|Western Earth Surface Processes Team|
San Andreas Fault ZoneGeneral Statement
The main strand of the San Andreas fault in southern California consists of two segments (fig. 1): (1) the Mojave Desert segment that mainly separates rocks of San Gabriel Mountains-type from rocks of San Bernardino Mountains-type, and (2) the Coachella Valley segment that separates rocks of Peninsular Ranges-type from rocks of San Bernardino Mountains- and San Gabriel Mountains-type. In plan view the main strand has a left-stepping geometry: the Mojave Desert segment has been stepped left (west) about 15 km from the Coachella Valley segment, with the step occurring in the San Gorgonio Pass region of the southeastern San Bernardino Mountains (Matti and others, 1985). This left step forms a structural knot in the San Andreas that has influenced the evolution of the entire transform-fault system and has led to multiple fault strands that evolved sequentially. Multiple strands also have developed along the Mojave Desert segment of the San Andreas (Barrows and others, 1985, 1987). This strand complexity has made it difficult to identify the distribution and displacement history of faults in the San Andreas family, and has led to variable fault nomenclature and to conflicting fault-movement scenarios.
The complex pattern of late Cenozoic right-lateral faults in southern California has led to nomenclature for the San Andreas fault that is more complex than for central California. There, most if not all Miocene and younger displacement on the San Andreas has occurred within a narrow, singular (although complex) zone that extends the length of central California to Hill and Dibblee's (1953) "Big Bend" at the latitude of the Garlock fault. The name "San Andreas fault" has been used by most workers in central California (Hill, 1981), and little confusion has arisen with regard to which geologic structure bears the name "San Andreas fault" or whether multiple fault strands have generated sequential displacements. By contrast, the San Andreas fault in southern California consists of multiple strands, each representing some portion of the geologic history allocated to the more singular zone in central California (Noble, 1932; Dibblee, 1954, 1968a; Allen, 1957; Crowell, 1962; Woodburne, 1975; Matti and others, 1985). This structural complexity has presented two challenges: (1) to document the distribution of the various fault strands and to determine their sequencing and amount of right-lateral displacement, and (2) to establish nomenclature that provides a logical framework for understanding faulting history and for relating fault segments having similar and/or dissimilar movement histories.
A nomenclatural framework for the San Andreas fault in southern California has evolved through the efforts of many workers (see the historical review by Hill, 1981). Early workers recognized that the characteristic geomorphic and geologic features of the San Andreas rift zone in central California extend southeast beyond Hill and Dibblee's Big Bend region and intervene between the Mojave Desert and the massifs of Liebre Mountain and the San Gabriel Mountains. On this basis, the name "San Andreas fault" originally was extended into southern California. However, Noble (1926, 1932) was among the first to observe that the San Andreas in southern California splays into several major branches some occurring close to each other within the narrow Mojave Desert zone, others forming discrete strands that follow independent traces many kilometers apart. As Allen (1957) pointed out, this strand complexity creates a dilemma: which strand should bear the name San Andreas fault and which strand generated the large displacements proposed by Hill and Dibblee (1953) for the fault in central California?
Whenever possible, we refer to the San Andreas fault proper in southern California by one of the specific strand names identified in Table 1. However, because each of these strands merged with or fed into the San Andreas fault in central California, each southern California strand during its lifetime represented the San Andreas before it was abandoned and succeeded by the next "San Andreas." This iterative pattern has culminated in the modern San Andreas fault in southern California the genetically related set of strands that most workers believe originated onshore in response to Pliocene opening of the Gulf of California by sea-floor spreading and transform faulting (Atwater, 1970). The name San Andreas fault usually is applied to this modern strand. Within this complex geologic and nomenclatural framework, we use the term "San Andreas (sensu lato)" for all strands of the San Andreas in southern California that have fed into the central California segment of the fault and contributed to its total history; we use the term "San Andreas (sensu stricto)" for those strands in southern California that have contributed to displacements on the fault only in Pliocene and Quaternary time. Thus, we can refer to the San Andreas fault generically without reference to a particular strand, but at the same time distinguish between broader vs. narrower interpretations of the name.
San Andreas Fault Zone, Mojave Desert Segment
The Mojave Desert segment of the San Andreas fault extends from Tejon Pass to the San Bernardino valley, where it passes into the San Bernardino strand. The modern break forms a singular trace that runs the entire length of the segment (Ross, 1969) and describes a gently-curving arc having a regional strike of about N 60 W; in Tejon Pass the trace is deflected into the big bend of Hill and Dibblee (1953, p. 453). Ground rupture associated with the 1857 earthquake on the San Andreas fault occurred along the Mojave Desert segment from Tejon Pass to about Wrightwood (Sieh, 1978a), and the modern trace has been the site of recurring Holocene ground rupture (Sieh, 1978b, 1984; Weldon and Sieh, 1985).
The Mojave Desert segment has been described thoroughly by Barrows and others (1985, 1987) on the basis of detailed mapping by the California Division of Mines and Geology (Barrows, 1975; Kahle, 1975; Barrows, 1979, 1980; Barrows and others, 1976; Beeby, 1979; Kahle and others, 1975; Kahle, 1979; Kahle and Barrows, 1980). Their mapping shows that from Tejon Pass southeast to Elizabeth Lake the modern trace coincides with older traces to form a narrow fault zone having a relatively simple faulting history. By contrast, between Elizabeth Lake and Cajon Pass the modern trace is but one of several fault strands that form a zone several kilometers wide (see similar interpretations by Noble, 1926, 1933, 1954a,b; Wallace 1949; Dibblee, 1967a, 1968a, 1975b). Within this wide zone, fault strands like the Punchbowl, Nadeau, and Little Rock faults have evolved sequentially and been abandoned, culminating in the modern trace that apparently evolved no earlier than the Pleistocene (Barrows and others, 1985, p. 105-106).
Geologic Setting in the San Gabriel Mountains
In the San Gabriel Mountains Barrows and others (1985, 1987; Barrows, 1987) recognize five discrete strands of the San Andreas fault including, from west to east, the Punchbowl, Nadeau south, Nadeau north, Mojave Desert, and Little Rock strands (we use the term "Mojave Desert strand" to designate the "main San Andreas trace" of Barrows and others, 1985, 1987). Careful mapping by Barrows and others (1985) shows that the Punchbowl, Nadeau (north and south branches), and Little Rock faults all are truncated on their northwest and southeast ends by the Mojave Desert strand of the San Andreas. Thus, they can be viewed as anastomosed and abandoned strands of the San Andreas that probably are related to strands of the fault farther to the southeast in the San Bernardino Mountains and (or) Salton Trough.
Regional correlation of the Punchbowl fault is of particular interest because the fault appears to be a significant strand of the San Andreas, and yet its isolated position outboard (west) of the main San Andreas trace makes its role in the overall history of the San Andreas difficult to evaluate. The fault extends for about 75 km along the northeast flank of the San Gabriel Mountains (fig. 1; Barrows and others, l987). The strand is characterized by its locally sinuous trace and west-dipping reverse dips, a geometry attributed by Barrows and others (l987) to deformation following its right-slip history. The Punchbowl fault originally was identified by Noble (1953), who interpreted it as a reverse dip-slip fault and implied that it merged with the San Jacinto fault by way of the Glen Helen fault (Noble, 1954a,b). Subsequently, Dibblee (l967a, 1968a) showed that the Punchbowl fault is a right-lateral fault that he interpreted as an old strand of the San Andreas family. Because it is truncated on the southeast and northwest by younger strands of the San Andreas zone, the Punchbowl should be viewed like any other linear structure that has been displaced by a strike-slip fault: its northwest and southeast terminations have been displaced from cross-fault counterparts that should be identifiable after right slip is restored on younger strands of the San Andreas. This interpretation motivated us to search southeastward for a possible displaced counterpart for the Punchbowl fault on the opposite side of the main trace of the San Andreas.
Faulting chronology. The multiple San Andreas strands in the San Gabriel Mountains evolved sequentially about 5 m.y. ago (Barrows and others, 1985, 1987). The Punchbowl fault appears to be the oldest strand, although Barrows and others (1987, p. 2, 84, 86) indicate that sequencing relations between the Punchbowl and Little Rock strands are not completely clear. According to Barrows (1987, p. 149-153, figs. 6-8), stratigraphic relations between the Punchbowl fault and sediments of the Juniper Hills Formation (of Barrows, 1987) indicate that the Punchbowl fault may have generated right slip continuously between early Blancan (5 Ma) and late Blancan time (2 Ma). However, the age of the Juniper Hills Formation is poorly constrained (Barrows, 1987, p. 129-130), and relevant beds in the unit cannot be confidently assigned to a particular part of the Blancan (A.G. Barrows, oral communication, 1990). The Nadeau faults appear to be Pliocene in age, while the Mojave Desert strand may have originated in middle Pleistocene time.
Displacement history. Barrows and others (1985, Table 4, 1987 p. 86) were able to document no more than 102 km of right slip on all strands of the San Andreas fault in the eastern San Gabriel Mountains: 21 km on the main or Mojave Desert strand of the San Andreas; 16 km on the Nadeau fault (north and south branches); 21+ km on the Little Rock fault; and 44 km on the Punchbowl fault. Their estimates for the Punchbowl fault are comparable with those proposed by other workers. Dibblee (1967a, fig. 72; 1968a, p. 263-264, fig. 1) recognized between 32 and 48 km of displacement on the Punchbowl based on cross-fault correlations between the San Francisquito and Fenner faults, between marine rocks of the San Francisquito Formation, and between the Sierra Pelona and Blue Ridge windows of Pelona Schist. Farley and Ehlig (1977) and Ehlig (1981, fig. 10-4) proposed about 40 km of displacement based on their suggestion that the Punchbowl fault has displaced strata in Ridge Basin that contain polka-dot granite clasts from strata in the Punchbowl Formation that contain similar clasts. Barrows and others (1985, 1987) acknowledge that their total displacement for the San Andreas (102 km) is considerably less than the widely accepted displacement (240 km), and they point out that displacement on the Little Rock strand may be greater than the 21 km they were able to document. Their displacement estimate for the main San Andreas strand (21 km) also is considerably less than most workers would infer for the strand.
