SPATIAL-DISTRIBUTION OF EARTHQUAKES IN AFTERSHOCK ZONES OF THE GARM REGION, SOVIET CENTRAL-ASIA

We examined the spatial distribution of earthquakes in aftershock zones of eight M4.5 to M6.3 earthquakes that occurred during the period January 1969-February 1987 in the Garm region of Soviet Central Asia. This seismically active region is a part of the collisional plate boundary between the Indian and Eurasian plates. Spatial patterns in earthquake distribution were studied using the technique of pair analysis, which is based on a quantitative comparison between observed and randomly simulated distributions of interevent distances. Similar to our previous findings from other seismically active areas of various scales and tectonic settings, we show that the spatial distribution of earthquakes in all of the aftershock zones studied is characterized by significant non-random features. Most prominent of the spatial patterns observed is spatial clustering at short interevent distances. The degree of non-randomness is usually higher for the aftershocks than for the background seismicity (i.e., the remainder of the catalogue, after space-time clusters have been removed). In some of the study areas the spatial distribution of aftershocks appears to be self-similar in magnitude, but in most cases is not self-similar with respect to the size of the study areas. More specifically, the seismic activity in the smallest subareas (of only few kilometres in size) exhibits a higher degree of spatial clustering than the activity in larger subareas. Furthermore, some of the spatial patterns observed change systematically in time. One such change is the gradual increase of the degree of pre-shock clustering as the main event approaches. Other temporal features are observed during the aftershock sequences themselves, when the degree of spatial non-randomness often increases to a peak value and then gradually decreases reaching a reduced 'background level' at some time after the sequences are over. Some of the temporal changes observed are similar to patterns previously identified in central California aftershock areas.