San Andreas Fault

The San Andreas fault has been recognized as a major geologic feature of California and of North America for nearly a century. It is hoped that research into this seismically active fault will help geologists to develop a simple, characteristic model to explain the behavior of the fault and possibly forecast potentially destructive earthquakes in California.

California's Earthquake History

The San Andreas fault is the longest fault in California and perhaps the longest in North America. Its total length is more than 1,600 kilometers. Because about three-quarters of California's population lives within 80 kilometers of the fault, its existence—and its potential for a major unannounced earthquake—is responsible for formulating many public policy decisions and for establishing the design criteria of many engineering projects in the state. In spite of much being known about the fault, geoscientists who have studied it readily admit that their knowledge remains far from complete.

California's earthquake history is short. The earliest recorded earthquake in the state was on July 28, 1769. It was experienced by the Spanish explorer Gaspar de Portolá while he camped along the Santa Ana River southeast of Los Angeles. One of the state's earliest recorded large earthquakes occurred on the Hayward fault, a branch of the San Andreas system in the East Bay on June 10, 1836, during which surface breakage occurred along the western base of the Berkeley Hills. Another large earthquake on October 21, 1868, also centered on the Hayward fault, caused 30 fatalities and ground breakage for about 32 kilometers, accompanied by as much as 1 meter of right-slip between San Leandro and Warm Springs. Other notable nineteenth-century earthquakes along the San Andreas system in the Bay Area occurred in June 1838, on October 8, 1865, and on April 24, 1890. Unfortunately, there are no seismographic records of these early earthquakes because it was not until 1887 that the first seismographs began being used in the United States.

Since 1850, California has experienced three great earthquakes of Richter magnitude 8 or greater. The Richter scale, which was introduced in 1935, defines earthquakes of magnitude 2 as “felt,” those of magnitude 4–4.5 as “causing local damage,” those of magnitude 6–6.9 as “moderate,” those of magnitude 7–7.5 as “major,” and those exceeding magnitude 7.5 as “great.” Two of these three great earthquakes—the San Francisco earthquake of 1906 and the Fort Tejon earthquake of 1857—resulted from movement of the San Andreas fault.

Recognition of the San Andreas Fault

It was not until after the Great San Francisco earthquake of April 18, 1906, that the San Andreas was first recognized as a continuous regional geologic structure of California. This earthquake caused a sudden right-slip of up to 5 meters, ground rupturing for 420–470 kilometers, an estimated 700 fatalities, and damage estimated at between $350 million and $1 billion. The earthquake is generally considered as having attained a magnitude of 8.3 on the Richter scale. Much of the property loss in the San Francisco disaster was caused by the extensive fires following the earthquake that resulted from ruptured gas lines and a lack of water from broken water mains. The strongest earthquake in the Bay Area associated with the San Andreas fault since 1906 was the magnitude 7.1 earthquake of October 17, 1989, which resulted in 63 casualties and caused more than $1 billion in damage.

The other important California earthquake of magnitude comparable to the 1906 San Francisco disaster caused by the San Andreas fault was the Fort Tejon earthquake of January 9, 1857. This earthquake had an estimated magnitude of 8.3 and caused ground breakage for 360–400 kilometers on the southern part of the San Andreas fault zone, from the Cholame Valley in the Coast Ranges to the Transverse Ranges as far south as the present site of Wrightwood. Ground motion was felt from north of Sacramento to Fort Yuma on the lower Colorado River. The ground accelerations during this earthquake were so great that mature oak trees in the Fort Tejon area were toppled, buildings collapsed, fish in a lake near the fort were tossed out of the water, and the Los Angeles River was thrown out of its banks. An informative account of the effects of the 1857 earthquake, taken from correspondence and newspaper accounts of the time, was published by D. Agnew and Kerry Sieh in 1978. Right-slip movement ranging from 4.5 to 4.8 meters occurred in the 1857 event. Although the 1857 earthquake had the same estimated magnitude as the 1906 earthquake, the 1857 earthquake was potentially more destructive than the 1906 disaster because the ground shaking is reported to have lasted from 1 to 3 minutes. Compare this time interval with the short 10- to 30-second shaking accompanying the moderate 1971 San Fernando earthquake, magnitude 6.3, in which 60 lives were lost, and the 1989 San Francisco earthquake of 7.1, which lasted for approximately 15 seconds and whose damage, although severe in certain local areas, was minor in comparison to that of earlier, larger-magnitude, longer-lasting earthquakes. (Evidence, although inconclusive, points to the possibility that the 1989 event was a strike-slip earthquake along the San Andreas fault.) Such long duration and low seismic frequencies associated with events such as the 1857 earthquake occurring in California's highly populated regions are more likely to cause serious damage to large buildings and claim more lives than are moderate earthquakes.

