San Andreas Fault

A map displaying each of the seven major faults in the San Francisco Bay Area, and the probability of an M6.7 earthquake or higher occurring on each fault between 2003 and 2032.

The northern segment of the fault runs from Hollister, through the Santa Cruz Mountains, epicenter of the 1989 Loma Prieta earthquake, then up the San Francisco Peninsula, where it was first identified by Professor Lawson in 1895, then offshore at Daly City near Mussel Rock. This is the approximate location of the epicenter of the 1906 San Francisco earthquake. The fault returns onshore at Bolinas Lagoon just north of Stinson Beach in Marin County. It returns underwater through the linear trough of Tomales Bay which separates the Point Reyes Peninsula from the mainland, runs just east of Bodega Head through Bodega Bay and back underwater, returning onshore at Fort Ross. (In this region around the San Francisco Bay Area several significant "sister faults" run more-or-less parallel, and each of these can create significantly destructive earthquakes.) From Fort Ross, the northern segment continues overland, forming in part a linear valley through which the Gualala River flows. It goes back offshore at Point Arena. After that, it runs underwater along the coast until it nears Cape Mendocino, where it begins to bend to the west, terminating at the Mendocino Triple Junction.

The central segment of the San Andreas Fault runs in a northwestern direction from Parkfield to Hollister. While the southern section of the fault and the parts through Parkfield experience earthquakes, the rest of the central section of the fault exhibits a phenomenon called aseismic creep, where the fault slips continuously without causing earthquakes.

Los Angeles

San Diego

Mexicali

Salton Sea

Tijuana

Alamo River mouth

Colorado River mouth

Laguna Salada

Yuma

Gulf of California

Colorado River Valley

Cerro Prieto volcano

Southern San Andreas Fault. The red lines are simplified faults. Right-lateral direction of motion of the transform fault is shown (pink arrows). The red rhombs are pull-apart basins; the northern one is the site of the Niland geothermal field, the southern the Cerro Prieto geothermal field.

Aerial photo of the San Andreas Fault in the Carrizo Plain

The Vasquez Rocks in Agua Dulce, California are evidence of the San Andreas Fault and part of the 2,650-mile Pacific Crest Trail.

The southern segment (also known as the Mojave segment) begins near Bombay Beach, California. Box Canyon, near the Salton Sea, contains upturned strata associated with that section of the fault.[4] The fault then runs along the southern base of the San Bernardino Mountains, crosses through the Cajon Pass and continues northwest along the northern base of the San Gabriel Mountains. These mountains are a result of movement along the San Andreas Fault and are commonly called the Transverse Range. In Palmdale, a portion of the fault is easily examined at a roadcut for the Antelope Valley Freeway. The fault continues northwest alongside the Elizabeth Lake Road to the town of Elizabeth Lake. As it passes the towns of Gorman, Tejon Pass and Frazier Park, the fault begins to bend northward, forming the "Big Bend". This restraining bend is thought to be where the fault locks up in Southern California, with an earthquake-recurrence interval of roughly 140–160 years. Northwest of Frazier Park, the fault runs through the Carrizo Plain, a long, treeless plain where much of the fault is plainly visible. The Elkhorn Scarp defines the fault trace along much of its length within the plain.

The southern segment, which stretches from Parkfield in Monterey County all the way to the Salton Sea, is capable of an 8.1-magnitude earthquake. At its closest, this fault passes about 35 miles (56 km) to the northeast of Los Angeles. Such a large earthquake on this southern segment would kill thousands of people in Los Angeles, San Bernardino, Riverside, and surrounding areas, and cause hundreds of billions of dollars in damage.[5]

The fault was first identified in Northern California by UC Berkeley geology professor Andrew Lawson in 1895 and named by him after the Laguna de San Andreas, a small lake which lies in a linear valley formed by the fault just south of San Francisco. Eleven years later, Lawson discovered that the San Andreas Fault stretched southward into southern California after reviewing the effects of the 1906 San Francisco earthquake. Large-scale (hundreds of miles) lateral movement along the fault was first proposed in a 1953 paper by geologists Mason Hill and Thomas Dibblee. This idea, which was considered radical at the time, has since been vindicated by modern plate tectonics.[12]

Seismologists discovered that the San Andreas Fault near Parkfield in central California consistently produces a magnitude 6.0 earthquake approximately once every 22 years. Following recorded seismic events in 1857, 1881, 1901, 1922, 1934, and 1966, scientists predicted that another earthquake should occur in Parkfield in 1993. It eventually occurred in 2004. Due to the frequency of predictable activity, Parkfield has become one of the most important areas in the world for large earthquake research.

