Please Note: This is an archived website. It is recommended to search on-line for revised or newer information.
USGS Banner with Coachella Valley as seen from Keyes View in Joshua Tree National Park
Western Earth Surface Processes Team

San Andreas Fault System in the Inland Empire and Salton Trough

Neotectonic Framework of the South-Central Transverse Ranges and Vicinity

Seismicity, Slip Rates, and Strain

Seismicity

Early studies of seismicity in the vicinity of the south-central Transverse Ranges were published by Dehlinger (1952), Richter and others (1958), Allen and others (1965), Brune and Allen (1967), Cheatum and Combs (1973), Hadley and Combs (1974), Thatcher and others (1975), and Fuis and Lamanuzzi (1978). Recent studies have been conducted by Green (1983), Nicholson and others (1983, 1984a, b, 1986), Williams and others (1984), Corbett and Hearn (1984), Sanders and Kanamori (1984), Webb and Kanamori (1985), Jones (1988), and Jones and others (1988).

Richter and others (1958) evaluated the 1948 Desert Hot Springs earthquake (ML=6.5), which they attributed to the Coachella Valley segment of the San Andreas fault (their Mission Creek fault). However, epicenters for this earthquake and for associated shocks did not align with the surface trace of the fault but instead formed a lineament parallel to and several kilometers north-northeast of the trace, and surface ruptures were not reported for the earthquakes. First-motion studies for the 1948 earthquake suggest oblique right-lateral displacement having a thrust component along a fault plane dipping northeast (Allen, 1957, p. 342; Richter and others, 1958). Thus, it is unclear if the 1948 earthquake sequence can be attributed to the Coachella Valley segment.

Allen and others (1965) evaluated the seismicity of the southern California region, including the south-central Transverse Ranges and vicinity, and concluded that although seismicity patterns were consistent with some of the major fault systems, much of the seismicity is diffuse and is not associated with known faults. Brune and Allen (1967) and later workers have emphasized this conclusion with regard to the San Andreas fault system.

Comprehensive analysis of focal mechanisms and hypocentral plots in the south-central Transverse Ranges and vicinity refined the results of earlier studies and offered several new conclusions (Green, 1983; Jones, 1988). (1) Seismicity is not associated with strands of the San Andreas fault (also see Allen and others, 1965; Brune and Allen, 1967; Nicholson and others, 1983). (2) A deep wedge of seismicity in San Gorgonio Pass yields reverse and thrust mechanisms but includes left-lateral and right-lateral mechanisms, and defines the deepest seismicity known from southern California (22 km; also see Fuis and Lamanuzzi, 1978; Nicholson and others, 1983, 1986; Corbett and Hearn, 1984; Webb and Kanamori, 1985). This wedge is bounded on the north by the Mission Creek fault, on the west by the Banning Canyon-Burro Flat segment of the San Bernardino strand of the San Andreas, on the south by the Banning fault, and on the east by a transitional boundary with an area of low seismicity in the northern Coachella Valley. (3) Pure dip-slip mechanisms and oblique dip-slip mechanisms with a left-lateral component occur in the vicinity of the Crafton Hills horst-and-graben complex (also see Nicholson and others, 1983, 1986), and normal mechanisms occur locally within the San Bernardino valley region (Jones, 1988; also see Webb and Kanamori, 1985, fig. 5a). (4) Left-lateral mechanisms appear to define northeast-oriented seismicity lineaments that traverse the San Bernardino valley region (also see Nicholson and others, 1983, 1986). (5) No seismicity is associated with the Coachella Valley segments of the San Andreas and Banning faults.