Definition of the San Andreas Fault

The name “San Andreas” was first used in 1895 by Andrew C. Lawson, a geologist at the University of California, Berkeley. Later, it was Lawson who headed the California Earthquake Investigation Commission, whose monumental report on the San Andreas system was published in 1908. This report describes surface ruptures that developed in the 1906 earthquake that extended for more than 420 kilometers from Humboldt County to San Juan Bautista in San Benito County and continued to follow the fault southward through the Coast Ranges into San Bernardino County. The report used the term “San Andreas rift” for the rift-valley surface expression in the San Andreas Lake area on the San Francisco Peninsula, and from this term have evolved the terms “San Andreas fault,” “San Andreas fault zone,” and “San Andreas system.”

Curiously, in spite of the ample evidence of right-slip motion described for the segment that ruptured in the 1906 earthquake, the 1908 report maintained that the dominant characteristic movement on the San Andreas fault throughout its geologic history had been vertical (up and down) by as much as thousands of meters. This idea of vertical motion became the prevailing idea of motion for the San Andreas for almost fifty years. It has since been agreed, however, that activity along the fault has shown dominantly horizontal motion, with the North American plate moving southward relative to the Pacific plate. This kind of horizontal motion is termed right-lateral strike slip, or, more simply, right-slip, because the sideways displacement of the block across the fault from the observer appears to be to the right. Thus, if the historically observed motion is typical of the movement of the San Andreas fault in the geological past, rocks that once faced each other across the fault should now be separated. Because the amount of displacement of the faulted rock unit is known, it is possible to establish the ages of many rocks that were once a single unit. Scientists are therefore able to determine the movement rate (called slip rate) during the geologic history of the fault.

Current usage defines the San Andreas fault proper as the strand of the San Andreas zone that reveals surface rupturing produced by recent movements within the zone. The definition of the term “San Andreas zone” incorporates numerous parallel to subparallel related fractures that may be separated by 10–15 kilometers. Many of these subordinate fractures record that the locus of movement in the zone has shifted from one branch to another through time. Some examples of active or formerly active subordinate strands are the Pilarcitos, Calaveras, and Hayward faults in northern California and the Punchbowl, San Gabriel, San Jacinto, and Banning faults in southern California.

Extent of the San Andreas Fault

On land, the San Andreas fault extends from Shelter Cove in Humboldt County southward, crossing the Golden Gate west of the Golden Gate Bridge, and under southwestern San Francisco, where the fault trace is obscured by housing tracts in Daly City. It appears that city planners, developers, and residents have ignored the hazards of living on this infamous fault. The fault continues southward along the length of the Santa Cruz Mountains of the San Francisco Peninsula, where it forms the rift valley occupied by the San Andreas and Crystal Springs reservoirs, and through the Coast Ranges past its intersection with the Garlock fault at Frazier Mountain near Tejon Pass. Beginning here, the fault turns more eastward—geologists call this segment of the fault the “big bend”—and marks the approximate northern boundary of the San Gabriel Mountains. It passes through Cajon Pass, which separates the San Gabriel and San Bernardino mountains, and continues southeastward along the southern margin of the San Bernardino Mountains to the vicinity of San Gorgonio Pass. Here, the San Andreas divides into the North and South branches, with the North Branch joining the Mission Creek fault and the South Branch joining the Banning fault. North of Indio, the branches rejoin, forming what some geologists call the “southern big bend,” and the San Andreas fault resumes its more southerly trend to border the eastern edge of the Salton Sea trough and continue south into the Gulf of California.