The Coachella Valley segment of the San Andreas fault is a relatively simple fault zone, although locally it is complicated by en-echelon strands and lateral splays (Clark, 1984). The segment is relatively straight and generally has a uniform regional strike of about N 45° W, although Bilham and Williams (1985) identified alternating segments 9 to 14 km long that have trends of N 40° W and N 48° W, respectively. Youthful tectonic landforms and faulted Holocene alluvial deposits indicate that the Coachella Valley segment is a modern neotectonic element (Keller and others, 1982), but the antiquity and tectonic significance of the segment cannot be judged because Quaternary and late Tertiary sediments of the Salton Trough conceal older rocks bearing evidence for its full history.
The southeast and northwest terminations of the Coachella Valley segment of the San Andreas involve complex interactions with other fault zones. To the southeast, surface expression of the segment terminates near the southeast margin of the Salton Sea, where it interacts with the Brawley seismic zone and the Imperial fault (Sharp, 1982; Johnson and Hill, 1982). Some workers view this segment of the San Andreas fault as the northwesternmost of a series of right-stepping transform faults that extend from the Gulf of California onshore into the Salton Trough (Moore and Buffington, 1968, fig. 4; Elders and others, 1972, fig. 1; Crowell and Ramirez, 1979; Lonsdale and Lawver, 1980, fig. 1; Crowell, 1981, fig. 18-4; Johnson and Hill, 1982, fig. 6; Curray and Moore, 1984, figs. 1, 9).At its northwest end the regional strike of the Coachella Valley segment is deflected westward, and it splays into the Mission Creek and Mill Creek strands of the southeastern San Bernardino Mountains. We cannot prove that these multiple strands merge beneath the Coachella Valley because they are buried by unfaulted Quaternary alluvium where they exit the San Bernardino Mountains. However, their map pattern in the mountains strongly suggests that the strands coalesce southeastward, and by the simplest interpretation they ultimately form a single fault zone the Coachella Valley segment of the San Andreas fault.
Northwest of Desert Hot Springs the Coachella Valley segment of the San Andreas fault loses its clear surface expression, and Holocene displacements have not been demonstrated for the segment. The fault forms conspicuous scarps in Quaternary alluvium southeast of Desert Hot Springs, and discontinuous scarps can be traced northwestward where they disrupt young (but not youngest) alluvium in the center of town (Clark, 1984). However, to the northwest, latest Quaternary alluvial fans that flank the Little San Bernardino Mountains are not disrupted by the fault. The late Quaternary history of the Coachella Valley segment has not been worked out in this region.
Displacement history. Although small youthful displacements on the Coachella Valley segment of the San Andreas fault have been recognized on the basis of cross-fault correlations between Quaternary alluvial materials (e.g., Keller and others, 1982; Matti and others, 1985), large older displacements recognized on the basis of cross-fault correlation between Cenozoic and pre-Cenozoic units generally have not been recognized. In part this reflects the fact that older rocks of appropriate age containing evidence for large-scale displacements largely are buried by young Quaternary sediment that has filled the Salton Trough. In addition, the widely cited model for 240 km of Pliocene and Quaternary displacement on the San Andreas fault in effect has dampened the search for pre-Pliocene cross-fault counterparts in the Salton Trough because the 240-km model requires that pre-Pliocene rocks have been displaced completely out of the Salton Trough region.
A study by Dillon (1975) suggests that this may not be the case. In his study of the southern Chocolate Mountains, Dillon (1975, fig. 70, p. 334-365) proposed that rocks in the southeastern San Bernardino Mountains have been displaced from the southern Chocolate Mountains by 180 ± 20 km of right slip on the Coachella Valley strand of the San Andreas (Dillon's north branch of the San Andreas). Dillon's proposal is based on three cross-fault correlations: (1) crystalline rocks of San Gabriel Mountains-type northeast of San Gorgonio Pass correlated with similar rocks in the vicinity of Mammoth Wash in the Chocolate Mountains (Dillon, 1975, p. 59-60, 351-353); (2) the Miocene Coachella Fanglomerate in the Whitewater area correlated with the fanglomerate of Bear Canyon in the southern Chocolate Mountains (Dillon, 1975, p. 341-346); and (3) the inferred strandline position of the marine Imperial Formation in the Whitewater area correlated with the inferred strandline position for the marine Bouse Formation in the Chocolate Mountain region (Dillon, 1975, p. 347-350, fig. 69). Dillon's reconstruction contrasts with that of Peterson (1975), who restores the Coachella Fanglomerate farther south in the Salton Trough and calls for 215 km of displacement on the Coachella Valley strand of the San Andreas.
The San Andreas fault within and adjacent to the San Bernardino Mountains consists of several strands that evolved sequentially and then were abandoned; to the northwest and southeast, these strands merge to form the Mojave Desert and Coachella Valley segments.
In his geologic reconnaissance of the eastern San Bernardino Mountains, Vaughan (1922) recognized that the region was traversed by two important faults the San Andreas fault, which he projected through San Gorgonio Pass and eastward into the Coachella Valley, and a fault he named the Mission Creek fault, which he projected through the southeastern San Bernardino Mountains. Hill (1928) discussed the San Andreas fault and related faults in the vicinity of the San Bernardino Mountains and named the Pinto Mountain and Mill Creek faults.
Following the observation by Noble (1932) that the San Andreas fault zone southeast of Cajon Pass splits into several major branches, Allen (1957) and Dibblee (1968) addressed the distribution and nomenclature of faults in the vicinity of the San Bernardino Mountains. Allen (1957, fig. 1, p. 336-343) reviewed fault nomenclature used by earlier workers in the San Gorgonio Pass region and clarified the distribution of the Mill Creek fault (adopted from Hill, 1928), the Mission Creek fault (modified from Vaughan, 1922), the Banning fault (recognized originally by Vaughan, 1922, and defined and extended by Hill, 1928), and the modern San Andreas fault. Allen showed that the Mill Creek and Mission Creek faults are strands of the San Andreas fault zone, although he was unable to confirm or refute large right-lateral displacements on the strands. Between Cajon Pass and San Gorgonio Pass, Allen followed tradition by applying the name San Andreas fault to "the most aligned and obvious prolongation" of the San Andreas that extends southeast of Cajon Pass that is, "the fault that lies at the foot of the San Bernardino Mountains and continues into San Gorgonio Pass. Within the Pass area, however, various structural complications make the continuity of the fault through this region doubtful" (Allen, 1957, p. 337). Allen's last sentence epitomizes the notion that San Gorgonio Pass is the site of a knot in the modern San Andreas fault system.
Dibblee modified the distribution and nomenclature of Allen's fault strands in the southeastern San Bernardino Mountains. In his geologic map of the San Gorgonio Mountain quadrangle, Dibblee (1964) equated his north branch of the San Andreas fault with Allen's Mill Creek fault, an interpretation he reiterated later (Dibblee, 1968a, p. 269). Dibblee apparently viewed Allen's Mission Creek fault as a splay of his north branch (Dibblee, 1968a, figs. 3, 4; 1975a, p. 127, figs. 1, 2) and it is not clear from these papers which structure has the greater displacement. Dibblee's recent work (Dibblee, 1982, p. 152, 161-165, figs 1, 4, and 8) clearly equates his north branch with Allen's Mill Creek fault and states that it generated large right-lateral displacements. Dibblee (1968a, p. 268) applied the term south branch of the San Andreas fault to the modern trace along the foot of the San Bernardino Mountains, and projected this trace southeastward through San Gorgonio Pass to join his south branch in the Coachella Valley (Dibblee, 1954). Dibblee's discussion of offsets on the south branch is confusing: he indicates minor displacements on the south branch between San Gorgonio River and San Gorgonio Pass but large displacements on the same strand along the base of the San Bernardino Mountains (Dibblee, 1968a, p. 268), without elaborating on this discrepancy.
Our concept of basement terranes and their bounding faults in the southeastern San Bernardino Mountains borrows some elements from both Allen and Dibblee, but departs from both in the details of fault distribution and movement history. Our view of fault relations is closer to Allen's scheme than to Dibblee's. Despite the unifying appeal of Dibblee's nomenclature, his concept of north and south branches of the San Andreas fault is too simplified and partly is in error as shown by later mapping (Ehlig, 1977; Farley, 1979; Matti and others, 1982a, 1983). Allen's (1957) interpretation of the Mill Creek and Mission Creek faults as independent strands that follow separate routes through the San Bernardino Mountains more accurately delineates the boundaries of crystalline terranes native and exotic to the southeastern San Bernardino Mountains, even though Allen was unable to use the significance of these terranes to reconstruct a comprehensive movement history for his fault strands. Although many workers have adopted Dibblee's nomenclature for the San Andreas fault in the south-central Transverse Ranges, we recommend that this nomenclature be abandoned because it does not accurately reflect the geologic relations.
