CENTRAL CREEPING SECTION [c5, p125-126]

Densely aligned epicenters mark the nearly straight creeping section of the San Andreas fault that separates the south end of the 1906 break near San Juan Bautista from the north end of the 1857 break near Cholame ( Figure 5.6B). Densely aligned epicenters follow the Calaveras fault northward to a point just east of the south tip of the San Francisco Bay, where the Hayward fault branches to the west and the Greenville fault zone branches to the east. Few epicenters fall along the northward extension of the Calaveras fault beyond this branching point although a diffuse cluster of epicenters coincides with the right-stepping offset between the north end of the Calaveras fault and the Concord fault ( Figure 5.6A). This dilatational offset was the site of pronounced earthquake swarms in June 1970 and August 1976 (Lee and others, 1971; Weaver and Hill, 1978/79).

Although these dense alignments of epicenters coincide closely with the mapped surface trace of the San Andreas fault system as first documented by Eaton and others (1970a), the coincidence is not everywhere perfect ( Figure 5.6B). In the region where the Calaveras fault branches from the main trace of the San Andreas fault for example, the densely aligned epicenters appear to be systematically displaced 3 to 4 km westward of the surface trace of the San Andreas fault and a somewhat smaller distance eastward of the surface trace of the Calaveras fault. Much of this apparent offset results from a strong contrast in rock type and P-wave velocity across the faults that is not adequately accounted for in routine hypocenter locations (Mayer-Rosa, 1973; Pavoni, 1973 Spieth, 1981). When the hypocenter locations are recalculated using a more appropriate, two-dimensional structural model, these offsets are much reduced but not completely eliminated. The remaining offsets reflect deviation of the faults from the vertical, with the San Andreas fault dipping 700 W. (Pavoni, 1973; Spieth, 1981) and the Calaveras fault dipping 800 E. (Reasenberg and Ellsworth, 1982).

Wesson and others (1973) recognized the close correlation between active creep and persistent microseismic activity along the 200-km-long section of the San Andreas fault north of Parkfield and proposed that this correlation may hold for other branches of the fault in central California as well. Allen (1981) and Schultz and others (1982) pointed out that this correlation holds for the Calaveras-Hayward fault system, but creep measurements have yet to be made on the subparallel Rodgers Creek-Healdsburg-Maacama and Green Valley-Bartlett a Springs faults. Creep is the dominant process for shallow slip along the central section of the San Andreas fault and geodetic measurements spanning this section of the fault indicate that the long-term slip rate of 32 mm/yr along the fault accommodates nearly all of the local plate motion. Because it appears that little, if any, shear strain 15 accumulating in the blocks on either side of the fault, most seismologists believe that this section of the fault is unlikely to rupture in a great earthquake (see chap. 7).

The creeping section of the fault system, however, has produced several moderate earthquakes during historical time (see chap. 6). The most recent of these events, which occurred along the right-branching segments northeast of the San Andreas fault, where creep rates average several millimeters per year, include (1) the M=5.9 Coyote Lake earthquake of August 6, 1979, and the M=6.2 Morgan Hill earthquake of April 24, 1984, which ruptured adjacent 20-km-long segments of the Calaveras fault south of its junction with the Hayward fault (Reasenberg and Ellsworth, 1982; Bakun and others, 1984); and (2) the M=5.5 and 5.8 Livermore events of January 24 and 27, 1980, which ruptured a 20-km-long stretch of the Greenville fault north of Livermore (Bolt and others, 1981).

The most noteworthy sequence of moderate earthquakes along the central section of the San Andreas fault involve the five virtually identical M= 6 events that have ruptured the same 30-km-long stretch near Parkfield at nearly 22-year intervals since 1881 (Bakun and McEvilly, 1984). This stretch of the fault is defined by a 1-km right-stepping offset on the south and a 5° W bend on the north; it coincides with the transition between the south end of the creeping section of the fault and the north end of the 1857 break (see Bakun and Lindh, 1985, Figure 1). The most recent of these characteristic Parkfleld earthquakes occurred in 1966, and an intensive monitoring experiment is underway to capture a detailed instrumental record of the next Parkfield earthquake, which is predicted to occur sometime within a 10-year window centered on 1987-88 (Bakun and Lindh, 1985).

