California (open arrows; see fig. 8.7) and somewhat different motion of lithospheric mantle below (solid arrows) (modified from Humphreys, 1985, and Humphreys and Hager, in press). Mantle convection cell is envisioned between the Salton Trough and the Transverse Ranges. Crust and lithospheric mantle appear to be moving independently of one another, as the San Andreas fault trends obliquely across region of inferred, sinking lithospheric mantle beneath the Transverse Ranges (see fig. 8.7). Small arrows, relative fault motion; sawteeth, upper plates of crustal thrust faults; crosslines, subduction zones in lithospheric mantle.
To summarize, block motions in the region between the Big Bend and the Salton Trough result in only minor interblock convergence in the crust. In contrast, major convergence in the lithospheric mantle is indicated by the presence of an inferred, sinking lithospheric slab.
STRUCTURE OF THE UPPER MANTLE
In addition to the Transverse Ranges and Salton Trough, other regions in California show mantle velocity anomalies that imply structure within the lithospheric mantle and even the asthenosphere. The seismic networks in California (see chap. 5) provide an abundant source of regional earthquake and teleseismic arrivals that have been used to determine this upper-mantle structure.
A detailed study of the compressional-wave velocity of the uppermost mantle in central California reveals a normal velocity of about 8.0 km/s and no evidence for velocity anisotropy (Oppenheimer and Eaton, 1984). A similar study in southern California finds nearly the same average velocity, 7.95 km/s, with evidence for 2-percent velocity anisotropy (Vetter and Minster, 1981; Hearn, 1984). The fast direction is N. 75° W., approximately parallel to the San Andreas fault in southern California. Seismic-velocity anisotropy in the upper mantle has been reported elsewhere, notably in oceanic crust, and is commonly attributed to alignment of olivine in the mantle along a shear-stress direction (Bamford and others, 1979). In southern California, this shear would presumably be that associated with the motion of crustal blocks above the lithospheric mantle.
Lithospheric thickness along the San Andreas fault has been investigated by using delay times of teleseismic arrivals and thermal models (Zandt and Furlong, 1982). These studies indicate a lithospheric thickness of only 30 to 60 km for much of western California, and as little as 20 km for northern California just south of Cape Mendocino. These lithospheric thicknesses contrast with averages of 60 to 80 km for the Western United States and 120 to 170 km for the Central and Eastern United States (Iyer and Hitchcock, 1989). The thinness of the lithosphere in northern California south of Cape Mendocino is due to the creation of the San Andreas fault system itself: The transform fault is lengthening as the Mendocino triple junction migrates northward. As this junction migrates northward, the west edge of North America is sliding off the edge of the northward-moving, subducting Gorda plate, thereby creating a "window" where no subducting lithospheric slab is present (Dickinson and
Synder, 1979). In this slabless window, the North American crust is initially in direct contact with the asthenosphere that has welled upward to fill the hole left by the Gorda plate (Zandt and Furlong, 1982). This geometry produces the thinnest lithosphere in California and, probably, in North America. In contrast, the lithosphere is abnormally thick (250 km) in the Transverse Ranges, where "subduction" of lithospheric mantle is occurring, as discussed above.
Velocity anomalies appear to extend even into the asthenosphere beneath western California. Aki (1982) summarized the results of Cockerham and Ellsworth (1979) and Raikes (1980) in a combined velocity-anomaly model for a depth range of about 100-225 km in the mantle (fig. 8.9). Aki suggested that the low-velocity region in central California is hot, mobile material associated with the slabless window. Such an association appears likely for the northwest-trending prong of this anomaly, as refined by the recent work of Benz and others (1990); however, the center of the anomaly, located near Long Valley caldera (figs. 8.2, 8.9), apparently has a different origin. Low-velocity regions are also associated with the Salton Trough, where asthenospheric upwelling is inferred, and the eastern Mojave Desert, where crustal extension has occurred. The high-velocity region that crosses the San Andreas fault in southern California is similar to the one discussed above (fig. 8.7).
Figure 8.9 - Seismic-velocity anomalies in upper mantle (chiefly asthenosphere), derived from teleseismic delay-time data (Aki, 1982), for depth ranges 125-225 km (central California; Cockerham and Ellsworth, 1979) and 100-180 km (southern California; Raikes, 1980). H, high velocities (contours solid); L, low velocities (contours dashed). Contour interval, 2 percent. High seismic velocities that cross the San Andreas fault in southern California are similar in pattern to those shown in figure 8.7. Heavy lines, major faults.