Wilson Creek and Mission Creek faults. The Wilson Creek and Mission Creek faults are major strands of the San Andreas fault in the south-central Transverse Ranges: both juxtapose exotic, far-traveled crystalline and sedimentary rocks against rocks native to the San Bernardino Mountains. Our recognition of two major fault strands in a region where only the Mission Creek fault has been mapped previously is based on relations in the Raywood Flat area of the southeastern San Bernardino Mountains (Matti and others (l983) and in the Mill Creek area (Matti and others, 1992).
Wilson Creek and Mission Creek faults in the southeastern San Bernardino Mountains. Within crystalline rocks of the southeastern San Bernardino Mountains, the Wilson Creek and Mission Creek faults are closely spaced and traverse the range along a gently bowed, east-trending arc. At the southwest end of this arc, the Wilson Creek and Mission Creek faults diverge and follow separate paths within and adjacent to the mountains. At the east end of this arc, the two faults apparently coalesce to form a single fault that continues southeastward beneath alluvium of the Coachella Valley (Matti and others, 1982a). There, the combined Wilson Creek-Mission Creek strands merge with the Mill Creek strand to form the Coachella Valley segment of the San Andreas fault.
Throughout the southeastern San Bernardino Mountains, the Mission Creek fault occurs outboard (south) of the Wilson Creek fault and bounds a distinctive terrane of crystalline rocks similar to those of the lower and upper plates of the Vincent thrust in the San Gabriel Mountains. These relations can be seen best in the headwaters of San Gorgonio River, where rocks outboard of the Mission Creek fault form two suites separated by a steeply dipping fault that probably originated as a low-angle segment of the region-wide Vincent-Orocopia-Chocolate Mountain thrust system. Lower-plate rocks are Pelona Schist that consist mainly of chlorite-albite-actinolite greenstone; upper-plate rocks are hornblende-bearing granitoid rocks and granitic gneiss that have strongly foliated and layered fabrics created by ductile and brittle-ductile deformation. Small bodies and lenses of the Triassic Lowe plutonic complex (of Joseph and others, 1982) occur locally in the upper-plate sequence (Farley, 1979). These lower- and upper-plate rocks are similar to those that occur in the same structural block in the nearby Crafton Hills, and both sequences in turn are similar to those in the eastern San Gabriel Mountains (Ehlig, 1968b, p. 301, fig. 1, locs. 6-8). These correlations form the basis for our interpreting rocks in the southeastern San Bernardino Mountains as San Gabriel Mountains-type.
Relations between the Mission Creek and Wilson Creek faults can be seen best in the vicinity of Raywood Flat north of San Gorgonio Pass, where the two faults parallel each other closely and separate three different assemblages of crystalline rock. The Wilson Creek fault separates granitoid and gneissic rocks native to the San Bernardino Mountains to the north from a slice of crystalline rocks of unproven provincial affinity to the south; we refer to this enigmatic slice as the Wilson Creek block. The Mission Creek fault bounds the Wilson Creek block on the south and separates it from rocks of San Gabriel Mountains type.
Our interpretations in the Raywood Flat area differ in detail from those of Ehlig (1977) and Farley (1979), who recognized that the Mission Creek fault zone is an important structural boundary but suggested a different interpretation for geologic features that we use to identify the Wilson Creek fault. Critical to any interpretation is the structural position of Pelona Schist and associated crystalline rocks that we believe are bounded by the Wilson Creek and Mission Creek faults.
Pelona Schist forms two bodies in the headwaters of San Gorgonio River (map sheet 2). We agree with Farley (1979), who includes the main Pelona Schist body with rocks of San Gabriel Mountains-type and separates this package from non-related rocks to the north by the Mission Creek fault (map sheet 2; Matti and others, 1983). A second slice of Pelona Schist and associated crystalline rocks just west of Raywood Flat is more problematical, and forms the basis for our recognition of the Wilson Creek fault as a structure distinct from the Mission Creek fault. Farley (1979) viewed the Raywood Flat Pelona Schist body as a fault-bounded slice caught up between two branches of his Mission Creek fault zone, and he concluded that this and the main Schist body belong together as one coherent terrane south of the main Mission Creek strand. In contrast, we believe that the two Pelona Schist bodies and associated crystalline rocks constitute two different assemblages bounded by two major strike-slip faults. The schist bodies differ in their pre-metamorphic stratigraphy, and crystalline rocks in fault contact with the Raywood Flat body are different from rocks of either San Gabriel Mountains-type or San Bernardino Mountains-type.
We use these differences to suggest that the Raywood Flat Pelona Schist body is not a slice caught up in a right-lateral fault zone (Farley, 1979) but instead is a window through crystalline rocks of the Wilson Creek block. By this interpretation, high-angle faults now bounding the window have reactivated and modified a thrust surface (the Vincent-Orocopia-Chocolate thrust) that formerly had a low-angle geometry. Upper-plate rocks of this window constitute rocks of the Wilson Creek block that appear to be unlike those native to the San Bernardino Mountains or those in the upper-plate of the Vincent-Orocopia thrust to the south. These observations lead to our conclusion that two exotic crystalline terranes separated by two major strands of the San Andreas fault occur in the southeastern San Bernardino Mountains: (1) a terrane of San Gabriel Mountains-type rocks outboard (south) of the Mission Creek fault and (2) the Wilson Creek block between the Mission Creek and Wilson Creek faults.
At the west margin of the southeastern San Bernardino Mountains the Wilson Creek-Mission Creek couplet has been modified by low-angle faulting and cannot be recognized in its original configuration. In the headwater region of San Gorgonio River, crystalline rocks native to the San Bernardino Mountains have been thrust southeastward across the Wilson Creek-Mission Creek fault zone and across exotic crystalline rocks outboard of the zone (Farley, 1979; Matti and others, 1983). Low-angle faulting probably evolved as the Wilson Creek-Mission Creek fault couplet was rotated from its original northwest strike into the anomalous east to northeast strike it presently displays in this vicinity. Although the low-angle faults obscure the original course of the Wilson Creek and Mission Creek faults, our mapping and palinspastic reconstructions indicate that they originally continued through and beyond Banning Canyon.
Wilson Creek fault along the southwest margin of the San Bernardino Mountains. Faults we assign to the Wilson Creek strand of the San Andreas occur along the southwest base of the San Bernardino Mountains between Wilson Creek and Cook Creek (map sheets 1, 2). Along this reach, the strand forms a curving, locally sinuous trace sandwiched between the San Bernardino and Mill Creek strands of the San Andreas. Southeast of this reach, between Wilson Creek and Banning Canyon, the Wilson Creek fault either is concealed by Quaternary alluvium or is truncated and displaced by the younger Mission Creek or San Bernardino strands of the San Andreas fault. To the northwest, beyond Cook Creek, the Wilson Creek strand is truncated by the Mission Creek and San Bernardino strands.
Near Wilson Creek, its namesake, the fault enters the San Bernardino Mountains and traverses Yucaipa Ridge before descending into Mill Creek Canyon. To the west, the Wilson Creek fault adopts a more westerly trend and converges with the younger Mill Creek fault. The two faults parallel each other for a few miles before the Wilson Creek strand diverges in the Cook Creek drainage near San Bernardino and exits the San Bernardino Mountains. Along this segment the fault dips between 35° and 65° to the southwest. The 35° dip occurs near the Cook Creek drainage, where the Wilson Creek strand juxtaposes gneissic crystalline rocks against Tertiary sedimentary rocks. There, the low-angle movement zone has been interpreted as a landslide surface (Gary Rasmussen and Associates, unpublished geotechnical report on file with San Bernardino County; G. Rasmussen, oral commun., 1984); however, we interpret it as a segment of the Wilson Creek fault that has acquired a low-angle geometry. West of City Creek, isolated masses of gneissose rock mapped as landslide masses by Morton and Miller (1975, fig. 1c) may be klippe of Wilson Creek block emplaced along a south-dipping and locally north-dipping thrust segment of the Wilson Creek fault.
The distribution, movement sense, and significance of the Wilson Creek fault are interpreted in different ways by different workers. In the vicinity of Yucaipa Ridge and Mill Creek Canyon, faults we assign to the Wilson Creek strand were mapped by Owens (1959), and by R.E. Smith (1959) who identified two fault segments (his South fault and Yucaipa Ridge fault) that he thought were unrelated. Matti and others (1985) combined Smith's South and Yucaipa Ridge faults into their throughgoing Wilson Creek fault. The distribution and displacement history of the fault were examined by Demirer (1986), by West (1987) who referred to the structure as the Yucaipa Ridge fault and restricted the name "Wilson Creek" to another fault zone in the Yucaipa Ridge area, and by Hillenbrand (1990) who mapped the Wilson Creek fault northwest to the Cook Creek area. All previous workers conclude that faults we assign to the Wilson Creek fault are minor structures having a dip-slip geometry.
Stratigraphic comparisons between Tertiary nonmarine sedimentary rocks traversed by the Wilson Creek fault have influenced investigators who view the fault as a minor structure. These rocks originally were grouped by Vaughan (1922) into his vaguely defined Potato Sandstone. In the Mill Creek area the rocks have been mapped and investigated by various graduate students (Smith, 1959; Owens, 1959; Gibson, 1964; Demirer, 1986; West, 1987; Hillenbrand, 1990). Recently, Matti and others (1992) proposed that Vaughan's Potato Sandstone can be separated into the Mill Creek Formation (modified from Gibson, 1964, 1971) and the Formation of Warm Springs Canyon, and they indicate that the two units are separated by the Wilson Creek fault (map sheet 2 of this report). Most workers (Dibblee, 1982; Demirer, 1986; Sadler and Demirer, 1986; West, 1987) recognize strong lithostratigraphic similarities between the units that we separate into two formations, and therefore see no need to juxtapose them along a fault having significant strike-slip movement. We believe that stratigraphic differences between the Mill Creek and Warm Springs Formations outweigh stratigraphic similarities between them (Matti and others, 1992). Moreover, basement rocks of the Wilson Creek block that underlie the Mill Creek Formation are dissimilar to those in the San Bernardino Mountains east of the Wilson Creek and Mill Creek faults, and if the Mill Creek fault is a relatively minor strand of the San Andreas (discussed below), then either the Wilson Creek fault or some other (unrecognized) fault is required to bring the Wilson Creek block into the region.