The scattered seismicity within the Coast Ranges surrounding the San Andreas fault system is distinctly more intense in the Franciscan terrane east of the fault than in the granitic Salinian block to the west. The large, dense cluster of epicenters along the eastern margin of the Coast Ranges adjacent to the south end of the creeping section represents the aftershocks of the M=6.7 Coalinga earthquake of May 2, 1983, and the M=5.7 Kettleman Hills earthquake of August 4, 1985, both of which involved reverse slip on northwest-striking planes subparallel to the adjacent section of the San Andreas fault (Stein and King, 1984; Eaton, 1989). Scattered clusters of epicenters within the Franciscan terrane show crude northwest-trending alignment with the southwest edge of the Coalinga-Kettleman Hills aftershock zone and the Ortigalita fault, the north end of which passes beneath the San Luis Reservoir (LaForge and Lee, 1982). This weakly defined lineation is essentially colinear with the Greenville fault, east of the San Francisco Bay (Figure 5.4).

The Salinian block west of the central section of the San Andreas fault and east of the Sur-Nacimiento fault forms a broad, nearly aseismic swath along the west flank of the Coast Ranges. These two seismically active fault zones separate the granitic Salinian block from the Franciscan terrane on either side (see chap. 3). The Rinconada fault, within the Salinian block, appears to be nearly aseismic except, possibly, toward the south where it approaches the Sur-Nacimiento fault zone. The small cluster of epicenters just east of the midpoint of the Rinconada fault represents a persistent spot of microearthquake activity at depths of 8 to 10 km near the San Ardo oil field (Poley, 1988).

In the cross section along the central section of the San Andreas fault (L-L', Figure 5.7), the actively creeping section of the fault shows up as a densely mottled distribution of hypocenters within the upper 12 to 15 km of the crust. The density of hypocenters within this creeping section tends to decrease with depth, and the denser clusters generally are concentrated at depths of less than 5 to 8 km. The base of the seismogenic zone undulates about an average depth of some 13 km beneath most of the creeping section, but it deepens to 15 km beneath both the northern and southern transitions to the locked sections of the fault. In contrast to the creeping section of the fault, the sparse seismicity associated with the locked segments that ruptured in 1906 (north) and 1857 (south) tends to be concentrated toward the deeper parts of the seismogenic crust. Note that the 1989 Loma Prieta M=7.1 earthquake ruptured the 45-km-long section of the San Andreas fault with a pronounced U-shaped gap in shallow earthquakes immediately north of the creeping section of the fault (cross sec. L-L', Figure 5.7; see chap. 6).

The cross sections transverse to the central San Andreas fault system (G-G', Figure 5.8A; I-I', J-J', K-K', Figure 5.8B) reveal the seismically active, creeping branches of the fault as narrow, near-vertical hypocenter distributions coincident with the fault plane. The broadened distribution in cross sections I-I' and J-J' (Figure 5.8B) reflects the oblique projection of earthquakes along the Calaveras fault zone and the clustering northeast of the fault in the Bear Valley region (Ellsworth, 1975), respectively. These transverse cross sections also emphasize the quiescence of the Salinian block relative to the Franciscan terrane on either side (note, however, the isolated cluster of deep events beneath San Ardo in the Salinian block in cross sec. K-K'), and the fairly uniform depth of 12 to 15 km to the base of the seismogenic zone that persists throughout the central Coast Ranges. As in the area farther north, however, maximum focal depths increase rather abruptly to 25 km beneath the eastern margin of the Coast Ranges and the Great Valley. This Increase in focal depth is particularly pronounced beneath the dense cluster of hypocenters associated with the 1983 Coalinga earthquake and its aftershocks (cross sec. G-G', Figure 5.8B).