Near the big bend at the Frazier Mountain intersection of the San Andreas and the Garlock faults are also the ends of the San Gabriel and Big Pine faults and the Frazier Mountain thrust. This area is known to geologists as the “knot,” where both right-slip faults, the San Gabriel and San Andreas, are intertwined with left-slip faults, the Garlock and Big Pine. An important parallel branch of the San Andreas system is the San Jacinto fault. It splays off the San Andreas fault at the northeastern end of the San Gabriel Mountains and runs southeastward under San Bernardino, San Jacinto, and Hemet to El Centro.

The northern and southern ends of the San Andreas fault terminate in ways that are still not entirely clear to geologists. At its northern end, the San Andreas disappears into the Pacific Ocean at Shelter Cove near Point Mendocino, apparently to join the Mendocino fracture zone and the Juan de Fuca oceanic trench at a point known to geologists as a “transform-transform-trench triple junction.” The southern end of the San Andreas disappears beneath the waters of the Gulf of California, where it appears to develop into a number of northwest-trending transform faults that offset the East Pacific Rise. The San Andreas system, therefore, links the East Pacific Rise to the Juan de Fuca Ridge as a transform fault and marks the approximate position of a very major structural feature of the earth's crust, the sliding boundary between the Pacific and North American crustal plates. Most of California rests on the North American plate, but much of southern California is on the Pacific plate.

Seismology and Field Study

Virtually every technique available to geoscientists has been used in studying the San Andreas fault. Most prominent, perhaps, is the work of seismologists and geophysicists who study the waves of energy generated in earthquakes. Seismologists in California study the activity of the state's fault systems by monitoring all seismic activity in the state with seismometers, instruments that record the passage of earthquake waves. The data from these instruments allow seismologists to determine the epicentral location, depths of foci (hypocenters), surface area, and movement directions of faults responsible for given earthquakes.

The geologic history of the San Andreas fault zone has been learned by fieldwork, which began in earnest following the San Francisco earthquake of 1906. Fieldwork is the process of working in the field by foot, car, or aircraft to locate and plot on aerial photographs and topographic base maps the locations of the rock units, faults, and folds in an area of interest. A primary goal of field mapping is correlation—that is, the determination of age relationships between rock units or geologic events in separate areas. For decades following the 1906 disaster, field mapping accumulated data on the structures and rock units exposed along the fault. For many years, the field data were interpreted to support a model supposing that the movement history of the San Andreas was primarily vertical. A tentative proposal of 38 kilometers of right-slip motion for the fault was advanced in 1926, but it was not widely accepted.

In 1953, in what may be the most important benchmark paper ever presented on California geology, two California geologists, Mason L. Hill and Thomas W. Dibblee, Jr., suggested that as much as 560 kilometers of movement had taken place along the San Andreas fault in a period of 150 million years. Their proposal resulted from careful analysis of field observations that allowed correlation of rocks and fossils in separate exposures on opposite sides of the fault. They illustrated that not only had the oldest rocks studied been offset by 560 kilometers, but younger rocks—rocks that had not existed long enough to be offset as much—had also been displaced by progressively lesser distances. For example, they showed that the correlation of uniquely similar rocks in the Gabilan and San Emigdio ranges illustrated 282 kilometers of offset in about 25 million years, that offset beds of gravel in the Temblor Hills and San Emigdio Mountains revealed 22 kilometers of offset in less than 1 million years, and that the Big Pine fault had been displaced from its eastern extension, the Garlock fault, by 10 kilometers in the last 200,000 years.

