Lateral variations in upper-mantle structure around India as obtained from Gauribidanur seismic array data

Abstract

Summary. In order to understand the tectonic processes responsible for the motion of lithospheric plates, it would be desirable to resolve as much information as possible on the detailed vertical and lateral structure that may be present in the upper mantle. From the point of view of continental drift, the Indian subcontinent is of particular interest because of the manner that it has rapidly moved northwards and collided with the Asian continent. The Himalayan mountains to the north, the mid-oceanic ridges to the south and the earthquake belts surrounding the Indian plate all show that the uppermantle structure in that area must have some very significant lateral variations. The Gauribidanur medium aperture seismic array (GBA) in southern India is a very well located array for obtaining data for upper-mantle research. In this paper we present the results from a study of over 300 earthquakes (distance 14°-36° and azimuth 0°-360°) recorded at this array. Adaptive processing techniques were employed on the first 36 s of the short-period P-wavetrains in an attempt to not only determine lateral variations in velocity but also to locate and identify the positions of the triplications on the upper-mantle travel-the branches which are associated with the major phase discontinuities or high-velocity gradient zones of the upper mantle. Four seismic zones defined by the following azimuthal ranges were considered: (a) Himalayan (0°-90°), (b) Java trench (90°-180°), (c) Midoceanic ridge (180° -280°) and (d) Hindukush (280°-360°). Inversion of the data has revealed that there is a considerable evidence for a larger lowvelocity or low-Q layer along the oceanic paths as compared to the continental paths. Further evidence for the existence of ‘400-km’ and ‘650-km’ discontinuities was found with the exception of the Himalayan region. In this region the replacement of the latter discontinuity by a broad high-velocity gradient zone was much more consistent with the observed data. Consistent slowness anomaly patterns, presented in the form of an array diagram, were observed for the Java trench and the Mid-Indian oceanic ridge. The most significant feature of the azhuthal anomaly pattern was the distinct absence of any positive anomalies from events occurring in the Ridge. Although upper-mantle velocity measurements were obtained over all directions about the array, it was not possible to carry out multi-directional velocity-measurements in the same region. However, it is felt that this study will compliment any other future work on anisotropy around India. There is very little evidence from our analysis for the existence of anistropy at depths below 500 km.