No abstract

A network of eight University of Texas ocean bottom seismographs (OBS) operated for 6 weeks in 1978 about 50 km offshore of Adak Island, Alaska, and nearby islands. In 1979 a similar network of nine instruments was deployed for 7 weeks farther offshore within and up to 100 km seaward of the Aleutian trench. For shallow earthquakes on the outer trench slope, for shallow earthquakes in the thrust zone, and for intermediate‐depth events we have analyzed the OBS and island‐based network data and evaluated the island network's capabilities for earthquake detection and location and for focal mechanism determination. Our three major conclusions are presented. The first concerns shallow earthquakes on the outer trench slope. In 1979 about 30 earthquakes occurred within the Aleutian trench and up to 60 km seaward of the trench axis. The island network located none of these events and detected P phases for only three of them. Ray tracing shows that the islands lie in a geometric shadow zone for events on the outer trench slope. The best located events are shallower than 20 km and exhibit first motions consistent with normal faulting. Several authors have suggested that these events are caused by bending of the oceanic lithosphere at the outer rise prior to subduction. If so, then the event locations reported here show that the bending stresses exceed the strength of lithosphere only in a narrow zone extending about 10 km landward and 60 km seaward of the trench axis. The second conclusion concerns shallow earthquakes in the thrust zone. Epicenters determined by island stations alone are virtually identical to epicenters determined using data from both island and OBS stations. The locations are similar whether they are determined using flat‐layered velocity models or a more realistic model and a ray‐tracing scheme. Nearly all the earthquakes detected by the OBS network are also recorded by island stations. Three composite focal mechanism solutions that use data from both island and OBS stations have P axes parallel to the direction of plate convergence. The thrust zone events are separated from events on the outer trench slope by a gap in seismicity about 50 km wide situated on the inner trench slope. The third conclusion concerns earthquakes deeper than 70 km. Epicenters determined using island network stations alone lie 10 to 80 km south of those determined using OBS and island stations, with the differences between epicenters depending both on event depth and on the velocity model used. Ray tracing with a realistic velocity model shows that for events at 200‐km depth locations are about 80 km north of locations determined from island data alone. These results do not support the observed increase in the dip of the Benioff zone beneath 100‐km depth that is suggested by locations determined from island network data alone using a flat‐layered velocity model.

No abstract

No abstract

Application of the method of Joint Hypocenter Determination (JHD) to carefully determined arrival time data from a small network of sensitive and well‐distributed local and teleseismic stations provides very precise locations of deep earthquakes. The precision of locations attained by this method—the relative locations are good to about 5 km—is limited mainly by the accuracy and precision of the measurements of arrival times. Application of the technique to a small and very active part of the deep earthquake zone of the Tonga island arc (depths near 600 km) reveals probable fault planes associated with the occurrence of deep earthquakes. The spatial distribution of precisely relocated hypocenters is closely related to the orientation of the focal mechanism solutions of two deep earthquakes. Hypocenters cluster in a plane parallel to one of the nodal planes for each of the two solutions examined. The planar features are yearly vertical and have linear dimensions of about 40 km. In one case the shocks defining the fault plane also tended to cluster in time near the occurrence of the largest event. These and other data suggest the existence of several distinct fault planes in the area studied. Our study is being extended to examine the structure of the entire Tonga deep seismic zone.