In contrast to the San Andreas fault, active seismicity is associated with the trace of the San Jacinto fault, which in the vicinity of the south-central Transverse Ranges has generated at least three historic earthquakes of magnitude (ML) 6 or greater. Two of these occurred in the southern San Jacinto Valley (Sanders and Kanamori, 1984, p. 5874-5875, fig. 2; note slight differences in epicentral position compared to Thatcher and others, 1975, fig. 1); the third probably occurred near Loma Linda on either the San Jacinto fault or the nearby Loma Linda fault (Sanders and Kanamori, 1984, p. 5873-5874, fig. 2). Both macro- and microseismicity define relatively quiet and active segments of the fault. For example, rates of microseismicity are relatively high in the southern part of the San Jacinto Valley but are significantly lower in the Reche Canyon area (Brune and Allen, 1967, locs. 13 and 37 of Table 1 and fig. 3; Cheatum and Combs, 1973, p. 5). Green (1983, pl. 1) indicates that microseismicity is low along the San Jacinto fault between Reche Canyon and Cajon Pass, but she reported a dense cluster of seismicity presumably associated with the Loma Linda fault 2 km east of the San Jacinto fault near Loma Linda. This cluster is associated with the convex-west bow in the regional strike of the San Jacinto fault. Thatcher and others (1975) used microseismicity patterns to suggest that a gap in seismic slip occurs on the San Jacinto fault between Reche Canyon and the Cajon Pass region.

Slip Rates

San Andreas fault, Coachella Valley segment . Studies by Keller and others (1982) in the southern Indio Hills indicate that the Coachella Valley segment of the San Andreas fault (their Mission Creek fault) has late Quaternary slip rates of between 10 and 35 mm/year, with a best estimate of 25 to 35 mm/year; this rate presumably applies to the Holocene as well as the late Pleistocene. During late Pleistocene time, this amount of slip probably carried up the Coachella Valley segment and through the San Bernardino Mountains on the Mill Creek fault. However, as discussed below, during Holocene time the 25- to 35-mm rate may have carried no farther northwest than the vicinity of Desert Hot Springs.

San Gorgonio Pass fault zone . Ongoing studies by the U.S. Geological Survey (J.C. Tinsley and J.C. Matti, unpubl. data) indicate that the strain budget for thrust faults of the San Gorgonio Pass zone may be quite complex. A thrust-fault scarp trenched on the Millard Canyon fan disrupts young alluvial deposits along fault planes that dips north at less than 20°. Preliminary age determinations from detrital charcoal suggest that some of the faulted deposits may be as young as 2,850 to 3,600 yr (14C ages uncorrected). An intense scarp-building period during the early part of this period was followed by a period during which only one scarp-forming event may have occurred in the last 2,850 years. These data suggest that modern seismicity patterns may be spatially and temporally cyclic, with high-strain periods followed by low-strain periods.

San Jacinto fault . Sharp (1981) indicates a minimum Quaternary long-term slip rate of about 8 to 12 mm/year for the fault in the vicinity of Anza. South of metropolitan San Bernardino, Wesnousky and others (1991; Prentice and others, 1986) determined a minimum slip rate of 1.7 to 3.3 mm/yr for the San Jacinto fault, although their studies did not capture all strands of the San Jacinto zone. Prentice and others (1986) propose that the long-term Quaternary slip rate on the northern San Jacinto fault probably averages about 10 mm/yr. However, Morton and others (1986) discuss San Bernardino Mountains-derived clast populations in Pleistocene deposits of the San Timoteo Badlands that may require accelerated long-term slip on the San Jacinto fault particularly during geologic intervals when strands of the San Andreas fault in the San Bernardino Mountains region were tied up within the San Gorgonio Pass structural knot.

San Andreas fault, San Bernardino strand . In the Cajon Pass region the San Bernardino strand has a Holocene slip rate of about 25 mm/year (Weldon and Sieh, 1985). Rasmussen (1982) indicates a similar rate farther southeast in the vicinity of Highland.

The San Bernardino strand southeast of Santa Ana River may have a variable slip-rate history that reflects the complex geologic evolution of the strand. In order to determine long-term slip rates for the fault at its southeast extent, Harden and Matti (1989) examined soil profiles of alluvial units in the Yucaipa area that contain clast populations derived from various distinctive bedrock units that crop out on Yucaipa Ridge. Throughout Holocene and latest Pleistocene time these alluvial-fan units have been displaced right-laterally away from their source areas. Within the limits of uncertainty posed by the soils data and by ambiguities in the displacement paths of the alluvial-fan deposits, Harden and Matti (1989) reached two main conclusions: (1) Holocene slip rates on the San Bernardino strand here approach the 25 mm/yr rates proposed for the fault in the Cajon Pass region (Weldon and Sieh, 1985; McFadden and Weldon, 1987); and (2) late Pleistocene slip rates appear to be considerably lower than Holocene rates, on the order of 6 to 13 mm/yr. These conclusions suggest that long-term slip on the San Bernardino strand has accelerated with time. We believe that this slip-rate scenario is compatible with gradual inception of the San Bernardino strand by reactivation of the abandoned Mission Creek strand starting in late Pleistocene time (discussed above).

