Figure 1.1

      This chapter briefly describes the San Andreas fault system, its setting along the Pacific Ocean margin of North America, its extent, and the patterns of faulting. Only selected characteristics are described, and many features are left for depictions on maps and figures. The other chapters in this volume elaborate on the history and evolution of the fault system, and the behavior of the Earth's crust and upper mantle within the fault system.

      Because of the extent and complexity of the San Andreas fault system, it is helpful to distinguish between the broad, complex feature as seen on a map of the Western United States and the individual faults on which displacements occur to produce single earthquakes. From larger to smaller features, the terms "fault system," "fault zone," "fault," and "fault branches, splays, strands, and segments" are useful.
      The term "San Andreas fault system" refers to the network of faults with predominantly right-lateral strike slip that collectively accommodate most of the relative motion between the North American and Pacific plates. The boundaries of this fault system are poorly defined, but to separate the San Andreas fault system from other tectonic provinces and systems, it is useful to limit the term to the set of faults along the Pacific rim of North America, both on land and off shore. Accordingly, at the latitude of San Francisco, the system is approximately 80 km wide, and at the latitude of San Diego approximately 150 km wide (see fig. 1 and maps at front of book).
      The term "fault zone" refers to the complex zone of sheared rock that may be from 0.5 to more than 1 km wide and hundreds of kilometers long. The fault zone has developed over a period of millions of years while growing in width and complexity. The terms "fault," "fault branches," and "fault strands" refer to smaller elements and can be applied as needed. For example, surface rupture accompanying an earthquake commonly produces a complex pattern of fractures, and detailed elements can be discussed more effectively by using such terms as "fault branch, splay, or strand." The term "fault segment" recognizes that the fault is not completely continuous but is in sections or parts with poorly defined boundaries, as discussed below in the subsection entitled "Segmentation."

      The San Andreas fault first came into prominence only after it was fully understood by geologists as the cause of the great San Francisco earthquake of 1906. The name had been first used only 9 years previously by A.C. Lawson (1895) for a small segment of the fault on the San Francisco peninsula, where he reported that "a remark- ably straight fault * * * has conditioned the San Andreas and Crystal Springs Valley" (p. 439). Lawson applied the name "San Andreas fault" almost incidentally in a discussion of "subsequent streams that flowed in the valleys." He suggested vertical displacement on the fault but apparently was not convinced of that, inasmuch as he failed to show the fault or, with one exception, displaced strata on several cross sections in his report (for example, pl. 7). Clearly, neither the amount of displacement on the fault nor its great regional and tectonic significance was appreciated at the time of Lawson's work.
      In one of the first reports about the 1906 earthquake, G.K. Gilbert (1907) accurately described the fault and its characteristic displacement, but he did not use the name "San Andreas fault." In his field notes for April 28, 1906, just 10 days after the great earthquake, Gilbert described 20 ft of right-lateral displacement of a road where it crosses the fault at the head of Tomales Bay. He had been following the surface rupture for several days, and on April 26 he recorded in his notes that along the west side of Bolinas Bay "some of the cracks were clearly secondary; others may have been primary." By the time the final report of the State Earthquake Investigation Commission was published (Lawson, 1908), the name "San Andreas" had been adopted, and its characteristics and role in causing the earthquake were clear. That report, which contains a remarkably extensive and accurate account, constitutes a major milestone in our understanding of the San Andreas fault and of strike-slip faults as a class.

      The San Andreas fault system is part of a complex system of faults, isolated segments of the East Pacific Rise, and scraps of plates lying east of the East Pacific Rise that collectively separate the North American plate from the Pacific plate (fig. 1.2). On a more generalized or global scale, the North American plate can be thought of as lying across and partly covering the northern part of the Pacific system of plates. In simplified terms, the Pacific system of plates includes three elements: a westward expanding plate (the Pacific plate), an eastward-expanding plate (the Juan de Fuca plate), and a spreading center (the East Pacific Rise) from which the plates expand as new material is added. To the north, the Pacific plate is underriding, or being subducted under, the North American plate along the Aleutian thrust.

Figure 1.2 - Northeastern Pacific Basin, showing relation of the San Andreas fault as one element in the complex boundary between the North American and Pacific plates. Modified from Drummond (1981).

