Deep creep as a cause for the excess seismicity along the San Jacinto fault

Since 1890, the San Jacinto fault in Southern California has been the site of eleven earthquakes of moderate magnitude (6 < M < 7) and tens of thousands of small earthquakes, but none of large magnitude1,2. This activity contrasts sharply with the seismic quiescence of the nearby southern San Andreas fault. Although this fault slips at a rate higher than that associated with the San Jacinto fault—23–27 mm yr−1 versus 12–22 mm yr−1 (refs 3, 4)—it has produced very few earthquakes and no moderate or larger events in historical times. Here I use recent seismic and geodetic data to reveal that at depths of 10–17 km within the seismogenic (brittle) crust, the San Jacinto fault is creeping and releasing elastic strain by many small earthquakes. As a result, the accumulation of strain along this fault occurs mostly in its upper 10 km; moderate earthquakes are likely to be sufficient to release such strain. In contrast, the southern San Andreas fault accumulates elastic strain throughout its vertical extent in the seismogenic crust, which will most probably be released by stronger earthquakes. Earthquakes are the most familiar and best studied mechanism of crustal elastic strain release, but not the only one. Other strain release mechanisms include postseismic after-slip, slow slip events and aseismic creep. These mechanisms indicate that a significant amount of elastic strain can be released steadily or episodically in between moderate or large earthquakes. Here I present another mechanism of elastic strain release, deep creep, occurring within the lower section of the seismogenic crust at depths of 10–17 km. This study relies on quantitative estimates of elastic strain accumulation and release along the southern San Andreas fault (SSAF) and the San Jacinto fault (SJF) as derived from geodetic and seismic observations and summarized in Table 1. The distribution and rate of strain accumulation are estimated from geodetic observations of interseismic crustal movements combined with dislocation models (see the Methods section). Elastic strain release is best determined from seismic observations, whereas aseismic strain release can be determined indirectly using geodeticmeasurements. Interseismic crustal movements in Southern California indicate a high rate of shear strain accumulation along the active fault segments of the San Andreas Fault system5 (SAFS). South of the Big Bend, where the SAFS branches into three parallel segments, the SSAF, SJF and the Elsinore fault, high-velocity gradients occur only along the SSAF and the SJF (Fig. 1 and Supplementary Fig. S1). Using dislocation models, various studies estimated the strain accumulation rate along the SJF at 12–22mmyr−1 and along the SSAF at 23–27mmyr−1 (Table 1 and Supplementary Table S1). Another important parameter calculated by these models is the locking depth, which is found to be 11± 2 km for the SJF and is 16±2 km for the SSAF. Elastic strain release can be estimated from earthquake catalogues of recorded seismicity and historical documents. The most