It is important to remember that the plate boundary is quite wide in some places and consists of a host of structures, and the the San Andreas fault is only one major fault in the complex. This is partly evident in the photo to the right taken from a commercial flight looking N. The red dots trace the San Andreas fault and the San Jacinto fault. The Mojave desert is the flat area to the right. The yellow dots trace a plunging syncline of Miocene sandstones and conglomerates. The mountains to the left are the San Gabriel mountains. The faults, the fold and the mountains are all part of the plate boundary. The associated deformation continues all the way to the coast and offshore. Thus the mobile zone that is the plate boundary here is quite wide. The Landers EQ tells us that deformation also occurs to the east of the San Andreas within the Mojave desert. Pallett Creek is also evident as the thin dark green riparian corridor in the right central portion of the image. If you look carefully (at the full image) you can detect the dextral offset of the drainage. This is also the site where trenching analysis was done to work out the history of recent large earthquakes on this stretch of the San Andreas fault.
San Andreas fault and the San Francisco Bay Area - good site, with detailed information and plenty of additional links.
Remember that there is plenty of evidence that a subduction zone used to exist off the coast of California (blue schists and batholiths). How did it change into a transcurrent plate boundary?
In the greatly simplified map view and velocity triangle diagram above you can see both the utility of using velocity triangles and the answer to the above question. We know that subduction was occuring off the west coast between North America and the Farallon plate. The Farallon plate was and is being subducted underneath North America. This process is still going on further north off the coast of Washington state. Fitch fault structures indicate that the motion was dextral oblique, as represented by the F vs. NA vector above. If we add to this the spreading vector of the Pacific vs. the Farallon, then by default the Pacific is moving dextrally with respect to North America. So, when the ridge gets subducted the motion across the boundary is strike-slip.
Image from USGS showing details of history of how the subduction of the spreading ridge from the USGS. Diagram from http://geomaps.wr.usgs.gov/socal/geology/geologic_history/san_andreas_history.html .
History of motion on the San Andreas fault:
Kinematics and character change along strike
In this remarkable view of the southern portion of the San Andreas plate boundary the large scale locking bend and releasing bends can be seen. Image source USGS site - http://earthquake.usgs.gov/research/salton/.
San Francisco locked area.
Anomalies associated with the San Andreas fault?
What needs to be explained?
SAFOD = San Andreas Fault Observatory at Depth - http://earthquake.usgs.gov/research/parkfield/safod_pbo.php . Much new info came from the Cajun Pass drill hole along the San Andreas.
Some of the details of the SAFOD drill hole. Image source: https://commons.wikimedia.org/wiki/File:San_Andreas_Fault_SAFOD_Project.jpg .
Decoupling along the San Andreas - Mount, V. S. & Suppe, J., 1987, State of stress near the San Andreas fault: Implications for wrench tectonics; Geology, v. 15, p. 1143-1146.
San-Andreas low-strength surfaces, produced by traces of weak phyllosilicates - Diane E. Moore and M.J. Rymer, 2007, Talc-bearing serpentinite and the creeping section of the San Andreas fault; Nature, 448, 795-797 - https://pubs.er.usgs.gov/publication/70030842 .
This is a photograph of recent Pallett Creek alluvium that has been trenched. Gravels, sands and some muds with dark charcoal layers can be seen. Note the small fault that cuts the sediment and the tilting of the sediment. This locality is directly astride the San Andreas fault, and this small rupture represents movement and an earthquake that has occurred since they have been deposited. By analyzing what layers are cut by a fault strand and what layers truncate a fault strand the recent history of motions can be worked out.
Course materials for Plate Tectonics, GEOL 3700, University of Nebraska at Omaha. Instructor: H. D. Maher Jr., copyright. This material may be used for non-profit educational purposes with appropriate attribution of authorship. Otherwise please contact author.