We compare relocated, teleseismically recorded seismicity before and after the May 7, 1986, Andreanof Islands earthquake (Mw= 8) to the spatial and temporal distribution of seismic moment release during the mainshock. Relocated seismicity from 1964 through April 1987 clusters strongly along the main thrust zone, in similar locations for the pre- and post-mainshock periods.

We performed simultaneous iterative inversions of broadband teleseismic GDSN and NARS P-waveforms using the method of Kikuchi and Fukao [1987]. Fault plane parameters (e.g., strike, dip, depth, and extent of fault plane) consistent with regional tectonics and aftershock locations were specified. The results of the inversions are spatial and temporal maps of moment release on the fault plane. Most of the moment found by the inversion is localized in several regions of the fault plane. Resolution of the location and timing of the major subevents is good. The rupture was bilateral, with a highly variable rupture velocity. A shallow, updip region of the fault plane ruptured toward the end of the mainshock process.

The mainshock moment release tended to occur in regions of the fault plane in which no or few aftershocks or preshocks are located. This tendency suggests that regions of high moment release in the 1986 Andreanof Islands earthquake resulted from rupture of mechanically strong regions (i.e., asperities) on the fault plane, which exhibited little seismicity before or after the mainshock that ruptured them. We infer that the asperity regions were locked and quiescent before the mainshock, while surrounding seismically active areas did not store as much strain and slipped mostly aseismically.