Crustal Structure in East Antarctica from Surface Wave Dispersion

Abstract

Summary The structure of the Earth’s crust and the upper mantle in East Antarctica are investigated using the dispersion properties of both the Love and the Rayleigh waves. The group velocities of surface waves between 15 and 70 sec periods are measured over several paths crossing Antarctica from Sandwich Arc and Drake Passage to stations of Wilkes and Mirny. Theoretical dispersion curves are computed for various crust and upper mantle models. The structure that fits the Rayleigh and Love wave data equally well is charactized by a 42 km thick crust and low (7.85 km/sec) sub-Moho velocities. The average thickness of the ice layer over East Antarctica, between Princess Martha Coast and Wilkes Station, is determined to be about 3 km. The general problem of non-uniqueness in group velocity inversion is also discussed. The dispersive properties of seismic surface waves have been successfully utilized to determine the structure of the crust and the upper mantle in many regions of the world. In Antarctica, however, the use of the technique has been hindered by the general lack of earthquakes located so as to provide purely continental paths between the epicentres and the few Antarctic stations equipped with long period seismographs. In determining group velocities across the Antarctic continent, paths with considerable oceanic portions had to be used, and errors due to uncertainties in proper corrections for the oceanic fractions of the paths have been introduced. Furthermore, computed epicentral distances and travel times have been less reliable in the Antarctic area than in other continents where higher station density makes the determination of epicentres and origin times more accurate. In this study an attempt is made to construct a crustal-upper mantle model of East Antarctica that fits data chosen so as to minimize the problems described above. Surface waves from several earthquakes in the same region were analysed. The waves had traversed nearly identical and mostly continental paths. The oceancontinent boundary is well-defined in the area where these paths cross it. Both Love and Rayleigh waves were recorded. The number of cases used should reduce the errors resulting from random inaccuracies in epicentre determination ; though a