DEPARTMENT OF THE INTERIOR TO ACCOMPANY MAP MF -307 UNITED STATES GEOLOGICAL SURVEY ACTIVE FAULTS AND PRELIMINARY EARTHQUAKE EPICENTERS (1969-1970) IN THE SOUTHERN PART OF THE SAN FRANCISCO BAY REGION By R. D. Brown, Jr. and W. H. K. Lee INTRODUCTION miles. This narrow map pattern indicates that earth- quake activity is localized along zones that are Earthquakes are both abundant and frequent in the vertical or that dip very steeply, for although many San Francisco Bay region. Most of them are so of the earthquakes represented by the plotted epicent- small that they are detected only by sensitive seis- ers are miles below the surface, all of them in any one mographs. Only rarely is one large enough to be belt fall within a mile or so horizontally of a cen- felt, and still more rarely is one large enough to trally located vertical plane. cause noteworthy damage. Great earthquakes like The three zones of earthquake activity approximately that which struck northern California in 1906 are coincide with four known fault zones. These are, from so infrequent that only 2 or 3 are known from southwest to northeast, the San Andreas, the Sargent, California in the last 150 years--the time span for and the Calaveras and Hayward faults. The San which we have reliable historic records. Andreas, the Calaveras, and the Hayward faults were The abundance of small, instrumentally detectable known to be active even before the 1906 San Francisco earthquakes permits a wide range of seismologic and earthquake. Each of these has generated major earth- geologic studies to proceed without waiting for quakes with measurable amounts of surface fault dis- more damaging earthquakes. This opportunity is be- placement in historic time. Locally, there is move- ing exploited in the San Francisco Bay region by ment, or creep, along each fault within the area of several cooperating scientific organizations that this map. Although average creep rates are only have installed and maintain sensitive seismometers fractions of an inch a year, they are great enough to to record small earthquakes. The data from this adversely affect some manmade structures. Those seg- network of seismograph stations are the basis for ments of the faults that exhibit tectonic creep are many research investigations, most of which have a chiefly within the zones of greatest small-earthquake direct bearing on earthquake risks in the San activity. Francisco Bay region. Not shown on the map, but determinable from the same seismic data as that on which the map is based, EPICENTER DISTRIBUTION AND ITS are several other earthquake characteristics, among RELATION TO ACTIVE FAULTS them depth to the hypocenter and the nature of the movement that produced the earthquake. These charac- This map is one product of the continuing teristics are known for a representative sample of the research on small earthquakes in the San Francisco plotted points. They show that most of the earth- Bay region. It shows both the preliminary epicenter quakes are no deeper than 10 miles below the surface location and the magnitude of earthquakes greater of the earth, and none are deeper than 15 miles. The than 0.5 magnitude on the Richter scale for the fault movements producing most of the earthquakes are period January 1, 1969 to December 31, 1970. A few right lateral and in the same sense as the 1906 move- of the symbols on the map may indicate two or more ment on the San Andreas fault--that is, rock masses epicenters located at exactly the same point. Also, seaward of such northwest-trending faults as the San because of technical problems that are discussed Andreas, Hayward, and Calaveras move relatively to- later, a large number of epicenters near Danville ward the northwest and move almost purely in the are not shown. Their inclusion would not signifi- horizontal plane. Locally, as along the Sargent fault cantly alter the patterns shown on the map nor the and at a few other points where faults branch or conclusions drawn from that pattern. splay, the fault movements may be more complex and The map also shows the location of faults that show different orientations and substantial components are historically active or that show evidence of of vertical movement. Nevertheless, the dominant pat- geologically young surface displacement (Brown, tern of movement recorded from earthquakes agrees with 1970). These faults can be recognized by such that observed in active fault displacements and from linear but discontinuous topographic anomalies as displaced rock units in the San Francisco Bay region. narrow elongate depressions, sag ponds, offset San Andreas fault.