Few geologists at the time could accept the idea of such great right-lateral displacement along the San Andreas fault. Nevertheless, so much interest was aroused by the tentative conclusions of Hill and Dibblee that many geologists began to conduct their own investigations. Consequently, subsequent field studies clearly confirmed the thesis. The work of Hill and Dibblee has proved to be the fundamental framework to which explanations of the geology of the Coast, Transverse, and San Bernardino ranges; the Salton Trough; and the Great Central Valley must conform. In retrospect, a large cumulative right-lateral offset by the San Andreas fault is a central component of the global tectonics theory. Work began in 2004 just north of Parkfield in central California on the San Andreas Fault Observatory at Depth (SAFOD), funded by the National Science Foundation (NSF), the US Geological Survey and the National Aeronautics and Space Administration. Parkfield is a small town on the San Andreas fault that has experienced moderate earthquakes (magnitude 6) about every 22 years. By 2007, a borehole was angled into the fault at a depth of about 3 kilometers. Instruments have been placed in the hole at several points to monitor the behavior of the fault.

Forecasting the Next “Big One”

Estimates are that a repeat of the 1906 San Francisco earthquake would cause tens of thousands of casualties and tens of billions of dollars in damages. A repeat of the 1857 event would jolt southern California with a similarly destructive earthquake, which could cause a similar amount of casualties and damages and leave hundreds of thousands of homes unfit for habitation.

Scientists are regularly asked to predict when the next “Big One” will occur. To arrive at answers to such questions, a different type of fieldwork was undertaken in the 1960s and 1970s. Studies by Tanya Atwater, a geologist who worked at Scripps Institute of Oceanography, showed that plate tectonics, which was at the time considered to be strictly an ocean-bound concept, was sometimes critical in understanding the deformation of crustal rocks within a continent. In terms of global tectonics, the San Andreas fault plays a very important role in understanding the geology of the western United States. Within the framework of plate tectonics, the San Andreas fault is a transform fault along which strain is occurring between the two moving plates.

To measure strain across the San Andreas fault, precision geodetic surveying began in the 1970s, using tellurometers, geodimeters, and long baseline interferometry (LBLI), a “space-geodetic” technique developed in the late 1970s. LBLI relies on extraterrestrial reference points, such as quasars, to measure distances to an incredible precision of 1 centimeter or less in 1,000 kilometers. Precision geodetic surveying and LBLI reveal that the relative motion between the Pacific and North American plates is about 5.5 centimeters per year, right-slip. About 4 centimeters of this motion is taken up each year by the San Andreas fault, with the remaining motion being accommodated by the slippage of other faults and in other ways within the San Andreas zone.

Thus, seismology, geologic field mapping during the first half of the twentieth century, large-scale tectonic and geodetic evidence of the 1970s and 1980s, and the recorded history of earthquakes along the San Andreas fault all suggest that Californians can expect the fault to move again—but questions remain about when, where, and with what frequency great earthquakes will occur along the San Andreas fault.

Determining Average Recurrence Interval

Researchers who estimate the frequency of earthquakes on an active fault use a still different kind of fieldwork: logging and mapping the walls of 20-foot-deep trenches dug across a fault in association with carbon-14 isotopic age dating and statistics. Through these efforts, geologists establish an average recurrence interval (RI) of earthquakes for that particular fault. The reason for attempting to establish the history of a fault's activity is that much of geological interpretation is based on the premise that geologic events occurring today result from the same processes that caused them in the past, or uniformitarianism. Uniformitarianism holds that the activity of the San Andreas fault in the geological past is the best clue to forecasting the fault's behavior in the future.

To develop a regional overview of the San Andreas fault, the fault has been divided into four segments based on existing knowledge of seismic activity: northern, central, south-central, and southern. At least two recorded large earthquakes have occurred on the northern segment, the 1838 and the 1906 events. Although little detailed work has been done on the northern segment, it is believed that the frequency of earthquakes here is similar to that on the south-central segment. The central segment has a historic record of as much as 3 centimeters of offset per year. The strain along this segment of the fault, however, appears to be released slowly by aseismic creep (very slow, incremental movement along the fault that is unaccompanied by earthquake activity). The southern segment, from Indio south, is considered dormant, with no record of seismic activity since 1688.

