Analysis of Foreshock Sequences in California and Implications for Earthquake Triggering

Abstract

We analyze foreshock activity in California and compare observations with simulated catalogs based on a branching aftershock-triggering model. We first examine foreshock occurrence patterns for isolated \(M \ge 5\) earthquakes in southern California from 1981 to 2011 and in northern California from 1984 to 2009. Among the 64 \(M \ge 5\) mainshocks, excluding 3 swarms and 3 doubles, 53 % of the rest are preceded by at least one foreshock within 30 days and 5 km. Foreshock occurrence appears correlated with mainshock faulting type and depth. Foreshock area is correlated with the magnitude of the largest foreshock and the number of foreshocks, however, it is not correlated with mainshock magnitude. We then examine the occurrence pattern of all seismicity clusters without a minimum magnitude requirement, and the possibility that they are “foreshocks” of larger mainshocks. Only about 30 % of the small clusters lead to a larger cluster. About 66 % of the larger clusters have foreshock activities, and the spatial distribution pattern is similar to \(M \ge 5\) mainshocks, with lower occurrence rates in the Transverse Range and central California and higher occurrence rates in the Eastern California Shear Zone and the Bay Area. These results suggest that foreshock occurrence is largely controlled by the regional tectonic stress field and fault zone properties. In special cases, foreshock occurrence may be useful for short-term forecasting; however, foreshock properties are not reliably predictive of the magnitude of the eventual “mainshock”. Comparison with simulated catalogs suggest that the “swarmy” features and foreshock occurrence rate in the observed catalogs are not well reproduced from common statistical models of earthquake triggering.