Where the San Andreas goes to get away from it all

by Chris Rollins, Caltech

If one were to start at the Golden Gate Bridge and follow the San Andreas Fault down the San Francisco peninsula, passing Silicon Valley and garlic fields in Gilroy and entering San Benito County to the southeast, a curious match might be observed between the behavior of the fault at depth and the behavior of life at the surface.


San Francisco, California
San Francisco, California


Motion along a fault such as the San Andreas falls somewhere between two endmember behaviors. The first is stick-slip behavior, where the two sides of the fault become stuck together, gradually building up strain for decades or centuries as the surrounding crust bends around the stuck portion, and then this strain is suddenly released in a matter of seconds as the two sides slip past one another – an earthquake. The second is creep, where the two sides of the fault gradually grind past one another without much resistance, with little to no strain accumulating and (presumably) no earthquakes.

In the Bay Area, geodesy (monitoring of the gradual motion of the Earth’s surface) tells us that the two sides of the San Andreas are as stuck as cars in the morning commute in Silicon Valley, and the system is accumulating strain as the North American and Pacific Plates move past one another on the global scale but are unable to do so locally along the stuck fault. Some portion of this strain will presumably be released in a matter of seconds in a future earthquake, as it was in the great 1906 M=7.8 San Francisco earthquake and the 1989 M=6.9 Loma Prieta shock, and then this portion of the fault will presumably become re-stuck and begin accumulating seismic strain for the next earthquake – stick-slip behavior.

Proceeding southeast from the Bay Area, as the congestion and hubbub of life in tech country are replaced by two-lane roads and rolling hills, the traffic jam on the fault also curiously disappears: around the small town of San Juan Bautista, the San Andreas transitions away from stick-slip and enters what is known as the “creeping section.”


Temblor map of San Andreas Fault
Temblor map of San Andreas Fault.


The creeping section


Here, for approximately 100 miles of fault, the signature of strain accumulation in geodesy (technically an arctangent function in surface velocity across the fault) somewhat disappears and no evidence of large-magnitude earthquakes has been found. The North American and Pacific plates appear to slide past one another comparatively quietly as the fault parallels the quiet and little-traveled State Route 25, long called the Airline Highway as it lies beneath the principal air corridor between San Francisco and Los Angeles. (Both faults and airlines, by nature, take the shortest distance between two points.) The only tourist draw along the Airline Highway, about an hour south of San Juan Bautista, is the eastern approach to Pinnacles National Park. The park’s spectacular rock spires and caves are the product of a volcanic complex that was split in two by the relative motion along the fault; the western half of the complex is Pinnacles, and the eastern half are the Neenach Volcanics, located hundreds of miles to the southeast along State Route 138 just north of Los Angeles. The offset between the two is in fact one of the best measurements of the cumulative slip along the San Andreas system.


Pinnacles National Park
Pinnacles National Park


The creeping section follows the Airline Highway from San Juan Bautista – not coincidentally the southeastern endpoint of the 1906 earthquake’s rupture – past Pinnacles East to the highway’s southern terminus at State Route 198 (which itself is subtly referenced as the highway that gets split by the giant chasm in “San Andreas”) and from there extends through even remoter country to the small ranching town of Parkfield. There, amidst a local phenomenon of magnitude 6 earthquakes that has itself long captured the attention of the geophysics community, the fault’s behavior returns to stick-slip and one passes into the southern section of the San Andreas, home to more great earthquakes such as the M=7.9 Fort Tejon tremor.


Location of last Monday's magnitude 4.6 shock and previous earthquakes along the creeping section, which extends southeast from San Juan Bautista (top center).
Location of last Monday’s magnitude 4.6 shock and previous earthquakes along the creeping section, which extends southeast from San Juan Bautista (top center).


Earthquakes on the creeping section


Although no evidence of large earthquakes such as the 1857 and 1906 shocks has been found in the creeping section, smaller earthquakes up to about magnitude 5 have occurred there in the instrumental period. One of them rattled the central California coast last Monday morning, a magnitude 4.6 earthquake that originated north of Pinnacles near the small town of Paicines. Although the earthquake was fortunately too small to cause any damage or injuries, its occurrence is eye-catching for multiple reasons. The first is that it may have actually ruptured a portion of the creeping section that may be somewhat locked, more like the Bay Area and southern sections – called an asperity. Monday’s earthquake occurred about one-quarter of the way from San Juan Bautista to Parkfield, close to the location where Jolivet et al [2015, their Fig. 4] inferred an asperity on the creeping section from geodetic data.

The second eye-catching aspect of this earthquake is that it brings to mind a debate about why the creeping section is creeping. It is commonly thought that the fault properties of the creeping section favor stable sliding and essentially kill off earthquakes that propagate into it – as evidenced by the 1906 earthquake’s end at San Juan Bautista. These same fault properties should in theory prevent earthquakes from nucleating (originating) in the creeping section, or at least from reaching magnitudes larger than 5 or so if they do. However, the shallowest portion of the Japan Trench offshore east Honshu was thought to have the same earthquake-killing fault properties, and to slip gradually and stably, until it slipped 50 meters in a few seconds during the 2011 M=9.0 Tohoku-oki earthquake, directly causing the devastating tsunami that followed. Following this, Noda and Lapusta [2013] used an advanced computational method to simulate an earthquake originating on a nucleation-friendly stick-slip section of a fault and propagating into a neighboring stable-sliding section. They found that if the models accounted for the shear heating of pore fluids in the fault, a surprising behavior resulted: although most earthquakes did indeed die off at the transition into the stable-sliding section, one in every few propagated through the stable-sliding section with extremely large slip – a possible analog to the 2011 Tohoku-oki earthquake. The question in California, if this hypothesis is true, is whether an earthquake rupturing towards San Juan Bautista from the north, such as the 1906 San Francisco earthquake, could rupture clean through the creeping section and continue past Parkfield into southern California as a result of the same mechanism.

Perhaps even a weaker earthquake nucleating in the middle of the creeping section, such as Monday’s 4.6 temblor, could break through to one side and keep going. Future studies will help pin down whether these mechanics may be applicable to the creeping section. In the meantime, one can take some comfort that excavations of the San Andreas Fault along the creeping section have revealed no evidence of past large ruptures [Cashman et al, 2007]. Monday’s earthquake, however, is a reminder that the creeping section is not always as quiet as the highway it parallels.



Cashman, S., et al (2007), “Microstructures developed by coseismic and aseismic faulting in near-surface sediments, San Andreas Fault, California,” Geology 35 (7), 611-614.

Jolivet, R., et al (2015), “Aseismic slip and seismogenic coupling along the central San Andreas Fault,” Geophysical Research Letters 42.

Noda, H., and N. Lapusta (2013), “Stable creeping fault segments can become destructive as a result of dynamic weakening,” Nature 493, 518-521.

Harris, R. A. (2017), Large earthquakes and creeping faults, Rev. Geophys., 55,
169-198, doi:10.1002/2016RG000539.