Transfer functions between teleseismic data components

Abstract

SUMMARY Different data components of teleseismic waveforms are related by transfer functions that depend only on receiver-side structure. This is the common basis of a number of teleseismic techniques, including receiver functions and shear wave splitting analysis. Common trace misfits used in these analysis techniques are shown to be equivalent to band-limited comparisons of real and synthetic transfer functions. The data deconvolution used in receiver function analysis leads to reduced structural resolution compared to direct trace-based misfits such as cross-convolution, with direct transfer function modelling of a data trace having the particular advantage of a physically meaningful misfit. Having established that the intertrace transfer function contains all available structural information, the sensitivity of transfer functions to structure is examined for a series of teleseismic scenarios. Transfer functions for the teleseismic P coda show a strong sensitivity to shallow low-velocity structures such as sedimentary basins; the Sp precursors used in S receiver functions are less affected. Examination of transfer functions for shear wave splitting shows that response complexities occur at frequencies too high to be observable in teleseismic studies, and that the dominant control on the response is the splitting intensity.