Cucamonga fault zone . The Cucamonga fault zone at Day Canyon in the east-central part of the zone has a minimum convergence rate of about 5 mm/year for the last 13,000 years (J.C. Matti, D.M. Morton, J.C. Tinsley, and L.D. McFadden, unpubl. data). Age control for this determination is based on correlation of pedogenic soils that cap faulted alluvial deposits; the faulted deposits could be younger (but probably not older), in which case the convergence rate would be greater. Matti and coworkers conclude that earthquakes with vertical displacements of about 2 m had an average recurrence of about 625 years. Seismic-moment calculations indicate expectable surface-wave magnitudes (MS) of 6.9 to 7.2 for fault-rupture lengths of 10 to 25 km.

Strain

Trilateration measurements in the Salton Trough indicate that the Coachella Valley segment of the San Andreas fault is accumulating strain at a rate of about 25 mm/year (Savage and others, 1979, table 3; King and Savage, 1983; Savage, 1983, figs. 2, 3). Line-length measurements by Savage and Prescott (1967) across a doubly braced quadrilateral spanning the San Jacinto fault in the San Jacinto Valley indicate that 25 mm/year of right-lateral shear strain is accumulating across the fault, assuming it is locked to depths of 20 km; this rate contrasts with a minimum long-term Quaternary rate of 8 to 12 mm/year determined by Sharp (1981) from geologic data further to the southeast.

Neotectonic Framework

Most of the neotectonic elements in the vicinity of the south-central Transverse Ranges occur within or adjacent to San Gorgonio Pass a structural knot in the modern San Andreas fault. We propose that many of these fault complexes owe their origin and kinematics to this structural knot.

The San Gorgonio Pass knot lies between the Coachella Valley and Mojave Desert segments of the San Andreas fault. Both segments approach the knot in a straightforward manner, but as Allen (1957, p. 337-339) originally demonstrated, neotectonic strands of these faults cannot be traced as continuous features through the San Gorgonio Pass region. Similar difficulties exist for the Coachella Valley segment of the Banning fault: the regional strike of this strand is aligned with the San Bernardino strand of the San Andreas fault, suggesting a geometric and kinematic relation between the two faults, but Allen (1957) showed that Quaternary thrust and reverse faults in San Gorgonio Pass obscure relations between them. Allen (1957, p. 337) explored five ways that the modern San Andreas might pass through the San Gorgonio Pass knot, of which three are major possibilities:

(1) The aligned San Bernardino strand and Coachella Valley segment of the Banning fault once formed a straight throughgoing trace within the San Gorgonio Pass region, but this trace subsequently has been deflected into a sharp bend that has created the San Gorgonio Pass knot. This interpretation is favored by Dibblee (1968; 1975a, p. 134; 1982, p. 166).

(2) The aligned San Bernardino strand and Coachella Valley segment of the Banning fault may form a straight throughgoing trace within the San Gorgonio Pass region, but this trace is concealed beneath a major thrust sheet of crystalline rock. This interpretation requires that right-slip occurs beneath the thrust plate and that the two strike-slip faults plunge beneath the thrust sheets as they approach San Gorgonio Pass from the northwest and southeast, respectively. Although such a relation can be mapped where the Coachella Valley segment of the Banning fault enters San Gorgonio Pass, similar relations have not been demonstrated for the San Bernardino strand where it would have to plunge southeastward beneath the thrust sheet.

(3) The San Bernardino strand and Coachella Valley segment of the Banning fault may be aligned and may interact kinematically, but the two strands never had a throughgoing connection. Thus, the San Bernardino strand dies out as it approaches San Gorgonio Pass, and neotectonic displacements on the Banning fault have been taken up by compressional convergence in the Pass region. Allen (1957, p. 338-339) favored this interpretation, and we concur.

