De-noising receiver function data using the Seislet Transform

Abstract

The converted wave data (P-to-s/S-to-p), commonly termed as receiver functions, contain noise of various origins. Such noises may influence the modeling and may sometimes lead to over interpretations of the data. In order to suppress noise, we use a robust sparsity enhancing tool, that is, the Seislet Transform, to process receiver function data by applying regularization in the seislet domain. The transform utilizes the multiscale orthogonal basis and the basis functions are oriented along the dominant seismic phases following local linearity. The inversion results of both the synthetic and field examples from the Hi-CLIMB network and station HYB from the Indian shield show an excellent performance over the original data sets.