Kazi, W.M., 1990, Velocity changes in the central Aleutian Islands determined from the analysis of earthquake doublets using cross-spectral analysis method: University of Colorado, Boulder, Ph.D. dissertation, 187 p.
The occurrence of the May 7, 1986 Andreanof Islands earthquake of Mw = 8.0 provided an opportunity to study changes in the stress field. The Central Aleutian Seismic Network (CASN) database was found to contain several doublets i.e. pairs of earthquakes that occur at identical locations and yield similar waveforms at all recording stations. Several of these doublets span the big earthquake while others do not. The cross spectral analysis method (CSAM) was applied to doublets to determine small (≈ 5 to 10 millisec) variations in travel-time arrivals of individual phases in the doublet pairs. These 'delay' times were used to relocate these doublet events to a much higher accuracy ≈ 100 to 300 m) than in the routine CASN locations. The fractional S-velocity change exhibited strong azimuthal variation with values ranging from 0.7% in a direction mostly north-northwest of the source to no change at 50° to 6° east or west of north. The direction of maximum compression is to the north or northwest, the direction of the subduction of the Pacific plate under the North American plate, in the vicinity of Adak Island. The observed velocity changes suggest that contrasting periods of stress changes may have occurred in the region. It appears that periods of quiescence may conform to periods to 'high' or 'low' velocity changes. In the absence of viscoelastic effects, it appears that periods of quiescence conform to periods of little or no stress change (periods of no velocity changes) and vice-versa. This in turn suggests that quiescence does not occur due to a 'locking' of plates but rather due to a period of little or no stress changes in the quiescent zone. A map representing the spatial distribution of the velocity changes suggests that contours of equal velocity change usually parallel the general strike of the plate interfaces and that the gradient is perpendicular to this direction, being roughly towards the direction of maximum stress change (north-northwest). Generally, the contours appear to be more complicated along the eastern boundary of Adak canyon than elsewhere, thereby supporting the concept of Adak Canyon marking an important inhomogeneity in the thrust zone. (Abstract shortened with permission of author.)
Theses and Dissertations
