Within 24 hr after the Landers earthquake, there were three magnitude 3.4+ events in western Nevada and an unexpected, widespread increase in the rate of small events. Based on combined catalogs for northern Nevada, southern Nevada, and southern California, and a model that assumes statistical independence of events in these regions, the probability of this happening in a 24-hr period by random chance is less than ∼10−12 per day. Therefore, there is high statistical confidence that the increased seismicity was triggered by the Landers event. Based on the statistical model, we develop a list of 227 earthquakes in the first 83 days following the Landers earthquake, each of which has no more than 10% probability of occurring by random chance. The suspect events are broadly distributed in regions that correlate with historical activity in the Great Basin. The events are not uniquely associated with known volcanic activity, or with zones that were previously active with microearthquakes or aftershock sequences. The magnitudes of the largest triggered events appear to decrease with distance. With time, the number of suspect events decreases at a rate comparable to the rate of decrease of aftershocks of the Landers and Big Bear earthquakes.
Three of the most significant triggered events that that occurred were as follows: Mina, 500 km from Landers, M 4, 36 min after Landers; Smith Valley, 590 km from Landers, M 3.4, 56 min after Landers; and Little Skull Mountain, 280 km from Landers, M 5.6, 22.3 hr after Landers. The evidence for triggering is particularly strong in the case of the Little Skull Mountain event, where an increased rate of microseismicity was evident as soon as small events could be identified in the coda of the Landers earthquake.
We evaluate the seismic history for the past 3 yr to understand what was unique about the Landers earthquake. This includes identifying the previous earthquakes most likely to have caused the strongest shaking in various frequency bands. Based on a simple screening model, the strains from the Landers earthquake were uniquely large at low frequencies, but at high frequencies they were exceeded frequently by small, more local events. Thus, we hypothesize that the cause of triggering is low-frequency dynamic strains, at periods of about 10 sec or greater, and that there is a regional threshold that must be exceeded, since previous events with calculated strains that were almost as strong were not causes of widespread triggering. This mechanism for triggering satisfies the criterion of being a relatively rare phenomenon, since it is likely to occur only when a large long-period wave is radiated into an area where strain has been building slowly toward the point where faults are unstable.