Rayleigh Wave Phase Velocity Maps at Regional Scale Inferring from SPAC of Ambient Noise at a Dense Array: A Case Study in Northeastern Tibetan Plateau

The SPAC (spatial auto-correlation) of the microtremor is widely used to infer a shallow structure at local scale. On the regional scale, based on the dispersion curves extracted from the interstation NCFs (noise cross correlation functions), the 2D lateral variation of the surface wave velocity is usually established via pure-path inversion. The retrospective study on seismic interferometry shows that the SPAC of the microtremor in the frequency domain is equivalent to the NCFs in the time domain. Therefore, benefitting from the deployment of large and dense networks, the array-based SPAC technique can be used to process the NCFs of the subsets of the network. The lateral variation of the velocity can thereby be obtained directly without pure-path inversion. In this study, an imaging scheme of SPAC with moving subarray is applied successfully to directly obtain the phase velocity maps of Northeastern Tibetan Plateau at regional scale using the interstation NCFs from ChinArray (Phase II). The array is divided into > 300 subarrays, and the dispersion curves of each subarray are then extracted by SPAC technique. The 2D phase velocity maps between 7 and 35 s are obtained by combining the results of all subarrays. The results are in good agreement with those obtained by classical surface wave tomography. Compared with beamforming method, the velocity at long periods measured by SPAC suffers less bias caused by the array configure at long periods. It is proved that the array-based SPAC technique with moving subarray can be used to image the earth crust structure at regional scale with quite stable measurements.