      Some investigators (Atwater, 1970; Atwater and Molnar, 1973) suggested that the North American plate has converged with and, indeed, slid over the Pacific system of plates, leaving only segments of the East Pacific Rise exposed, to which such names as "Juan de Fuca and Gorda Ridges" are applied. Similarly, related scraps of the eastward-expanding plate are the Juan de Fuca and Gorda plates (fig. 1.2). Absolute plate motions derived by Minster and Jordan (1980), and Jordan and Minster (1988) are shown in figure 1.2, along with the plate-motion vectors derived from the relative migration of mantle plumes or hotspots responsible for volcanic activity in Yellowstone National Park and the Hawaiian Islands.
      At its north end, the San Andreas fault joins the Mendocino Fracture Zone at a high angle, and there three plates are juxtaposed: one moving relatively north-westward, the second southeastward, and the third eastward and northeastward, to form a triple junction. At its southeast end, the San Andreas fault system merges more gradually with the set of transform faults underlying the Gulf of California. Just northwest of the area of merging, however, the trend of the San Andreas fault system changes to much more westerly, whereas a set of echelon faults accompanied by volcanism less than a million years old form a north-south-trending zone that extends northward across the Mojave Desert into Owens Valley of eastern California (Hill and others, 1985). This zone may be considered the East Pacific Rise overridden, and thus modified in pattern, by the North American plate.
      The San Andreas fault system may be viewed as forming the hypotenuse of a right triangle of which the northward extension of the East Pacific Rise and the eastward extension of the Mendocino Fracture Zone are the legs. The model of an overridden, subducted oceanic plate within this triangle and underlying the North American plate (Dickinson and Snyder, 1979) presents significant tectonic problems (see chap. 3).
      The San Andreas fault system has rearranged an assemblage of microplates, or terranes, some of which originated tens of degrees of longitude apart. During the fault's approximately 29-m.y. existence, an extremely complex pattern of rock distribution, has thus has been created (see chap. 3).

      The San Andreas fault system consists primarily of the San Andreas fault and several major branches, such as the Hayward and Calaveras faults in central California and the San Jacinto and Elsinore faults in southern California (fig. 1.3). In addition, in southern California the San Andreas fault splits into northern and southern branches in the eastern Transverse Ranges east of Los Angeles. These major faults accommodate about two-thirds of the right-slip motion between the North American and Pacific plates.

Figure 1.3 - The San Andreas fault system in California. Arrows on San Andreas fault (red) indicate direction of relative movement. Faults dotted where concealed; sawteeth on upper plate of thrusts. Mendocino Fracture Zone queried where uncertain.

      Numerous smaller branches, and extensions of segments of the fault, include in northern California such faults as the Rodgers Creek and Maacama faults, which may be considered northward extensions of the Hayward fault. The Green Valley and Bartlett Springs fault zones extend the Calaveras fault northward in a complex way (see maps at front of book). At the south end of the system, the Imperial fault represents a transition from the more continuous San Andreas fault to a more nearly echelon pattern characteristic of the faults under the Gulf of California. The Superstition Hills and Coyote Creek faults similarly represent a transition from the San Jacinto fault to a more segmented pattern to the south in Mexico.
      In this volume, the San Andreas fault system is considered to lie principally within a belt about 100 km wide by 1,300 km long, but this boundary is arbitrary. Indeed, part of the relative strike-slip motion between the North American and Pacific plates seems to be taken up as far as 1,000 km east of the coastline throughout the Great Basin province (Jordan and Minster, 1988). The name "San Andreas fault system," however, should be confined to the more limited belt with the highest concentration of right-lateral strike slip.

      With some notable exceptions, the fault trends about N. 35° -40° W. (fig. 1.3). In its central section between the latitudes of San Jose and Bakersfield, the fault is relatively simple and straight, but farther to the south and north several branches splay from the main active trace. Near San Jose, where the fault bends about N. 50° W., the Calaveras and Hayward faults splay to the east and trend between N. 20° and 35° W. South of the latitude of Bakersfield, the main fault changes most sharply in strike, in what commonly is referred to as the Big Bend reach of the fault. For 120 km or more the fault strikes about N. 60° W., where it bounds the Mojave block on the south. This bend has significant tectonic implications (see chaps. 2 and 3, and maps at front of book).
      South of the latitude of Los Angeles, the Elsinore and San Jacinto faults splay to the southeast, forming, in a general way, a reversed image of the splays in the San Francisco Bay region, although both faults trend about N. 50 W.