The fact that none of the neotectonic right-lateral faults of the Coachella Valley can be traced northwestward through the greater San Gorgonio Pass region raises a major question: how is right-slip in the Salton Trough passed through or around the San Bernardino Mountains? We address this question below.

The San Andreas fault system in the vicinity of the central Transverse Ranges: a product of left and right steps in a right-lateral fault zone

The neotectonic setting of the San Gorgonio Pass region owes its origin and kinematics to a bottleneck that gradually evolved in the San Andreas fault during the Pleistocene as the San Bernardino Mountains block was projected across the path of the San Andreas fault and multiple right-lateral strands successively were deformed and abandoned. The geometric effect of these events is apparent from a geologic map of the region: the northwest-oriented trace of the Coachella Valley segment of the San Andreas fault is offset or stepped left about 15 km from the northwest-oriented trace of the Mojave Desert segment (fig. 5). The modern neotectonic framework thus has inherited a bottleneck that must be accommodated in the late Quaternary strain budget of the region.

Any neotectonic model that attempts to distribute strain through or around the San Gorgonio Pass bottleneck must accommodate the following elements: (1) right slip on the Coachella Valley segment of the San Andreas fault falls off as the segment approaches the Transverse Ranges segment; (2) convergence is occurring in San Gorgonio Pass; (3) the San Bernardino valley region is undergoing extension; and (4) right slip on the Mojave Desert segment carries southeastward toward the San Bernardino Mountains segment by way of the San Bernardino strand, but does not carry simply or easily through that segment. To accommodate these elements we propose a speculative model (fig. 6) in which slip is carried around, not through, the Transverse Ranges segment of the San Andreas fault by a complicated series of left and right steps that have created compressional and extensional fault complexes in San Gorgonio Pass and the San Bernardino valley region.

We start with the premise that right-slip occurs on the San Andreas fault in the Coachella Valley but does not carry through the San Bernardino Mountains. We assume that about 25 mm of annual slip occurs on the Coachella Valley segment of the San Andreas between the Salton Sea and the northern Coachella Valley; this figure is consistent with a range of slip values indicated by geologic and geodetic data (Keller and others, 1982; Savage and others, 1979; Savage, 1983). Modern neotectonic slip accounts for the youthful tectonic geomorphology displayed by the San Andreas fault along this segment (Keller and others, 1982; Clark, 1984). However, northwest of Desert Hot Springs, the Coachella Valley segment loses its fresh tectonic geomorphology and our preliminary data suggest that late Quaternary alluvial units have not been displaced significantly by the fault. Farther northwest, the Wilson Creek, Mission Creek, and Mill Creek strands of the San Bernardino Mountains segment are paleotectonic faults that have been abandoned as throughgoing right-lateral strands of the San Andreas fault. Thus, we conclude that during late Quaternary time, most if not all right slip on the San Andreas fault in the northern Coachella Valley has stepped left onto the Banning fault. This process may account for two features. (1) As slip has been transferred across the gap between the two faults, the youthful Indio Hills have been squeezed into an anticlinal uplift. Thus, some percentage of right-slip on the San Andreas would be converted into compressional strain. (2) A left step between the San Andreas and Banning faults in the northern Coachella Valley may explain the absence of fresh tectonic geomorphology for the Banning fault near its junction with the San Andreas fault in the southern Indio Hills (Keller and others, 1982): youthful slip along this segment of the Banning fault would not be necessary if right slip were transferred to the strand farther to the northwest.

Between the Indio Hills and San Gorgonio Pass, late Quaternary right-slip on the Banning fault is indicated by youthful tectonic geomorphology and by right-lateral displacement of late Pleistocene fluvial gravels 2 or 3 km into the Pass (Sheet 2, F-F', G-G'). Moreover, Allen and Sieh (1983) report 2 mm of annual creep on the fault just east of San Gorgonio Pass. The Holocene history of the Banning fault in the Coachella Valley has not been documented, however, and modern right-slip may step still farther west (left) from the Banning fault onto the Garnet Hill fault. This speculation is based on two features: (1) Several youthful dome-like uplifts of Quaternary gravel that occur between the two faults in the vicinity of Whitewater River and Garnet Hill (Allen, 1957, fig. 1 of pl. 6; Dibblee, 1982, p. 166, oblique aerial photograph) may reflect compression within a left-stepping zone; and (2) geomorphic evidence suggests that late Quaternary fluvial gravels in the east part of San Gorgonio Pass may have been displaced right-laterally by the Garnet Hill fault. Whether or not latest right-slip has occurred on the Banning fault or the Garnet Hill fault, neither strand can be traced beyond the eastern San Gorgonio Pass area, and late Quaternary right-slip in the Coachella Valley must have been partly absorbed within the San Gorgonio Pass fault complex.