The crust along the San Andreas fault system thickens from about 16 km at Cape Mendocino, in northern California, to about 30 km in southern California and thus is significantly thinner than the average thickness (36 km) for the conterminous United States. Lithospheric thickness (20-60 km) is also substantially less along most of the San Andreas fault system than is typical for continental areas (60-170 km). The lithosphere is thinnest at both ends of the fault system, at the Mendocino triple junction on the north, where the North American plate is sliding off the edge of the Gorda plate as it moves northward, and in the Salton Trough on the south, where onshore spreading centers of the East Pacific Rise are generating new crust in a rift between the North American and Pacific plates. In contrast, the lithosphere is abnormally thick (250 km) in the Tranverse Ranges, where "subduction" of lithospheric mantle is occurring.
The crust of central California was formed at an Andean-type continental margin and has been modified by large offsets along strike-slip faults of the San Andreas fault system. East of the San Andreas fault, the Andean-marginal sequence includes a subduction complex (Franciscan rocks), a forearc basin (Great Valley sequence), and a magmatic arc (plutons of the Sierra Nevada). The subduction complex appears to have been emplaced as a tectonic wedge beneath sedimentary rocks of the forearc basin. West of the fault, displaced blocks constitute an Andean-marginal sequence that has been shortened by strike-slip faulting.
The tectonic wedge of Franciscan rocks east of the fault is reinterpreted to extend from its tip beneath the Great Valley all the way to the San Andreas fault. This interpretation is motivated by the apparent continuity between crystalline basement rocks beneath the Great Valley and mafic rocks at mid crustal depths in the Diablo Range, beneath the Franciscan rocks. The presence of extended crust atop the tectonic wedge (outliers of Coast Range ophiolite and Great Valley sequence) has led us to propose the following tectonic evolution for the wedge. (1) Franciscan rocks were uplifted and upper-plate rocks (those above the subduction zone) were extended during the Cretaceous (or, possibly, early Tertiary) well west of their current position in the Coast Ranges. (2) The Franciscan rocks and overlying extended crust were subsequently forced landward during one or more episodes in the form of a wedge that largely followed the contact between Great Valley basement and the Great Valley
sequence. (3) Wedge movement began during the subduction of the Farallon plate (or its derivative) beneath central California; however, it apparently is also occurring at present, in the San Andreas transform regime. Present movement is interpreted to result from compression across the San Andreas fault system coupled with differential motion between the upper and lower crust; this differential motion is interpreted to occur on thrust fault(s) at the base of the wedge that sole into the brittle-ductile transition zone.
The crustal structure in southern California shares several features in common with central California, including, west of the fault, an Andean-marginal sequence that has been shortened or, at least, shuffled by strike-slip faulting, and, east of the fault, subduction-complex rocks that are inferred to have moved landward as a tectonic wedge into the continental rocks. However, major differences are apparent in southern California. First, east of the San Andreas fault, the Andean-marginal sequence is incomplete: A forearc basin is absent, and the magmatic arc is diffuse. Second, new continental crust has formed in the Salton Trough, an active crustal pullapart basin, by a combination of rapid sedimentation, metamorphism, and magmatic intrusion at the onshore spreading centers. In addition, the motions of the crust and lithospheric mantle differ in southern California: The crust is moving as a collage of blocks, with only minor interblock convergence, whereas the lithospheric
mantle is converging and "subducting" beneath the Transverse Ranges.
The interpretations of (1) a midcrustal detachment in the brittle-ductile transition zone in central California and (2) a crust-mantle detachment in the Transverse Ranges of southern California would appear to require that the deformational style and (or) location of the San Andreas fault system change with depth in these regions.
The properties of the lithosphere along the San Andreas fault are not at all typical of continental areas, and further characterization of these properties presents a significant scientific challenge.
Stimulating discussions with C.M. Wentworth and M.C. Blake, Jr., led to some of our ideas concerning the configuration and emplacement history of the inferred wedge of Franciscan rocks in central and northern California. A. Griscom and R. C. Jachens provided insightful discussions regarding interpretation of potential-field models, and R.C. Evarts assisted us in drawing a cross section through the uppermost crust in the north-eastern Diablo Range. We are also indebted to E.D. Humphreys for providing pre prints of manuscripts on lithospheric motion in southern California.