Coincidence of Temporal Changes in Earthquake Clustering and Geodetic Velocity in the Basin and Range, western United States

Analysis of a catalog of all M ≥ 4 earthquakes in the Great Basin from 1932--2006 reveals that, in general, earthquakes are temporally more clustered than would be expected from a random distribution of independent events, even after removal of foreshocks and aftershocks from the catalog. These results are consistent with prior evaluation of foreshock probabilities in the western Great Basin, where the occurrence of a M ≥ 4.0 earthquake increased the probability of a M ≥ 5.0 earthquake by several orders of magnitude (Savage and dePolo, 1993). Both analyses suggest that many faults in the Great Basin are at or near critical stress for failure, and that triggering of seismic events may occur from local or even remote events (e.g. the 1992 Landers earthquake; Anderson et al., 1994). The exception to this pattern is in the distribution of regional seismicity from 1999--2004. During this five year period, temporal clustering of earthquakes was reduced, and a strong periodic element was observed in the temporal distribution of seismicity. This time period coincides with a sub-regional change in the geodetic velocity field of the western Great Basin, where 10 permanent GPS sites of BARGEN network west of the Nevada-Utah state line accelerated progressively eastward towards stable North America (Davis et al., 2006). One interpretation of the change in geodetic velocity is that it represents a regional slow earthquake, analogous to events that have been geodetically observed in subduction zones. One effect of such an event would be to release crustal strain across the western Great Basin, consistent with the observed decrease in temporal clustering of seismicity during the event. Since early 2005, geodetic velocities have accelerated westward and returned to pre-1999 levels. The number of M ≥ 4 earthquakes in the Great Basin since 2004 is too few to draw any conclusions about the present distribution of seismicity. Taken together, however, the temporal changes in geodetic velocity and earthquake clustering suggest that the time-dependence of seismic hazard may change over decadal timescales.