Shear velocities, Q , and the frequency dependence of Q in stable and tectonically active regions from surface wave observations

Abstract

Velocities and attenuation coefficient values of Rayleigh waves within the South American continent and Indian subcontinent have been determined in order to study velocity structure and Qβ structure, as well as their lateral variation, in those regions. Single-station and two-station data are combined to yield dispersion and attenuation data from periods of a few seconds to periods between 60 s and 100 s for each region. Both types of data are inverted using differential inversion schemes. This inversion process is found to sometimes allow the location of velocity discontinuities in the models directly from the dispersion data. Qβ models obtained from the inversion process are not constrained to have high values at large depths, a problem with earlier methods used to invert surface-wave attenuation data. The shear-velocity models for the Indian shield and eastern South America are similar to one another, and have somewhat slower velocities than those reported for shield regions of North America. The shear-velocity models for the Himalaya and Andes are similar to one another in crustal thickness but the velocities in the Andes are lower than those in the Himalayas in the uppermost mantle. A low-velocity zone can marginally be resolved in the upper mantle beneath the Himalaya, but not beneath the other regions in this study. Qβ models obtained for the stable regions of eastern North America and eastern South America exhibit similar Qβ values at mantle depths. The Indian shield, however, has much lower Qβ values, in both the crust and upper mantle, a result which may be related to the unusually high heat-flow observed for that stable region. Qβ for the tectonically active regions of western North America, western South America, and the Himalaya is considerably lower than Qβ for the stable regions throughout most of the crust and upper mantle. Minimum mantle values appear to be about 120 for stable regions and 40 or less for tectonically active regions. Thus, Qβ values in the upper mantle appear to vary regionally by a factor of 3 or more, whereas shear velocities in the same regions vary by no more than 13 per cent. The Qβ models obtained for the tectonically active regions adequately predict Q values reported from time domain measurements of Lg waves for those regions. The Q values for 1 Hz Lg waves predicted by those models of stable regions in eastern North and South America, and the Indian shield are, however, significantly smaller than the observed values for those regions. Our results, combined with reported values of Lg Q at 1 Hz, imply that Qβ in the crust is independent of frequency over the frequency range of this study in tectonically active regions, but varies with frequency in stable regions.