ON RUPTURE ZONES AND GEODYNAMIC PROCESSES OF GREAT EARTHQUAKES ALONG THE HIMALAYAN CONVERGENT PLATE MARGIN

Abstract

This thesis is a compilation of six distinct yet complementary studies of great and moderate earthquakes along the Himalayan convergent plate margin. In the first study the extant largely qualitative evidence for ground level changes during the 1897 earthquake is reinterpreted to suggest that a mid-crustal detachment was ruptured over an area of 170x100 km2 by thrust faulting. The rupture was located under the western Shillong Plateau and the Brahmaputra valley and does not extend upto the Himalaya several tens of kilometres further to the north. Geodetic observations of coseismic ground elevation changes during the Kangra earthquake of 1905 suggest, in a second study, that the causative rupture had an area of 280x80 km2 in the low angle thrust fault type detachment under the Outer and Lesser Himalaya. The same data set was analysed by trial and error and more rigorous minimum norm inversion methods. The lower bound of maximum slip on the detachment was 5-7 m. Interseismic levelling observations along a line in the Himalaya of central Nepal were analysed in the third study. It is concluded that aseismic permanent deformation and recoverable earthquake generating strains are accumulating in the region simultaneously. The latter will be released either fully or partially in the next great earthquake of central Nepal Himalaya. The fourth study is similar but based upon much more limited observations in the Dehradun region. It is concluded that preparations for a great earthquake are in progress in this region also though its date of occurrence cannot be predicted. The fifth study pertains to modelling of stresses for the occurrence of great and moderate earthquakes along the detachment under Himalaya. Within the framework of classical rock mechanics, we are forced to conclude that either very high pore pressures or very low frictional coefficients or both should occur along the detachment. Finally a statistical study of seismicity of the Himalaya was carried out to see if the Uttarkashi earthquake of 1991 in the Garhwal Himalaya could be retrodicted from the available seismicity data for the Himalaya and southern Tibet. The K function of CN algorithm proposed by Keilis-Borok and group to depict the spatio-temporal variation of seismicity was found to be a sensitive parameter for this purpose.

Regarding seismic hazards along the Himalayan convergent plate margin, we argue in several of above studies as follows. Firstly, the length of the Himalaya which has released strain through the four great earthquakes of the last 100 years is probably much less than considered by recent investigators. Secondly, the geodetic evidence reveals that earthquake generating strains are accumulating at least along two segments and probably throughout the Himalaya. Thirdly, the detachment lying at depths between 5 to 17 km beneath the Outer and Lesser Himalaya is probably the nearest active fault for most dam sites in the region. Two cumulative impressions from these studies are as follows. Firstly, the data analysed here are consistent with plate tectonics hypothesis as applied to the Himalayan convergent plate margin. Secondly, seismic hazards in the Himalaya and the adjoining Indo-Gangetic plains are much more severe than is being acknowledged by most Indian seismologists and civil engineers.