Micro-earthquakes have been systematically recorded with magnitudes down to -1.3 at more than 60 sites along the San Andreas fault system in southern California during intervals of 2 days to 1 year, representing more than 35,000 hours of usable records. Eight trailer-mounted instruments were operated with peak gains of 4-8 million at 20 cps with noise levels averaging about 0.1 mu amplitude of ground motion.
Observed micro-earthquake activity varies from virtually nil along the central section of the San Andreas fault to more than 75 shocks daily in the Imperial Valley. Quietest is the 300-km segment between Cholame and Valyermo; more than one year of recording at Lake Hughes indicates an average of only one micro-earthquake within 24 km every nine days. Activity increases northward from Cholame toward Hollister, and southward it increases abruptly near Valyermo and continues high along major branches of the fault southeast into Mexico, with the exception of the Banning-Mission Creek fault southeast of Desert Hot Springs. Most areas where regional strain or fault creep have been demonstrated by geodetic measurements are also areas of high micro-earthquake activity. Existence of an area of minimal micro-earthquake activity within a broad region of active tectonism, and indeed along the very segment of the fault that broke in the great 1857 earthquake, suggests that short-term micro-earthquake activity is not necessarily positively correlated with long-term activity and with earthquake hazard, and in some areas the relationship may be inverse. However, areal distribution of micro-earthquake activity is grossly similar to that of larger earthquakes (M ≧ 3) during the past 29 years, and in many areas micro-earthquake activity can be approximately predicted by extrapolation of 29-year recurrence curves based solely on larger earthquakes.