Mission Creek fault along the southwest margin of the San Bernardino Mountains. Matti and others (1985) inferred the existence of the Mission Creek fault along the southwestern base of the San Bernardino Mountains in order to explain the juxtaposition of San Gabriel Mountains-type basement rocks (upper- and lower-plate rocks of the Vincent thrust) against the Wilson Creek block and rocks of San Bernardino Mountains-type. Juxtaposition of these distinctive crystalline terranes can be documented in the southeastern San Bernardino Mountains where the Mission Creek fault separates rocks of San Gabriel Mountains-type from rocks native to the San Bernardino Mountains and from rocks of the Wilson Creek block (discussed above). Matti and others (1985) reasoned that the Mission Creek strand continues in the subsurface along the southwestern base of the San Bernardino Mountains, where the structure continues to separate outboard rocks of San Gabriel Mountains-type from rocks of the San Bernardino Mountains and the Wilson Creek block. Just as in the southeastern San Bernardino Mountains, where the Mission Creek fault is an older abandoned strand of the San Andreas (Matti and others, 1983, 1985), so too along the southwestern base of the San Bernardino Mountains the Mission Creek fault is an older fault either (1) buried under Quaternary surficial materials or (2) reactivated by the late Quaternary San Bernardino strand. In the latter case, the trace of the San Bernardino strand would mark the trace of the older Mission Creek strand.
Mill Creek strand. The Mill Creek fault was named by Hill (1928), but modern discussions of the fault date from Allen's (1957) clarification of its distribution and geologic setting; our mapping of the fault is similar to Allen's. Within the San Bernardino Mountains the Mill Creek fault occurs inboard (east) of the Wilson Creek and Mission Creek faults, and traverses the mountains along a relatively straight to slightly curving trace that has a regional strike of about N 70° W. The fault zone is relatively simple and narrow. Southeast of the San Bernardino Mountains the Mill Creek fault and the Wilson Creek and Mission Creek faults coalesce to form the Coachella Valley segment of the San Andreas fault. To the northwest the Mill Creek fault exits the mountains near San Bernardino and merges with the San Bernardino strand of the San Andreas fault.
At the southeast end of the San Bernardino Mountains, near the confluence of the North and South Forks of the Whitewater River, we map the Mill Creek fault differently than either Allen (1957) or Dibblee (1967b). Allen (1957, pl. 1) thought that the Mill Creek fault terminated in this vicinity after having traversed much of the San Bernardino Mountains to the west. Dibblee recognized that the fault (his north branch) must continue to the southeast; he (Dibblee, l967b, 1975a) projected the Mill Creek fault eastward beneath alluvium of the Whitewater River and suggested that it truncates the Pinto Mountain fault. We map the Mill Creek fault differently, and propose that it is deflected by the Pinto Mountain fault (Matti and others, 1982a, 1983). Where it obliquely exits the San Bernardino Mountains southeast of the Whitewater River, the Mill Creek fault flanks the mountain front and is located close to but north of the combined traces of the Mission Creek and Wilson Creek faults (Matti and others, 1982a, 1983).
San Bernardino strand. As defined by Matti and others (1985), the San Bernardino strand denotes the modern trace of the San Andreas fault in the vicinity of the San Bernardino Mountains. The fault extends for 60 km along the base of the Mountains from Cajon Pass southeast to the vicinity of Banning Canyon, and describes a gently curving arc that is convex to the south. To the northwest the San Bernardino strand is continuous with the modern trace of the Mojave Desert segment of the San Andreas; to the southeast the strand appears to terminate within the San Gorgonio Pass fault zone.
The San Bernardino strand coincides spatially with the projected trace of the Mission Creek fault, and probably evolved by reactivation of that older strand. By this interpretation, the Mission Creek fault is responsible for juxtaposing exotic far-traveled Pelona Schist bedrock of the San Bernardino valley against the Wilson Creek block to the north, and the San Bernardino strand is only a relatively recent break that developed within or close to the older fault.
We recognize three segments of the San Bernardino strand: a segment extending from Cajon Pass to the vicinity of Mill Creek, a segment extending from Mill Creek to the vicinity of Banning Canyon, and a segment extending from Banning Canyon to the Burro Flats area north of San Gorgonio Pass. Distinction between these three segments is based on contrasts in geologic structure and tectonic geomorphology.
Cajon Pass to Mill Creek segment. The San Bernardino strand between Cajon Pass and Mill Creek is characterized by its conspicuous geologic and geomorphic expression, by its overall simplicity, and by abundant evidence for youthful activity: (1) The segment is relatively continuous and only slightly curved, and is not sinuous. It does not have significant left or right steps along its trace, although minor steps between overlapping fault segments occur locally. (2) Youthful activity along the segment is indicated by well developed primary fault features (scarps, sag ponds, pressure and shutter ridges) and by youthful geologic and physiographic features (alluvial fans, landslides, drainage lines) that have been offset by the fault throughout Holocene time. The fault cuts and forms scarps in all Holocene alluvial units, except for the youngest active stream alluvium. In the active sediments, shallow ground water is backed up behind the fault and its trace is marked by linear vegetation lines. (3) The San Bernardino Mountain front along the segment generally has low topographic relief, particularly to the northeast (Weldon and Meisling, 1982; Weldon, 1983), although relief gradually increases southeastward toward Mill Creek. These relations indicate that this segment of the San Bernardino strand has not generated significant amounts of vertical displacement (Weldon, 1985b). (4) In Cajon Pass the segment has a well documented Holocene slip rate of about 25 mm/year (Weldon and Sieh, 1985). Rasmussen (1982) indicates a similar rate of Holocene slip near San Bernardino.
Mill Creek to Banning Canyon segment. Between Mill Creek and Banning Canyon the San Bernardino strand is characterized by its lack of clear continuous geomorphic expression and by its structural complexity locally.
Between Santa Ana River and the canyon mouth of Mill Creek, the San Bernardino strand maintains its regional strike of about N 70° W but becomes increasingly complex. Directly west of Mill Creek, bedrock and surficial units are traversed by several parallel fault strands, some of which are part of the San Andreas zone; locally, these strands have north-facing scarps (Matti and others, 1992). None of these faults breaks youngest deposits of Mill Creek wash, but on the east side of the wash vintage aerial photographs reveal a scarp that disrupts surficial deposits that have attributes of soil-stage S6 or S7 of McFadden (1982), and the unit probably is middle to late Holocene in age.
Southeast of Mill Creek the San Bernardino strand adopts a regional trend of about N 55° W, is complex structurally, and is complicated by landslide masses that have been shed from the Mill Creek Formation on Yucaipa Ridge (map sheet 2; Matti and others, 1992). These landslides have numerous crown scarps that resemble scarps created by faults, and it is difficult to associate any of these geomorphic features with the San Bernardino strand. The antiquity of the landslide masses is unknown. Whatever their age, right-slip on the San Bernardino strand has not displaced them laterally to any measurable degree. This reflects either (1) the youthfulness of the landslides or (2) the fact that the San Bernardino strand may step left around and beneath the landslide masses (Matti and others, 1992), in the process leaving most of them effectively attached to the Yucaipa Ridge block. We favor the latter hypothesis.
The left-stepping geometry is more conspicuous to the southeast toward Wilson Creek (Harden and Matti, fig. 2). In plan view, the fault pattern here consists of short northwest-trending fault segments that pass into northeast-trending fault-like scarps. The latter probably are north-dipping reverse faults, but we have not confirmed this speculation; in other fault systems, reverse dip-slip fault segments are common within left-stepping right-lateral strike-slip fault zones, and their occurrence in this vicinity would be compatible with the apparent left-stepping geometry of the San Bernardino strand along this reach. Locally, stream gullies and other geomorphic features have been displaced right-laterally by the northwest-trending fault segments. This evidence for youthful right-slip on the San Bernardino strand southeast of Mill Creek Canyon is complemented by convincing evidence for longer term right-slip throughout the latest Pleistocene and Holocene (Harden and Matti, 1989).
Between Wilson Creek and Banning Canyon the trace of the San Bernardino strand is not well defined, and it may have local left steps. Primary fault features generally are not common between Wilson Creek and Banning Canyon, and geologic and geomorphic evidence for youthful right-lateral activity is not obvious. The segment does not appear to cut or form scarps in youngest Holocene alluvial units, and the fault trace locally is overlapped by unfaulted landslide deposits. The San Bernardino Mountain front along the segment has considerable topographic relief, culminating in a 5,000-foot escarpment near Banning Canyon. This suggests that significant vertical movements have occurred along the segment.
