Improving receiver function imaging with high-resolution Radon transform

The receiver function (RF) method is one of the most widely used passive seismic approaches for imaging subsurface structures from shallow sedimentary basins to the deep Earth's interior. RF is typically computed with a deconvolution operation that enables separating structural response from source-time function embedded in teleseismic wavefields. However, seismic recordings of temporary arrays are often contaminated with strong random noise and even amplitude bias due to poor installation conditions. Both factors can cause instability to deconvolution and severely degrade the accuracy, thereby the imaging quality, of the RF method. This underscores the importance of effective denoising algorithms in RF processing. In this study, we explore the application of high-resolution Radon transform (RT) to improve the conventional RF imaging workflow. Contrary to the commonly implemented post-processing (i.e. after deconvolution) noise suppression schemes, we introduce the RT to data pre-processing (i.e. before deconvolution). This method seeks a sparse representation of teleseismic wavefields in the tau-p domain by iteratively solving a least-squares minimization problem with the conjugate gradient algorithm. Synthetic test with a 2-D step-Moho model shows that non-linear phase arrivals including incoherent noise and diffraction energy are effectively removed in resulting RFs, with the signal-to-noise ratio increased by as much as similar to 8 dB. Real data experiments using the Hi-CLIMB network in the Tibetan Plateau demonstrate the superior performance of the proposed workflow in regularizing the wavefield and improving the coherence of converted phases across the recording array. Consequently, imaging results of common conversion point stacking using a single and a group of teleseismic events both recover subtle converted phases from the Moho and potential lithospheric discontinuities that are otherwise obscured by noise arising from the conventional processing workflow. This study highlights the necessity of wavefield regularization in the RF method and calls for improved data processing techniques in array-based seismic imaging.