In 2004, work began just north of Parkfield on the San Andreas Fault Observatory at Depth (SAFOD). The goal of SAFOD is to drill a hole nearly 3 kilometres (1.9 mi) into the Earth's crust and into the San Andreas Fault. An array of sensors will be installed to record earthquakes that happen near this area.[13]

The San Andreas Fault System has been the subject of a flood of studies. In particular, scientific research performed during the last 23 years has given rise to about 3,400 publications.[14]

Radar generated 3-D view of the San Andreas Fault, at Crystal Springs Reservoir near San Mateo, California.[15]

A study published in 2006 in the journal Nature found that the San Andreas fault has reached a sufficient stress level for an earthquake of magnitude greater than 7.0 on the moment magnitude scale to occur.[16] This study also found that the risk of a large earthquake may be increasing more rapidly than scientists had previously believed. Moreover, the risk is currently concentrated on the southern section of the fault, i.e. the region around Los Angeles, because massive earthquakes have occurred relatively recently on the central (1857) and northern (1906) segments of the fault, while the southern section has not seen any similar rupture for at least 300 years. According to this study, a massive earthquake on that southern section of the San Andreas fault would result in major damage to the Palm Springs–Indio metropolitan area and other cities in San Bernardino, Riverside and Imperial counties in California, and Mexicali Municipality in Baja California. It would be strongly felt (and potentially cause significant damage) throughout much of Southern California, including densely populated areas of Los Angeles County, Ventura County, Orange County, San Diego County, Ensenada Municipality and Tijuana Municipality, Baja California, San Luis Rio Colorado in Sonora and Yuma, Arizona. Older buildings would be especially prone to damage or collapse, as would buildings built on unconsolidated gravel or in coastal areas where water tables are high (and thus subject to soil liquefaction). The paper concluded :

The information available suggests that the fault is ready for the next big earthquake but exactly when the triggering will happen and when the earthquake will occur we cannot tell ... It could be tomorrow or it could be 10 years or more from now.[16]

Nevertheless, in the 14 years since that publication there has not been a substantial quake in the Los Angeles area, and two major reports issued by the U.S. Geological Survey (USGS) have made variable predictions as to the risk of future seismic events. The ability to predict major earthquakes with sufficient precision to warrant increased precautions has remained elusive.[17]

The U.S. Geological Survey most recent forecast, known as UCERF3 (Uniform California Earthquake Rupture Forecast 3), released in November 2013, estimated that an earthquake of magnitude 6.7 M or greater (i.e. equal to or greater than the 1994 Northridge earthquake) occurs about once every 6.7 years statewide. The same report also estimated there is a 7% probability that an earthquake of magnitude 8.0 or greater will occur in the next 30 years somewhere along the San Andreas Fault.[18] A different USGS study in 2008 tried to assess the physical, social and economic consequences of a major earthquake in southern California. That study predicted that a magnitude 7.8 earthquake along the southern San Andreas Fault could cause about 1,800 deaths and $213 billion in damage.[19]

A 2008 paper, studying past earthquakes along the Pacific coastal zone, found a correlation in time between seismic events on the northern San Andreas Fault and the southern part of the Cascadia subduction zone (which stretches from Vancouver Island to northern California). Scientists believe quakes on the Cascadia subduction zone may have triggered most of the major quakes on the northern San Andreas within the past 3,000 years. The evidence also shows the rupture direction going from north to south in each of these time-correlated events. However the 1906 San Francisco earthquake seems to have been the exception to this correlation because the plate movement was moved mostly from south to north and it was not preceded by a major quake in the Cascadia zone.[20]