Banning Canyon to Burro Flats segment. Along the Banning Canyon-Burro Flats segment the regional strike of the San Bernardino strand turns abruptly southeastward toward San Gorgonio Pass. Southeast of Banning Canyon the fault zone is marked by springs, bedrock scarps, and lineaments, and Allen (1957), Dibblee (1975a, 1982), and Farley (1979) report that the fault forms a gouge zone in the crystalline rocks. Farther southeast, in an alluviated intermontane area known as Burro Flats, youthful-appearing northwest-trending en-echelon scarps that disrupt Holocene alluvial deposits and Pleistocene landslide debris may have been formed by the modern San Andreas fault. However, the origin of these scarps is questionable: they face northeast and have trends that are similar to other northeast-facing scarps in the region that are not part of the San Andreas fault zone (for example, the family of northeast-facing scarps on the Beaumont Plain), and it is possible that they formed within an extensional strain field rather than by right-slip along the trace of the modern San Andreas fault. If so, then there may be no evidence for recent activity on the San Bernardino strand (modern San Andreas) southeast of Banning Canyon. If they were formed by the San Andreas, these scarps are the southeasternmost evidence for a surface trace of the San Bernardino strand: between Burro Flats and the Banning fault, alluvial deposits are not cut by faults attributable to the San Andreas, and bedrock exposures are not traversed by major fault zones.
Lack of fault features led Allen (1957) to conclude that the modern San Andreas fault (our San Bernardino strand) dies out before reaching San Gorgonio Pass, although Dibblee (1968a, 1975a, 1982) concluded that the fault (his south branch) continues through San Gorgonio Pass and into the Coachella Valley. We do not know if the San Bernardino strand is continuous between Banning Canyon and San Gorgonio Pass or if it has generated the full right-lateral displacement (3 km) that we recognize in the vicinity of Mill Creek (discussed below). The total displacement gradually may fall off between Mill Creek and Burro Flats, so that the fault in the vicinity of San Gorgonio Pass has considerably less displacement and may not have been active recently. We tentatively agree with Allen's conclusion that the modern San Andreas fault dies out southeastward before it reaches San Gorgonio Pass.
The San Bernardino Mountains segment of the San Andreas fault consists of four separate fault strands that evolved sequentially; each strand generated right-lateral displacements during a specific period and then was abandoned and succeeded by a younger strand. We have determined the sequence in which the four strands evolved, but we have not confirmed the timing and amount of displacement for all of them. Even though some elements of their movement history are ambiguous, however, one fact seems clear: together, the four strands record the total history of the San Andreas fault (sensu stricto) since its inception 4 or 5 m.y. ago.
Sequencing relations. The relative sequence in which the four strands evolved can be determined from structural relations between them and by the alluvial units that overlap the strands or are broken by them. The Wilson Creek fault is the oldest strand, followed sequentially by the Mission Creek, Mill Creek, and San Bernardino strands.
The Wilson Creek fault is demonstrably older than the Mission Creek fault because the Mission Creek truncates faults we interpret as part of the Wilson Creek strand and because the trace of the Wilson Creek everywhere is more curving and discontinuous than that of the Mission Creek. These relations suggest that the Wilson Creek fault is an older strand that was deformed and then succeeded by the less sinuous Mission Creek fault. The Mill Creek fault is younger than either the Wilson Creek or Mission Creek faults because it displaces late Quaternary gravel units that are not broken by either of the older faults (discussed below). The San Bernardino strand is the youngest of the four strands, and forms the modern trace of the San Andreas fault in the San Bernardino Mountains segment.
Wilson Creek and Mission Creek faults. Our conclusion that the four strands of the San Bernardino Mountains segment record the full history of the San Andreas fault (sensu stricto) requires that the oldest strand the Wilson Creek fault developed when the San Andreas first evolved. The timing of initial movements on the Wilson Creek fault cannot be determined directly from relations in the San Bernardino Mountains but can be inferred from relations elsewhere in the region.
In the Whitewater area, south of all strands of the San Andreas fault but north of the Banning fault, Miocene and Pliocene sedimentary units all record about the same amount of right-lateral displacement from their probable depositional positions. Relevant units include the Coachella Fanglomerate, which is about 10 m.y. old (Peterson, 1975), and unconformably overlying beds of the Painted Hill and Imperial Formations, older parts of which are 6 m.y. old and older (Table 2). Dillon (1975, p. 334-365, fig. 70) proposed that the San Andreas fault has displaced the Coachella Fanglomerate and Imperial Formation from cross-fault counterparts in the southern Chocolate Mountains, a proposal we accept here. Moreover, conglomeratic beds of the Painted Hill Formation contain clasts that indicate source areas east of the southern Salton Trough. The fact that the Miocene Coachella Fanglomerate and unconformably overlying late Miocene and Pliocene units all record the same displacement requires that the San Andreas fault displaced them together during an episode of right-lateral faulting that commenced after 6 Ma. This timing is compatible with the proposed inception of the San Andreas fault (sensu stricto) in southern California, which most workers link with the onset of seafloor spreading in the Gulf of California and northwestward propagation of the East Pacific Rise (Moore and Buffington, 1968; Larson and others, 1968; Atwater, 1970; Elders and others, 1972; Moore, 1973; Crowell, 1979, 1981). The precise timing of this event is disputed, but about 4 or 5 Ma is a widely cited figure. Thus, initial movements on the Wilson Creek strand of the San Andreas fault occurred after 6 Ma and probably during the earliest Pliocene, 4 or 5 m.y. ago.
Initiation of the San Andreas fault (sensu stricto) also has been linked with termination of movement on the San Gabriel fault. According to Crowell (1982, p. 35, fig. 12), major right-lateral activity on the San Gabriel fault ceased about 5 m.y. ago when the Pliocene Hungry Valley Formation began to accumulate in Ridge Basin. Crowell suggests that right-slip was initiated on the San Andreas at or slightly before that time, although sequencing relations between deposition of the Hungry Valley Formation and initial movements on the San Andreas fault have not been documented.
Eclipse of the San Gabriel fault by the San Andreas fault about 5 m.y. ago corresponds nicely with the maximum age suggested for opening of the Gulf of California (4.9 m.y. ago according to Curray and Moore, 1982). Curray and Moore (1984) proposed a two-phase model for the geologic history of the Gulf: an early phase of diffuse extension, crustal attenuation, and rifting that may have been accompanied by formation of oceanic crust without lineated magnetic anomalies, followed time-transgressively by a later phase of opening accompanied by formation of lineated magnetic anomalies. The early extensional phase commenced 5.5 m.y. ago, but the actual opening phase commenced at about 4.9 Ma and culminated at about 3.2 Ma. Thus, the San Gabriel fault waned, the San Andreas fault (sensu stricto) evolved, and the Gulf of California opened all about 4.9 m.y. ago. Linkage between these events needs to be documented, however, and this scenario has problems. For example, Curray and Moore (1984, p. 29) conclude that 300 km of right-slip on the combined San Andreas and San Gabriel faults corresponds with the 300-km separation of Baja California from mainland Mexico, and that right-slip along the onshore faults commenced 5.5 m.y.B.P. This interpretation is in conflict with the onshore evidence (Crowell, 1981, 1982; Ehlig, 1981; Matti and others, 1986; Matti and Morton, 1992), and illustrates the lack of congruence between onshore and offshore histories.
We have not documented the timing of critical events like the duration of right-slip activity on the Wilson Creek and Mission Creek faults or when the Wilson Creek fault was eclipsed by the Mission Creek fault. These elements of faulting history require information about total displacement, rate of slip, and (or) the age of sedimentary units that date the critical events questions that presently are unanswered by our studies.
The timing of latest strike-slip displacements on the Wilson Creek and Mission Creek faults is clearer than the timing of their early activity. In the vicinity of Whitewater River and Mission Creek, both faults are buried by Pleistocene gravel deposits (Matti and others, 1982a) that bear stage-S2 soils of McFadden (1982) and thus are at least 0.5 m.y. old (McFadden, 1982, p. 55-64, 344-349, 352-354, fig. 16). Farther west, in the Raywood Flat area, the two faults are buried by Pleistocene gravels that appear to correlate with those in the Whitewater-Mission Creek region. Farther northwest, between Banning Canyon and the Cajon Pass region, the Wilson Creek and Mission Creek faults are buried wherever they are associated with Quaternary alluvium.
The absence of primary fault features provides additional evidence for inactivity on the Wilson Creek and Mission Creek faults during late Quaternary time: neither fault displays scarps, sag ponds, shutter or pressure ridges, or right-laterally offset drainage lines. A degraded north-facing scarp that traverses Pleistocene gravels in the Raywood Flat area (Matti and others, 1983) is a possible exception. The scarp is associated with a north-dipping fault noted by Ehlig (1977) and mapped by Farley (1979); the fault breaks the Raywood Flat gravels at their east margin and drops them down to the north. Both Ehlig (1977) and Farley (1979) cite this fault as evidence for youthful right-lateral activity within the Mission Creek-Wilson Creek fault zone. However, we interpret it as a normal dip-slip fault that is part of a family of late Quaternary dip-slip faults in the region that have northeast- or north-facing scarps. The fault coincides with, and has reactivated, the Wilson Creek-Mission Creek fault zone in the vicinity of Raywood Flat, but diverges from it farther west (Matti and others, 1983). Thus, we conclude that the Wilson Creek and Mission Creek strands are abandoned right-lateral faults that have been bypassed by the San Andreas system.
Mill Creek fault. The Mill Creek fault is a late Quaternary strand that is younger than the Wilson Creek and Mission Creek strands but older than the San Bernardino strand. The fault probably has not been a throughgoing right-lateral strand of the San Andreas during Holocene and latest Pleistocene time, and we propose that it has been abandoned and bypassed by the San Andreas system. Locally, the strand has been reactivated by dip-slip movements that have formed north-facing scarps in alluvium and in bedrock.
