Multichannel Sparse Deconvolution of Teleseismic Receiver Functions With f − x Preconditioning

Abstract

Teleseismic receiver functions (RFs) are an effective tool for structural imaging due to their simplicity and sensitivity to changes in rock elastic properties. Generally, RFs are estimated through trace‐by‐trace deconvolution. This individualized approach ignores the lateral coherency among RFs from neighboring raypaths, which could lead to rougher or incorrect seismic sections due to the presence of noise. This study presents a multichannel sparse deconvolution method that takes advantage of the cross‐trace coherency of RFs at individual stations for stable and accurate imaging outcomes. The proposed algorithm incorporates sparse inversion and frequency‐space prediction filters, which facilitate the retrieval of high‐resolution and spatially continuous conversion energy. Rigorous testing using synthetic and actual data typically suggests the higher robustness of multichannel deconvolution over single‐channel deconvolution, especially under low signal‐to‐noise ratios. The noise‐resistance property of the multichannel approach offers new opportunities to increase the volume of usable, high‐quality data for receiver function analysis.