--The most southwesterly of the streams, and scarps. Most of the recent surface three zones of earthquake activity extends from a fault breaks shown on the map have been verified by point about 5 miles north of Watsonville to the south- field investigations, but a few, shown as dashed on east corner of the map. Data from beyond the map area the map, are based solely or largely on photo- (Lee and others, 1970) show that this zone of small- geologic interpretation of aerial photographs. earthquake activity extends for many tens of miles to Earthquakes occur throughout the entire San the southeast and that, for all this distance, it fol- Francisco Bay region, but as the map shows they are lows or closely parallels the surface trace of the not randomly distributed. Most of them are within San Andreas fault. three northwest-trending, nonparallel belts. Each The belt of epicenters shown on this map does not of the plotted epicenters is, by definition, that exactly coincide with the line of mapped surface fault point on the earth's surface that is vertically breaks, but lies generally 2 or 3 miles southwest of above the source of an earthquake (fig. 1), so the it. The map location of the surface fault breaks is belts of epicenters shown on the map actually' known to be accurate and is the locus of past fault represent zones of earthquake activity that extend movements as well as of modern tectonic creep. Al- deep within the earth's crust. though the epicenters are not so precisely located, Where the epicenters are numerous the belts are they are thought to be accurate to within a mile of narrow and clearly evident, even though the belt their true position. If so, they indicate that the boundaries are somewhat irregular or indefinite. fault surface at depth lies to the southwest of its The belts are about 2 miles wide or less in most surface trace. This condition could result from a places, and nowhere are they wider than about 4 steep southwesterly dip on the surface fault, from a Figure 1. BLOCK DIAGRAMS SHOWING THE RELATIONSHIP BETWEEN: F, earthquake focus or hypocenter: that point, generally at some depth within the earth's crust, from which the seismic energy appears to radiate; E, earthquake epicenter: a point on the earth's surface that is vertically above the source of the earthquake; f, fault surface (shaded): a fracture surface along which failure and accompanying dislocation have displaced adjacent blocks of the earth's crust. In both diagrams arrows indicate a right-lateral strike-slip sense of movement, as on the San Andreas fault. 2 series of discontinuous vertical fault surfaces is added corroboration that the difference in abun- stepped deeper towards the southwest, or from a more dance and distribution of small earthquakes is real and complicated fault zone geometry. Some such varia- significant. tion with depth is reasonable, in view of the com- Calaveras and Hayward faults.--A second well-defined plexity and variations in trend of the mapped sur- belt of abundant earthquake epicenters follows or face fault breaks. However, we cannot entirely closely parallels surface fault breaks along the eliminate the possibility that much of the apparent Calaveras fault zone from its junction with the San discrepancy is due to systematic error arising from Andreas fault 6 miles south of Hollister to a point incorrect assumptions regarding seismic velocities about 5 miles east of Fremont. Farther north, the in the earth's crust near the fault zone. Current epicenters are much less abundant, but they appear to research on crustal structure is aimed at resolving define a trend diverging northwestward from the Cala- this problem. veras fault and merging with mapped surface breaks Whatever the reason for the inexact match between along the Hayward fault zone north of Fremont. This surface faulting and epicentral locations, the belt of less abundant earthquake activity roughly small-earthquake activity is clearly related to the follows the Hayward fault nearly to San Pablo Bay. San Andreas fault, showing that this segment of the Where earthquake epicenters are most abundant along fault is now moving and, at least to some extent, the Calaveras fault, between Pacheco Creek north of is relieving crustal strain. Corroborative evi- Hollister and Calaveras Reservoir, they define a linear dence of fault activity on this segment of the San belt that lies a mile or so northeast of the mapped sur- Andreas fault is afforded by measurements of surface fault breaks of Radbruch (1968b). The mismatch tectonic creep, which show that the blocks on between the belt of epicenters and the surface fault is opposite sides of the surface fault break are similar but in the opposite direction to the mismatch gradually moving apart at rates as great as a half observed on the San Andreas fault south of Watsonville. an inch per year (R. O. Burford, written commun., As with the San Andreas, the reason for lack of precise 1971). The sense of movement is such that the agreement between the surface fault and the belt of ep- block southwest of the fault moves horizontally icenters is not known.