If our observation is correct that it has displaced Pleistocene gravel deposits and pre-Pleistocene crystalline rocks by about the same amount (discussed below), then the Mill Creek fault generated all of its right-slip displacement during late Pleistocene time. Faulting occurred more recently than about 0.5 Ma, the minimum age for Pleistocene gravel deposits displaced the full amount by the fault: (1) Gravels in the vicinity of Mission Creek and Whitewater River that are displaced 8 to 10 km from possible cross-fault counterparts north of Desert Hot Springs bear stage-S2 soils (McFadden, 1982, p. 55-64) and are at least 0.5 m.y. old; (2) gravels in the vicinity of Raywood Flat that have been displaced about 8 km from Hell For Sure Canyon and the North Fork of the Whitewater River are late Pleistocene, and probably are the same age as those in the vicinity of Mission Creek.
The Mill Creek fault is not a modern neotectonic strand of the San Andreas fault. For most of its extent, the fault is buried by Holocene and latest Pleistocene alluvial deposits. This relation can be observed at several localities between Waterman Canyon and the head of Mill Creek Canyon. Moreover, the strand does not form primary fault features such as sag ponds, shutter and pressure ridges, and right-laterally offset drainage lines. Scarps occur along some segments of the fault: for example, in the Harrison Mountain and San Bernardino North 7.5' quadrangles (Morton and Miller, 1975; Miller, 1979), north- and northeast-facing scarps disrupt bedrock and older Pleistocene alluvial deposits (Allen, 1957, p. 343); the sense of displacement indicated by scarps and faulted contacts is down on the north. Latest displacements on the Mill Creek fault in this vicinity probably represent dip-slip reactivation of the strand, and the north-side-down movements may be related to partial subsidence of the San Bernardino Mountains following their uplift during the Pleistocene (Weldon and Meisling, 1982; Meisling, 1984; Weldon, 1983, 1985b).
North of Raywood Flat the Mill Creek fault zone may have been the site of modern movements, but we have not determined their recency or sense of displacement. At the head of the Middle Fork of Whitewater River (the Middle Fork Jumpoff), alluvial units overlying the crush zone of the Mill Creek fault are broken by a conspicuous fault plane that dips steeply to the north and drops alluvial deposits to the north against crystalline rocks to the south (Allen, 1957, pl. 1; Farley, 1979, p. 86-88, fig. 9). This fault appears to form a degraded north-facing scarp that can be traced a short distance west from the Middle Fork Jumpoff (J.C. Matti and J.W. Harden, unpubl. data). The faulted alluvial deposits are a north- and south-thinning wedge of colluvial sand and gravel derived from highlands to the north and south. Uppermost layers of the north-thinning wedge are youthful and have only incipient soil-profile development. It is not clear if these youngest deposits are faulted, or if they buttress depositionally against the degraded fault scarp.
Several problems plague this locality: (1) The age of the faulted alluvial deposits has not been determined, nor have their stratigraphic and paleogeographic relations to Pleistocene fluvial gravels of the nearby Raywood Flat area that have been displaced 8 km by the Mill Creek fault. (2) It is not clear that the obvious fault plane represents strike-slip movements on the Mill Creek fault most of whose crush zone is buried by the alluvial deposits or activity on one of the other numerous faults that traverse the Raywood Flat region. It is possible that the fault plane, with its north-facing scarp and down-on-the-north separation, represents dip-slip displacements like those that have occurred on the Mill Creek fault farther to the west. Available data do not confirm or refute Holocene right-slip displacements on the Mill Creek fault here or to the east. Farley (1979) explored the possibility that youthful displacements have occurred on the fault east of the Middle Fork Jumpoff, but have died out to the west. Our data cannot rule out this proposal. However, because of the absence of primary fault features throughout most of the alluvial cover here, we suspect that the fault in this region has not generated throughgoing right-lateral displacements since latest Pleistocene time.
San Bernardino strand. The San Bernardino strand is the modern neotectonic component of the San Andreas fault in the vicinity of the south-central Transverse Ranges, and has generated right-lateral displacements throughout Holocene and latest Pleistocene time. If the strand has maintained the 25 mm/year Holocene slip rate determined for the fault in Cajon Pass (Weldon and Sieh, 1985), and if it has generated no more than about 3 km of displacement, then right-slip on the modern neotectonic trace of the San Andreas fault in the Transverse Ranges segment commenced about 120,000 yr. ago.
Although the San Bernardino strand apparently did not generate ground ruptures during the 1857 earthquake (Sieh, 1978a,b), it should be viewed as a fault capable of generating large or even great earthquakes (Allen, 1968, 1981; Sieh, 1981; Rasmussen, 1981; Raleigh and others, 1982; Lindh, 1983; Nishenko and Sykes, 1983; Sykes and Seeber, 1982; Sykes and Nishenko, 1984; Ziony and Yerkes, 1985; Wesson and Wallace, 1985; National Earthquake Prediction Evaluation Council, 1988). Most earthquake scenarios for this part of southern California incorporate ground-rupture lengths of several hundred kilometers and moment magnitude of 8 or greater (Raleigh and others, 1982; Sykes and Nishenko, 1984; Lindh, 1983). Some workers suggest that an earthquake comparable to the 1857 event might lead to ground rupture on the San Bernardino strand northwest through Cajon Pass and onto the Mojave Desert segment, and possibly southeast through San Gorgonio Pass and into the Coachella Valley region where other neotectonic components of the San Andreas fault would be activated. Such scenarios need to be tested in a rigorous way, however, and other scenarios may be possible (Rasmussen, 1981; Weldon and Sieh, 1985, p. 811-812). Ground-rupture patterns within the San Andreas fault zone will become more predictable only when we understand the overall fabric of a region where the San Andreas fault is only one of several neotectonic elements.
Amount of displacement: previous interpretations. Following Crowell's (1962) proposal that the San Andreas fault in southern California has 210 km of right-lateral displacement, many workers have attempted to apportion this displacement among various San Andreas strands in the southeastern San Bernardino Mountains. Gibson (1964, 1971) inferred on the basis of paleocurrent and clast-provenance studies that the Mill Creek fault has displaced the Miocene Mill Creek Formation (of Gibson, 1971) about 120 km from its original position adjacent to the Orocopia Mountains. Dibblee (1968a, p. 269) concluded that, if Crowell's 210 km of right slip has occurred along strands of the San Andreas in the San Bernardino Mountains, then the largest movement probably occurred along his north branch (Mill Creek fault). Later, Dibblee (1975a, p. 134) proposed that the north branch generated about 96 km of right slip and displaced crystalline rocks in the southeastern San Bernardino Mountains from presumed cross-fault counterparts in the Orocopia Mountains. Dibblee (1982b, p. 164) subsequently increased this value to 120 km a displacement identical to Gibson's (1964, 1971) and presumably based on Gibson's palinspastic restoration of the Mill Creek strata to the Orocopia Mountains region.
Neither Gibson (1964, 1971) nor Dibblee (1968a, 1975a) accounted for the large difference between their proposed displacements on the Mill Creek fault (96 to 120 km) and Crowell's (1962) proposal for total displacement on the San Andreas (210 km). This difference presumably was made up by other faults in the region. However, Dibblee (1968a, p. 168) had ruled out his south branch of the San Andreas because the fault has only a few km of right slip in the San Gorgonio Pass area, leaving only the Banning and San Jacinto faults to take up the missing 90 km. However, neither of these faults qualifies for the following reasons: (1) the Banning did not generate right-lateral displacements at the same time as the San Andreas (sensu stricto) and therefore does not figure into the 210-km reconstruction proposed by Crowell (1962); (2) the San Jacinto has no more than 25 or 30 km of displacement (Sharp, 1967); and (3) neither fault figures into palinspastic reconstruction of rocks in the southeastern San Bernardino Mountains because both faults pass outboard of that region.
Dillon's (1975) proposal that the north branch of the San Andreas has displaced rocks in the southeastern San Bernardino Mountains from counterparts in the southern Chocolate Mountains (discussed above) provided a major alternative to the proposals by Gibson and Dibblee. Dillon's displacement on Dibblee's north branch (180±20 km) is considerably greater than the displacements proposed by Gibson and Dibblee (96 to 120 km), but both estimates seemed equally attractive based on the merits of their underlying cross-fault correlations.
This conflict was resolved by later studies in the San Bernardino Mountains that reinterpreted the bounding faults between distinctive basement terranes. Ehlig (1977), Farley (1979), and Matti and others (1983a, 1985) showed that the Mill Creek fault (Dibblee's north branch) does not form a major break between crystalline rocks as Gibson (1971) and Dibblee (1968a, 1975a, 1982a) believed. Instead, lithologic similarities between crystalline rocks on either side of the fault preclude large right-lateral displacements on this strand of the San Andreas (Matti and others, 1985, p. 9). Thus, the Mill Creek fault could not have displaced Gibson's Mill Creek strata by more than a few kilometers from their depositional position. Building on Ehlig's (1977) earlier work, Farley (1979) demonstrated that the major lithologic break between crystalline rocks is formed by the Mission Creek fault of Allen (1957) an interpretation refined by Matti and others (1983a, 1985). This fault not the Mill Creek strand (Dibblee's north branch) was shown to be the San Andreas strand that juxtaposed rocks of San Gabriel Mountains-type against rocks of San Bernardino Mountains-type as proposed by Dillon (1975; see Farley, 1979, p. 120-129; Matti and others, 1985, p. 9-10). This clarified the distribution of major basement terranes and their bounding faults and set limits on displacements for individual strands, but left two factors still unresolved: (1) the palinspastic position of the Mill Creek Formation alleged to have originated near the Orocopia Mountains and (2) the 40-km to 80-km discrepancy between Dillon's estimate of 180±20 km of displacement on the San Andreas and the widely cited estimate of 240 km (upgraded by Crowell, 1975a, 1981, and Ehlig, 1981, from Crowell's original estimate of 210 km).