      In addition to the right-lateral strike-slip faults that characterize the San Andreas fault system, faults displaying left-lateral strike slip, as well as thrust faults and reverse faults of many sizes, are present (see maps at front of book). Normal faults are less common but are present in some places, for example, in zones of extension at the crest of folds associated with the major faults, in the bordering ranges, and at jogs in the fault where local extension is to be found.
      Most conspicuous of the faults displaying left-lateral slip is the Garlock fault which intersects the San Andreas fault at about lat 35° N. and extends northeastward and eastward from there for 240 km (fig. 1.3; see maps at front of book). On the west side of the San Andreas fault, the southwest- and east-west-trending Big Pine fault joins the San Andreas fault a few kilometers northwest of the point where the Garlock fault joins the San Andreas fault.
      The left-lateral Pinto Mountain fault zone joins the San Andreas fault on its east side at about lat 34° N. and extends northeastward, in a pattern not unlike that of the Garlock fault. The Blue Cut fault is another left-lateral fault in the same general area.
      On a broad regional scale, thrust faults and detachments that accommodate subduction of the lower crust are significant, and various interpretations and speculations have been offered (fig. 1.4; Weldon and Humphreys, 1986; Namson and Davis, 1988). Intermediate-scale thrust faults that border the Transverse Ranges on the south side are characterized by such faults as the San Fernando fault zone (Grantz, 1971) and the Cucamonga fault zone (fig. 1.5). At the north end of the San Andreas fault system, where it joins the Mendocino Fracture Zone, such thrust faults as the Kings Range fault similarly accommodate crustal shortening. Smaller thrust faults that flank the San Andreas fault and dip toward it are common along many parts of the fault zone.

Figure 1.4 - Some interpretations suggest that the San Andreas fault bottoms at a detachment fault and that numerous thrust faults, as well as the San Andreas fault itself, are important elements which accommodate the relative displacements between the North American and Pacific plates (from Namson and Davis, 1988). Heavy lines, faults; arrows indicate direction of relative movement: A, away from observer; T, toward observer.

Figure 1.5 - Complex branching and changes in the style of faulting characterize the eastern branch of the San Andreas fault system in southern California. Although right-lateral slip strike characterizes the main San Andreas fault and its major branches, thrust faults, normal faults and left-lateral faults are also present (modified from Matti and others, 1985).

      A range of fault types, and the complexities that typify much of the fault zone, are well illustrated near Cajon Pass and southeast of there (fig. 1.5), where the San Andreas fault zone splits into a northern and a southern branch. The strike of the San Gorgonio Pass fault zone in the east-central part of figure 1.5 changes in several places, and depending on the trend of a given segment, strike or dip slip may predominate. In this same area, the San Andreas fault, which to the northwest is relatively continuous and linear, ends as a surface feature, and the style of deformation changes from strike slip to primarily dip slip at the surface. Left-lateral slip characterizes the Pinto Mountain fault in the eastern part of figure 1.5, and small normal faults can be found throughout this area.

      Different behavior patterns along different parts of the fault began to be recognized when Steinbrugge and Zacher (1960) found that continuous slip or "creep" occurred along the fault in central California. Allen (1968) delineated five different regions along the San Andreas fault, three displaying seismic activity and two displaying little or no current activity. Wallace (1970) described in more detail major differences in behavior along different segments.
      Large segments of the fault system that are believed to produce damaging earthquakes are illustrated in figure 1.6, which also shows an evaluation of the probabilities of earthquakes of different magnitude along these major segments of the San Andreas fault and three of its branches, the Hayward, San Jacinto, and Imperial faults. Both historical seismicity and paleoseismic evidence of large earthquakes and slip rates that characterize these different segments have been used in this assessment (Working Group on California Earthquake Probabilities, 1988).

Figure 1.6 - Segments of the San Andreas fault system display different behavior. Here are shown conditional probabilities for the occurrence of major earthquakes along segments of the San Andreas (A) and the Hayward, San Jacinto, and Imperial faults (B) for the 30-year interval from 1988 to 2018. From Working Group on California Earthquake Probabilities (1988).

      Segmentation at a scale of a few to several kilometers is shown in figure 1.7, and a cumulative plot of the segment lengths in figure 1.8. The maximum length of these segments is about 18 km, but a more significant upper range appears to be near 10 km. These mappable segments are based on relatively fresh geomorphic features considered to be "young," that is, probably less than several thousand years old.
      Both left (fig. 1.7)- and right-stepping echelon patterns are displayed, and combinations or transition zones are, also present. Complex patterns are common.

Figure 1.7 - Individual branches and strands of surface trace of the San Andreas fault are arranged in various patterns: A-C, left-stepping echelon arrangement; D-F, right-stepping echelon arrangement; G, both left- and right-stepping arrangements; H-J, complex arrangements of individual strands. Note that individual strands may be at an angle of as much as 12° to general trend of fault zone. Sources: A, D (Ross, 1969); B (Hope, 1969; see references in this chapter); C, J (Brown and Wolfe, 1972); E, H, I (Vedder and Wallace, 1970); F, G (Brown, 1970; see references in chap. 2).

Figure 1.8 - Census of lengths of individual segments of the San Andreas fault. Maximum lengths are in the range 10-18 km. Line is fitted by eye to show trend of values. From Wallace (1973).

REFERENCES CITED