Using this information as a starting point, Kerry Sieh and his associates at the California Institute of Technology have trenched three sites on the south-central and southern segments of the San Andreas fault to obtain the information necessary to determine the average RI. The first of these sites is at Pallett Creek near Palmdale, about 55 kilometers northeast of Los Angeles; another is just north of the big bend at Wallace Creek on the Carrizo Plain west of Bakersfield; and the third locality is on the southern segment near Indio. Ray Weldon, of the University of Oregon, has studied a fourth area astride the south-central segment at Lost Lake in Cajon Pass.

Results of Recurrence Interval Study

The results of the Pallett Creek work reveal the RI for ten episodes of faulting in a period of about 2,000 years to be about 132 years. Because the Pallett Creek area was affected by the great southern California earthquake of 1857, Sieh's research seems to suggest that this segment of the San Andreas should break again soon. Sieh is quick to point out, however, that the site's average RI could have limited meaning because of newly perfected methods of carbon-14 dating, developed by Minze Stuiver of the University of Washington and statistician David Brillinger of the University of California at Berkeley, that allows much tighter precision in age dating than has been possible in the past. Consequently, age determinations of the ten events reveal that the earthquakes are clustered in bunches of two or three, with individual earthquakes within the clusters separated by tens of years. Five of the intervals between clusters are less than a century, and three of the remaining four intervals span from 200 to 330 years. From this field, laboratory, and statistical information, Sieh's interpretation of the Pallett Creek record is that it is probably near the middle of a two-hundred-year-long dormant period. Thus, if the pattern of earthquakes at Pallett Creek is truly representative of the fault's behavior of the south-central segment, it might be eighty years or more before Los Angeles could expect another great earthquake along this portion of the San Andreas fault.

Weldon's investigation at Lost Lake, about 40 kilometers southeast of the Pallett Creek site, reveals six earthquakes in the last thousand years with an average RI of 150–200 years. Thus, both the Pallett Creek and Lost Lake sites reveal relatively short RI values that appear to be at odds with what is known about the frequencies of earthquakes on the segments of the San Andreas to the north and south. Sieh's work at Wallace Creek, about 100 kilometers northwest of the Pallett Creek site and north of the big bend, results in a longer average RI of 250-450 years, and investigations on the southern segment show that this segment has not had a major, or great, event since about 1688, even though the RI for this segment is about 220 years. Comparing the results from the Indio, Lost Lake, Wallace, and Pallett creeks, Weldon and Sieh have developed several alternate models of behavior of the south-central and southern segments of the fault. Two of their models suggest that the southern segment is the most likely site for the next large earthquake. This is a sparsely populated region at present, about 120 kilometers east of Los Angeles, far enough that a great earthquake on the southern segment is currently not a serious threat to the residents of Los Angeles.

In 2017, a research report published by US Geological Survey geologist Kate Scharer shared information discovered following a focused study on a 100-mile section of the southern part of the San Andreas Fault next to Frazier Mountain, not far from the intersection of Los Angeles, Ventura, and Kern counties. According to Scharer, her findings revealed that the land on either side of the fault had been pushing against each other at a rate of more than one inch per year since 1857. The amount of energy pent up through this movement over such a long period of time, she claimed, served as a warning sign that the section could experience a major earthquake in the near future. This information was based off of conclusions that ten major earthquakes had occurred in that section over a one thousand–year period.