Although right-slip on the San Andreas fault largely may have been absorbed by convergence within San Gorgonio Pass, some component of slip may step left through San Gorgonio Pass and onto the San Jacinto fault, where it would be added to the 10 mm/year average slip determined by Sharp (1981) for the fault in the Anza area (fig. 6). Local acceleration of slip on the San Jacinto might explain four features of the region. (1) Northwest-trending faults of the Beaumont Plain that appear to have normal dip-slip displacements may reflect extensional fragmentation created as slip steps left to the San Jacinto fault. (2) The San Jacinto Valley is a graben that is rapidly subsiding (Morton, 1977) between right- and left-stepping strands of the San Jacinto fault (Cheatum and Combs, 1973, figs. 2, 4). Rapid subsidence may reflect addition of right-slip acquired from the San Andreas fault. (3) The San Jacinto Valley has been the site of two earthquakes of magnitude (ML) greater than 6.5 during the last 85 years (Thatcher and others, 1975; Sanders and Kanamori, 1984), and the southern San Jacinto Valley has high rates of microseismicity (Brune and Allen, 1967; Cheatum and Combs, 1973); this may reflect an increased potential for seismic activity in response to locally accelerated slip in the San Jacinto Valley area. (4) The San Jacinto fault in the San Jacinto Valley is accumulating about 25 mm/year of right-lateral shear strain (Savage and Prescott, 1967); this departure from the long-term slip rate determined by Sharp (1981) may reflect local acceleration of strain accumulation due to slip acquired from the San Andreas fault.

The neotectonic framework of the San Bernardino valley region includes several distinctive features whose origin and kinematics may require transfer of slip from the San Jacinto fault back to the San Andreas (fig. 6). (1) The San Bernardino strand of the San Andreas fault appears to die out southeastward toward San Gorgonio Pass; (2) the greater San Bernardino valley region is the site of dip-slip fault complexes like the Crafton Hills horst-and-graben complex and the Peters and Tokay Hill faults, which appear to represent an extensional strain field; (3) south of the San Bernardino valley, the San Jacinto fault has a pronounced convex-west bend which may form an impediment to right-slip; (4) the San Jacinto fault between Reche Canyon and Cajon Pass may represent a seismic-slip gap (Thatcher and others, 1975); (5) Morton and Matti (1987) have shown that the San Jacinto fault in the southeastern San Gabriel Mountains does not rupture latest Quaternary alluvium; the youngest branch of the San Jacinto system in this vicinity appears to be the Glen Helen fault, and even this strand is concealed by youngest alluvial deposits in the Cajon Pass region; and (6) the San Bernardino valley region is traversed by northeast-trending left-lateral seismicity lineaments (Green, 1983; Nicholson and others, 1986) that may define the boundaries of clockwise-rotating blocks (Nicholson and others, 1986). In combination, these features may require a common explanation.