The slip discrepancy was reemphasized when Matti and others (1985, 1986) proposed that total right slip on the San Andreas fault in the vicinity of the San Bernardino Mountains may be no more than 160±10 km. This estimate corresponds nicely with the lower limit of Dillon's estimate of 180±20 km, and both proposals indicate that estimates of 210 to 240 km for total right slip on the San Andreas fault (sensu stricto) in southern California may be too large. We adopt this position, and use the 160-km displacement together with estimates for displacement on other strands of the San Andreas zone to reconstruct fault-movement histories for individual strands of the San Andreas in the vicinity of the San Bernardino Mountains. Return to Contents, San Bernardino Mountains segment, San Andreas Fault
Amount of displacement: new possibilities. Our conclusion that the Wilson Creek, Mission Creek, Mill Creek, and San Bernardino strands record the full history of the San Andreas fault (sensu stricto) requires that their individual displacements sum to the total displacement on the fault in southern California. Estimates for this total range from less than 100 km to 270 km, but the most widely cited studies have focused on a range of 210 km (Crowell, 1962) to 240 km (Crowell, 1975a; Ehlig and others, 1975; Ehlig, 1981, 1982). By contrast, Matti and others (1985, 1986; Frizzell and others, 1986) concluded that the San Andreas fault in southern California probably has no more than 160±10 km of total displacement. This smaller displacement is based on the proposal that distinctive bodies of Triassic megaporphyritic monzogranite that occur in the Mill Creek region of the San Bernardino Mountains (Monzodiorite of Manzanita Springs of Morton and others, 1980; unit Trm of Matti and others, 1992) and in the Liebre Mountain area on the opposite side of all strands of the San Andreas fault represent segments of a formerly continuous pluton that was severed and displaced by the fault. We adopt the 160-km displacement, and apportion it between the four strands of the San Bernardino Mountains segment of the San Andreas fault.
San Bernardino strand. We propose that the San Bernardino strand has no more than 3 km of displacement. This hypothesis is based on evidence from Pleistocene alluvial deposits and crystalline bedrock that occur south of the fault between Mill Creek and the Santa Ana River. The gravels were deposited by ancestral streamflows of Mill Creek, and contain several distinctive clast types traceable to bedrock sources drained by the modern stream Since their deposition, the gravels have been displaced no more than 3 km by the San Bernardino strand (Matti and others, 1992). The gravel deposits are capped by soil profiles that have thick, well developed, red argillic horizons comparable to those in old Pleistocene soils (stage S2 soils of McFadden, 1982). The bedrock is a slice of crystalline rock from the Wilson Creek block; although we cannot identify an exact cross-fault counterpart for this slice, it constrains right-lateral displacements on the San Bernardino strand to no more than 3 km.
Mill Creek strand. Available evidence suggests that the Mill Creek fault has no more than 8 to 10 km of right-lateral displacement. This conclusion is based on four separate geologic relations:
Mission Creek and Wilson Creek strands. Displacements on the Mission Creek and Wilson Creek faults presently cannot be determined directly on the basis of specific cross-fault correlations. Originally, we proposed that the two faults are major strands of the San Andreas that in combination have brought exotic rocks against the San Bernardino Mountains from original positions as much as 150 km farther southeast in the Coachella Valley region (Matti and others, 1985, p. 9-10). We proposed that the Wilson Creek strand had considerably more displacement than the Mission Creek strand, but we acknowledged the uncertainties in this model and pointed out that an equally likely reconstruction of displaced crustal blocks could be achieved if the Mission Creek fault, rather than the Wilson Creek fault, had the larger displacement. We now favor this role reversal between the two fault strands.
We assign right-slip estimates to the Mission Creek and Wilson Creek faults based on three arguments: (1) the hypothesis that these two faults, along with other strands of the San Andreas system, have displaced the Mill Creek and Liebre Mountain Triassic megaporphyry bodies by no more than 160 km; (2) cross-fault correlation between rocks of San Gabriel Mountains-type in the southeastern San Bernardino Mountains and similar rocks in the Chocolate Mountains (Dillon, 1975); and (3) correlation of fault segments in the San Bernardino Mountains with other fault strands of the San Andreas zone in the San Gabriel Mountains.
As discussed by Matti and Morton (in press), total displacement on the San Andreas fault (sensu stricto) in the vicinity of the San Bernardino Mountains is the sum of displacements on individual faults that jointly have displaced the Liebre Mountain megaporphyry body away from its cross-fault counterpart in the Mill Creek region:
A + B + C + D + E = 160 km,
where A = San Bernardino strand, B = Mill Creek strand, C = San Jacinto fault, D = Mission Creek strand, and E = Wilson Creek strand.
The slip equation can be partly solved using the following values:
Using slip data for A, B and C, the displacement equation is complete except for D and E, displacements on the Mission Creek and Wilson Creek faults:
3 km + 8 km + 20 km + D + E = 160 km.
Assuming values A, B, and C are correct and sum to 31 km, then values D and E (the combined displacements on the Mission Creek and Wilson Creek faults) sum to 129 km. Compare this value with the 150-km displacement we originally proposed for combined slip on the Mission Creek and Wilson Creek faults when we did not take into account slip contributed by the San Jacinto fault.
To arrive at values for D and E in the displacement equation, we turn to strike-slip faults in the San Gabriel Mountains that might be counterparts of either the Mission Creek fault or the Wilson Creek fault. Given our conclusion that the Mission Creek strand is middle Pliocene and younger in age and is a major fault that juxtaposed rocks of San Gabriel Mountains-type and San Bernardino Mountains-type, its counterpart in the San Gabriel Mountains should include faults of the same age that also juxtapose the two crystalline suites. By these criteria the Mission Creek strand is correlative with the Mojave Desert strand, which must be viewed as a major strand of the San Andreas that has a right-slip history extending from the present back to at least middle Pleistocene (Barrows and others, 1985, 1987). Although Barrows and others (1987, p. 2, 83, 85) indicated that the Mojave Desert strand (their "main trace") may be no older than 1 to 1.4 Ma, they (1987, p. 82) also pointed out that "... all rock units older than late Pleistocene that are juxtaposed along the main trace are dissimilar...." We take this to mean that the Mojave Desert strand is the major trace of the San Andreas in the vicinity of the San Gabriel Mountains, and is responsible for juxtaposing rocks of San Gabriel Mountains-type against rocks of San Bernardino Mountains-type. Like other workers, we suspect that the Mojave Desert strand has a longer-lived history than proposed by Barrows and others (1985, 1987). Thus, we correlate the Mission Creek strand (and the younger Mill Creek and San Bernardino strands) with the Mojave Desert strand (and potentially with the Pliocene Nadeau faults).
If the Mission Creek and Mojave Desert strands are the major throughgoing traces of the San Andreas, then restoration of significant slip on that strand together with restored slip on the Mill Creek and San Bernardino strands would bring the northwest termination of the Punchbowl fault near Palmdale (fig. 1) closer to the southeast termination of the Wilson Creek fault in the southeastern San Bernardino Mountains. We propose that the Wilson Creek and Punchbowl faults originally formed a single throughgoing structure that was severed by younger strands of the San Andreas fault; the displaced segments of the once-continuous structure now are situated in the San Bernardino and San Gabriel Mountains. Comparison of the Wilson Creek and Punchbowl faults is supported by three lines of evidence: (1) they generated right-lateral displacements during the same time period (5 or 6 Ma to about 3.5 Ma, although Barrows, 1987, would have activity on the Punchbowl fault be as young as 2 Ma); (2) they both have sinuous traces and west-dipping reverse dips; and (3) they both bound crystalline-rock slices (the Wilson Creek block and crystalline rocks of Pinyon Ridge in the San Gabriel Mountains) that have no documented affinity with particular crystalline terranes in southern California but are broadly similar to each other. We propose that the Punchbowl and Wilson Creek strands of the San Andreas once formed a continuous throughgoing right-lateral fault that had about 40 to 45 km of displacement based on estimates for the Punchbowl fault in the San Gabriel Mountains. In this report, we use the 40-km figure and insert this value for E in the displacement equation.
Using slip data for A, B, C, and E, the displacement equation
3 km + 8 km + 20 km + D + 40 km = 160 km
can be solved for D, yielding a displacement of 89 km for the Mission Creek strand of the San Andreas fault. Return to Contents, San Bernardino Mountains segment, San Andreas Fault
Available data lead to a preliminary history of the San Andreas fault in the vicinity of the Transverse Ranges (fig. 4):
(1) Following inception of the San Andreas fault 4 or 5 m.y. ago, about 40 km of right-slip on the throughgoing Wilson Creek-Punchbowl fault juxtaposed exotic crystalline and sedimentary rocks of the Wilson Creek block against the San Bernardino Mountains (fig. 4A). Rocks native to the San Bernardino Mountains, including Triassic megaporphyritic monzogranite discussed by Matti and others (1986; Frizzell and others, 1986), were displaced to the northwest by these offsets, and the Wilson Creek block and Mill Creek Formation were juxtaposed against rocks native to the San Bernardino Mountains. During this period, the Wilson Creek-Punchbowl fault probably was a straight throughgoing strand that was continuous to the northwest and the southeast with the Mojave Desert and Coachella Valley segments of the San Andreas fault.
