The San Andreas fault is a crack in the Earth's crust in California, some 1100 kilometers long. Many earthquakes have occurred along it, including famous ones in 1857, 1906 and 1989. It marks the boundary between the North American and Pacific lithospheric plates. Geologists divide it into several segments, each with its own distinct behavior. A research project has drilled a deep hole across the fault to study the rock there and listen for earthquake signals. And the geology of the rocks around it sheds light on the fault's history. This page gives information on all these topics.
The San Andreas fault is the foremost of a set of faults along the boundary between the Pacific plate on the west and the North America plate on the east. The west side moves north, causing earthquakes as it moves. The forces associated with the fault have pushed up mountains in some places and stretched apart large basins in others. The mountains include the Coast Ranges and the Transverse Ranges, both of which consist of many smaller ranges. The basins include the Coachella Valley, the Carrizo Plain, the San Francisco Bay, the Napa Valley and many others. The California geologic map will show you more about those.
The northern segment of the San Andreas fault extends from Shelter Cove to south of the San Francisco Bay area. This whole segment, about 300 kilometers long, ruptured on the morning of 18 April 1906 in a magnitude-7.8 earthquake whose epicenter was just offshore, south of San Francisco. In some places the ground shifted by 6 meters, ripping roads, fences and trees apart. "Earthquake trails" on the fault, with explanatory signs, can be visited at Fort Ross, Point Reyes National Seashore, Los Trancos Open Space Preserve, Sanborn County Park and Mission San Juan Bautista. Small portions of this segment ruptured again in 1957 and 1989, but quakes the size of 1906's are not considered likely today.
The 18 April 1906 earthquake occurred just before dawn and was felt in much of the state. Major downtown buildings like the Ferry Building (shown here), well designed by contemporary standards, came through the shaking well. But with the water system disabled by the quake, the city was helpless against the fires that followed. Three days later nearly all of San Francisco's center had burned out, and some three thousand people had died. Many other cities, including Santa Rosa and San Jose, also suffered severe destruction. During reconstruction, better building codes gradually came into force, and today California builders are much more careful about earthquakes. Local geologists discovered and mapped the San Andreas fault at this time. The event was a landmark in the young science of seismology.
The creeping segment of the San Andreas fault extends from San Juan Bautista, near Monterey, to the short Parkfield segment deep in the Coast Ranges. While elsewhere the fault is locked and moves in major earthquakes, here there is constant steady movement of about 3 centimeters per year and relatively small quakes. This kind of fault motion, called aseismic creep, is rather rare. Yet this segment, the related Calaveras fault and its neighbor the Hayward fault all exhibit creep, which slowly bends roadways and pulls buildings apart.
The Parkfield segment is at the center of the San Andreas fault. Hardly 30 kilometers long, this segment is special because it has its own set of magnitude-6 earthquakes that don't involve the neighboring segments. This seismological feature plus three other advantagesthe fault's relatively simple structure, the lack of human disturbance and its accessibility to geologists from both San Francisco and Los Angelesmake the tiny, colorful town of Parkfield a destination out of proportion to its size. A swarm of seismic instruments has been deployed for several decades to catch the next "characteristic earthquake," which finally came on 28 September 2004. The SAFOD drilling project pierces the fault's active surface just north of Parkfield.
The central segment is defined by the magnitude-8 earthquake of 9 January 1857, which broke the ground for about 350 kilometers from the hamlet of Cholame near Parkfield to Cajon Pass near San Bernardino. Shaking was felt over most of California, and motion along the fault was 7 meters in places. The fault takes a large bend in the San Emigdio Mountains near Bakersfield, then runs along the south edge of the Mojave Desert at the foot of the San Gabriel Mountains. Both ranges owe their existence to the tectonic forces across the fault. The central segment has been fairly quiet since 1857, but trenching studies document a long history of great ruptures that will not stop.
From Cajon Pass, this segment of the San Andreas fault runs about 300 kilometers to the shores of the Salton Sea. It splits into two strands in the San Bernardino Mountains that rejoin near Indio, in the low-lying Coachella Valley. There is some aseismic creep documented in parts of this segment. At its south end, the motion between the Pacific and North American plates shifts to a stairstep series of spreading centers and faults that runs down the Gulf of California. The southern segment has not ruptured since some time before 1700, and it is widely considered "overdue" for an earthquake of approximately magnitude 8.
Distinctive rocks and geologic features are found widely separated on both sides of the San Andreas fault. These can be matched across the fault to help unravel its history over geologic time. The records of such "piercing points" show that the plate motion has favored different parts of the San Andreas fault system at different times. Piercing points have clearly demonstrated at least 300 kilometers of offset along the fault system in the last 12 million years. We may find even more extreme examples as research continues.