We propose that slip on the San Jacinto fault gradually steps right onto the modern San Andreas fault throughout the San Bernardino valley region (fig. 6). By this interpretation, the San Bernardino valley has moved northwestward away from the San Gorgonio Pass region, and the crust between the two regions is pulling apart. This extension is manifested by faults like those in the Crafton Hills extensional complex. Northwestward movement of the San Bernardino valley has occurred along the San Bernardino strand, which extends as a youthful neotectonic feature southeastward from Cajon Pass to the Crafton Hills-Oak Glen region but may not necessarily continue through San Gorgonio Pass and on into the Coachella Valley to the Banning fault. In the Devore area southeast of Cajon Pass, slip may step right from the San Jacinto fault to the Glen Helen fault, which has scarps and sag ponds in the Devore area, and thence to the San Bernardino strand thereby creating an extensional strain field that gives rise to normal dip-slip displacements on the Peters and Tokay Hill faults. A right step from the San Jacinto to the Glen Helen may explain a distinctive seismicity lineament between the inferred traces of the two faults beneath the floodplains of Cajon and Lytle Creeks (Green, 1983, fig. 7). Extension created by a regional right step from the San Jacinto to the San Andreas fault may occur throughout the San Bernardino Valley region, and may create high heat flow that accounts for hot springs and subsurface hot-water zones that occur at several locations in the valley region.

Right-stepping transfer of slip and (or) accumulated strain from the San Jacinto to the San Andreas would create a right-lateral shear couple that could generate clockwise block rotations of the type proposed by Nicholson and others (1986). During the period between large earthquakes on either the San Jacinto or San Andreas faults (the interseismic period of Nicholson and others, 1986), accumulated shear strain within the San Bernardino valley region partly could be released by block rotations and extensional faulting; large earthquake events on the San Andreas and San Jacinto faults would release strain accumulated along the margins of the shear couple. Thus, two styles of seismicity might alternate through time.

We have not documented geometric and kinematic relations between the San Jacinto, San Andreas, and Cucamonga faults in the vicinity of Cajon Pass and the southeastern San Gabriel Mountains. However, one point is clear: at the surface, right-slip on the San Jacinto fault does not pass easily into the San Andreas fault. For example, northwestward migration of the Perris block by right-lateral displacements on the San Jacinto fault partly has been taken up by late Pleistocene and Holocene thrust-fault displacements within the Cucamonga fault zone (Morton and others, 1982; Matti and others, 1982; Morton and Matti, 1987; J. C. Matti, D. M. Morton, J. C. Tinsley, and L. D. McFadden, unpubl. data). Thus, the fault zone represents a zone of convergence between the Peninsular and Transverse Ranges Provinces: to the south, the Perris block and Peninsular Ranges are slipping northwestward along traces of the San Jacinto fault zone; however, this right-lateral migration apparently is impeded by the eastern Transverse Ranges, and the Perris block and alluviated lowlands of the upper Santa Ana River Valley apparently are being thrust beneath the eastern San Gabriel Mountains.

Convergence rates across the Cucamonga fault must be factored into the overall strain budget of the region. Here, the neotectonic San Andreas and San Jacinto faults have late Quaternary slip rates of 25 mm/year and 8 to 12 mm/year, respectively (Weldon and Sieh, 1985; Sharp, 1981). Our studies suggest a minimum convergence rate of 5 mm/year for the Cucamonga fault zone during latest Pleistocene and Holocene time a rate that could double to 10 mm/year if the faulted alluvial succession proves to be younger than we believe. Thus, if the Cucamonga fault zone represents convergence between the Peninsular and Transverse Ranges Provinces, then half to nearly all of the 8 to 12 mm of annual slip on the San Jacinto fault could have been taken up by latest Pleistocene and Holocene convergence within the Cucamonga fault zone. Such a model would imply that part or all of the slip on the San Jacinto fault has not contributed to slip on the San Andreas during latest Quaternary time. Viewed in this way, the Cucamonga fault may represent a major zone of convergence between large crustal blocks.

By contrast, Weldon (1984, 1985a,b) suggests that, even though the San Jacinto fault zone may not have a surface connection with the San Andreas fault (Morton, 1975b), the 8 to 12 mm of annual slip on the San Jacinto nevertheless feeds into the San Andreas and contributes to slip on that fault. If this interpretation is correct, then the annual 5-mm convergence rate within the Cucamonga fault zone may not reflect wholescale convergence between major crustal blocks of the Peninsular and Transverse Ranges but instead may simply reflect interactions between local small blocks in a region where the San Jacinto and San Andreas faults merge in a complicated manner. This interpretation might also account for the geographically segmented strain-release behavior that appears to have characterized the Cucamonga fault zone during latest Pleistocene and Holocene time (Morton and Matti, in press).



Return to the Inland Empire homepage.

FirstGov button