(2) Right-slip on the Wilson Creek-Punchbowl fault terminated when it was deformed into a sinuous trace and was bypassed by the Mission Creek fault (fig. 4B). This event occurred adjacent to the left-lateral Pinto Mountain fault, and left the Wilson Creek block stranded against the San Bernardino Mountains block. Thereafter, the San Bernardino and Wilson Creek blocks functioned as a single structural unit, with the Wilson Creek block behaving as though it was native to the region.
(3) The Mission Creek fault juxtaposed Pelona Schist and associated upper-plate rocks like those in the eastern San Gabriel Mountains against the Wilson Creek block (fig. 4C). The throughgoing Mission Creek fault was continuous to the northwest and southeast with the Mojave Desert and Coachella Valley segments of the San Andreas. Ultimately, the Mission Creek fault and adjacent rocks were deflected to the west and southwest as a left step developed in the San Andreas fault (fig. 4D), and the strand was abandoned and bypassed by the San Andreas transform system.
(4) The Mill Creek fault evolved inboard of the locked up Mission Creek fault (fig. 4E), and generated about 8 km of right-lateral displacement during the late Pleistocene (after about 0.5 Ma). Subsequently, the Mill Creek fault was deformed and apparently abandoned as a throughgoing right-lateral fault.
(5) The San Bernardino strand marks the trace of the modern neotectonic strand of the San Andreas fault within the Transverse Ranges segment (fig. 4F). This strand probably has reactivated the older Mission Creek strand in the vicinity of the San Bernardino valley.
Several questions can be asked about the history of the San Andreas fault in the vicinity of the south-central Transverse Ranges. Why did several strands evolve? By what mechanism was each strand deformed and succeeded by a younger strand, and how much time did such a transition involve? Were there gaps in right-slip activity on the San Andreas fault during these periods? If so, did right-slip transfer from the San Andreas to some other fault in the San Andreas transform system? What impact did deformation of the Wilson Creek and Mission Creek faults have on the region? Or alternatively, were these two faults deformed in passive response to other events in the region? Does the Pliocene and early Pleistocene history of the San Andreas fault provide a precedent for the modern neotectonic framework? Some of these questions can be addressed on the basis of existing data.
Several unique Pliocene-Pleistocene structural and physiographic elements coincide in the vicinity of the south-central Transverse Ranges a juxtaposition that implies a cause-and-effect relation. The most important elements are (1) the multiple deformed strands of the San Andreas fault; (2) the Pinto Mountain fault, a major left-lateral structure that generated about 16 km of displacement (Dibblee, 1968b); and (3) the San Bernardino Mountains, a major physiographic element created by Pleistocene orogenesis (Meisling, 1984; Meisling and Weldon, 1989). Various authors have discussed the history of the San Andreas fault in the context of these unique structural and physiographic elements. Allen (1957) contributed many original concepts, including (1) the idea that the present surface geometry of faults in the San Gorgonio Pass region differs from their original orientation parallel to the northwest regional strike of the San Andreas fault, and (2) the idea that faults of the San Andreas system are vying with east-oriented faults of the Transverse Ranges for structural control of the region.
Allen (1957, p. 344-346, fig. 3) evaluated the structural setting of the San Gorgonio Pass region from the viewpoint of a regional triaxial strain field incorporating a principal stress direction oriented N-S and an intermediate stress direction that alternates between the E-W and vertical axes. Allen proposed that right-slip on San Andreas-type faults predominated when the intermediate stress was oriented vertically, but thrust faulting (and presumably left-slip on structures like the Pinto Mountain fault) predominated when the intermediate stress was oriented E-W. Allen's analysis applied mainly to Quaternary thrust faulting in San Gorgonio Pass, but by implication extends to the paleotectonic history of the Banning and San Andreas faults. The triaxial-strain concept was embellished by Farley (1979, p. 115-120) and by Crowell and Ramirez (1979, p. 31-32, 39), who evaluated relations between the San Andreas fault and the Pinto Mountain fault from the viewpoint of conjugate shear within a region undergoing simple shear.
Dibblee (1975a) evaluated the history of the south-central Transverse Ranges by relating deformation of the San Andreas fault to left-lateral displacements on the Pinto Mountain fault, and by relating these events in turn to uplift of the San Bernardino Mountains. In his analysis, left-lateral displacements on the Pinto Mountain fault impinged on the San Andreas and created a bottleneck that impeded throughgoing right slip; as a consequence, the Banning fault and segments of his north and south branches in the San Gorgonio Pass region were bent into east-west orientations, and the San Bernardino Mountains evolved through compressional uplift (Dibblee, 1975a, p. 134-135; 1982b). He suggested that these events occurred in late Quaternary time. Farley (1979, p. 115-129) envisioned a similar scenario, although he interpreted strands of the San Andreas fault differently than Dibblee and suggested that deformation of the San Andreas and Banning faults by left-slip on the Pinto Mountain fault occurred during late Pliocene time.
We agree that structural and physiographic elements in the south-central Transverse Ranges are linked in their evolution, but we differ from earlier workers in our view of how the various elements interacted with each other. For example, Dibblee (1975a, p. 134-135) and Farley (1979, p. 115-129) both suggest that left-slip on the Pinto Mountain fault deflected the Banning fault and all strands of the San Andreas fault during a single period of deformation. By contrast, we believe that the Banning fault acquired its east-trending orientation prior to inception of the San Andreas fault 5 m.y. ago, and that the Wilson Creek, Mission Creek, and Mill Creek faults each were deformed separately during compressional episodes that were followed by right-slip on the succeeding strand.
It seems certain that deformation of the San Andreas fault in the south-central Transverse Ranges is linked with orogenic uplift of the San Bernardino Mountains, but linkage between the two events cannot be documented in detail because their timing is known only in a general way. Recent studies suggest that uplift of the mountains commenced less than about 2.6 m.y. ago (Sadler, 1982a,b; May and Repenning, 1982; Weldon and Meisling, 1982; Meisling and Weldon, 1982, 1989; Sadler and Reeder, 1983; Meisling, 1984). We can demonstrate that the Mission Creek fault was being deformed during the Pleistocene and was abandoned before 0.5 m.y. ago, thus linking uplift and strand deformation during later parts of the Quaternary.
The Pinto Mountain fault probably played a role in the interaction between the San Andreas fault and the San Bernardino Mountains although more likely as an effect than as a cause. The history of the Mission Creek fault may have bearing on the history of the Pinto Mountain fault. In the southeastern San Bernardino Mountains, the Mission Creek fault has a concave-south trace that occurs outboard of crystalline rocks native to the region. This curving geometry apparently was achieved as a bulge of San Bernardino Mountains basement was projected west and southwest across the path of the fault during the late stages of its activity an event that accompanied deformation of the strand into its curving trace (fig. 4B,C, fig. 5). Left-slip on the Pinto Mountain fault could have achieved this effect by displacing the southeastern corner of the San Bernardino Mountains west and southwest relative to the little San Bernardino Mountains. The Pinto Mountain fault need not have truncated and offset the Mission Creek fault by this process: instead, gradual projection of the San Bernardino Mountains across the path of the San Andreas merely could have deflected or bowed the Mission Creek strand into a curved trace (fig. 5). When the curvature became too great to sustain right-slip, the Mission Creek was succeeded by other faults in the San Andreas family, including the San Jacinto fault zone and the Mill Creek strand (fig. 4C).
It is unclear whether left-slip on the Pinto Mountain fault continued throughout the history of the Mission Creek fault, thereby causing slow progressive deformation of the strand, or whether the Mission Creek fault enjoyed an early history of right-slip unimpeded by regional compression. Ultimately, however, continued left-slip on the Pinto Mountain fault projected the San Bernardino Mountains farther across the path of the San Andreas fault, and the Mission Creek strand was abandoned and bypassed by the San Andreas transform system before 0.5 Ma (fig. 4D). The evolving intersection between the Mission Creek and Pinto Mountain faults probably resembled the curving intersection between the modern San Andreas fault and the Garlock fault in the Big Bend of the Tejon Pass region. Westward projection of the San Bernardino block created a significant geometric effect (fig. 5): the Mojave Desert and Coachella Valley segments of the San Andreas fault have been stepped left from each other by about 15 km, which is about the amount of left-lateral displacement on the Pinto Mountain fault (Dibblee, 1968b).
The Mill Creek fault did not exist when the Mission Creek fault was deformed, but subsequently broke in behind the barrier imposed by the bent and locked-up Mission Creek strand. This event occurred after about 0.5 Ma. Subsequent displacements on the Mill Creek fault sliced off the westward projection of the San Bernardino Mountains block created by left-lateral displacements on the Pinto Mountain fault. Ultimately, renewed left-slip on the Pinto Mountain fault kinked the Mill Creek fault in the vicinity of the Whitewater River Forks, and the Mill Creek fault, too, has been abandoned by the San Andreas system.
The history of the San Bernardino Mountains segment of the San Andreas fault may have included one or more slip gaps filled by other faults in southern California. For example, some authors have suggested that the San Jacinto fault evolved when right-slip on the San Andreas was impeded in the vicinity of the southeastern San Bernardino Mountains (for example, Crowell, 1981, p. 597), and that accelerated or decelerated slip on these two faults may have alternated through time (Sharp, 1981, p. 1761). The long-term slip rate for the San Bernardino Mountains segment of the San Andreas fault may well have varied through time, but such a scenario can be documented only when the timing and amount of right-lateral displacements on the Mission Creek and Mill Creek strands of the segment are better understood.