Principal Terms

epicenter: the point on the earth's surface directly above the focus of an earthquake

geodetic surveying: surveying in which account is taken of the figure and size of the earth and corrections are made for the earth's curvature

geodimeter: an electronic optical device that measures ground distances precisely by electronic timing and phase comparison of modulated light waves that travel from a master unit to a reflector and return to a light-sensitive tube; its precision is about three times as great as that of a tellurometer

right-slip (right-lateral strike-slip): sideways motion along a steep fault in which the block of the earth's crust across the fault from the observer appears to be displaced to the right; left-slip faults are exactly the opposite

seismic: pertaining to an earthquake

tectonics: a branch of geology that deals with the regional study of large-scale structural or deformational features, their origins, mutual relations, and evolution

tellurometer: a portable electronic device that measures ground distances precisely by determining the velocity of a phase-modulated, continuous microwave radio signal transmitted between two instruments operating alternately as a master station and a remote station; it has a range up to 65 kilometers

triple junction: a point on the earth's surface where three different global plate boundaries join

Bibliography

Abaimov, S. G. et al. “Earthquakes: Recurrence and Interoccurrence Times.” Pure and Applied Geophysics, vol. 165, 2008, pp. 777–795. doi.org/10.1007/978-3-7643-8757-0‗20. Accessed 15 Apr. 2023.

Agnew, D., and K. Sieh. “A Documentary Study of the Felt Effects of the Great California Earthquake of 1857.” Bulletin of the Seismological Society of America, vol. 68, 1978, pp. 1717–1729.

Clarke, Thurston. California Fault: Searching for the Spirit of State Along the San Andreas. Ballantine Books, 1996.

Collier, Michael. A Land in Motion: California's San Andreas Fault. Golden Gates National Parks Association, 1999.

Condie, Kent C. Plate Tectonics and Crustal Evolution. 4th ed., Butterworth Heinemann, 1997.

Davis, Craig A. “Los Angeles Water Supply Impact from a M7.8 San Andreas Fault Earthquake Scenario.” Journal of Water Supply: Research and Technology, vol. 59, no. 6–7, 2010, pp. 408–417. doi.org/10.2166/aqua.2010.006. Accessed 15 Apr. 2023.

Fradkin, Philip L. Magnitude 8: Earthquakes and Life Along the San Andreas Fault. U of California P, 1999.

Hill, M. L. “San Andreas Fault: History of Concepts.” Bulletin of the Geological Society of America, vol. 92, 1981, pp. 112–131. activetectonics.asu.edu/ActiveFaultingSeminar/Papers/Hill‗1981.pdf. Accessed 15 Apr. 2023.

Iacopi, Robert, ed. Earthquake Country: California. 4th ed., Sunset Books/Lane, 1996.

Jordan, T. H., and J. Minster. “Measuring Crustal Deformation in the American West.” Scientific American, vol. 259, 1988, pp. 48–58.

Kerr, R. A. “Take Your Choice: Ice Ages, Quakes, or Impacts.” Science, vol. 243, 1989, pp. 479–80. doi.org/10.1126/science.243.4890.479. Accessed 15 Apr. 2023.

Lin II, Rong-Gong. "A Section of the San Andreas Fault Close to L.A. Could Be Overdue for a Major Earthquake." Los Angeles Times, 7 Mar. 2017, www.latimes.com/local/lanow/la-me-ln-earthquake-san-andreas-grapevine-20170307-story.html. Accessed 15 Apr. 2023.

Lynch, David. "The San Andreas Fault." Geology, geology.com/articles/san-andreas-fault.shtml. Accessed 15 Apr. 2023.

Norris, R. M., and R. W. Webb. Geology of California. 2d ed., John Wiley & Sons, 1990.

Plummer, Charles C., and Diane Carlson. Physical Geology. 12th ed., McGraw-Hill, 2007.

Powell, Robert, et al., eds. The San Andreas Fault System: Displacement, Palinspastic Reconstruction, and Geological Evolution. Geological Society of America, 1993.

Sieh, K. E. et al. “A More Precise Chronology of Earthquakes Produced by the San Andreas Fault in Southern California.” Journal of Geophysical Research, vol. 94, 1989, pp. 603–623. doi.org/10.1029/JB094iB01p00603. Accessed 15 Apr. 2023.

Stoffer, Philip W. Where's the San Andreas Fault?: A Guidebook to Tracing the Fault on Public Lands in the San Francisco Bay Region. U of Michigan Library, 2006.

Walker, Bryce. Earthquake. Time-Life Books, 1982.