Magnetotelluric multi-reference array data processing algorithm

Abstract

Rational use of multiple remote reference data is an important way to improve the data quality of MT method. Based on the electromagnetic array data processing model, a data processing algorithm with multiple reference stations is proposed. Firstly, reference channels are selected from multiple remote stations to construct the natural field dominated reference data array; then the polarization parameters of natural sources are extracted from the reference data array by robust principal component analysis method; the parameters are then applied to the target array, which constructed by data from the stations to be processed, to estimate the target tensor impedances by robust estimation method. Finally, the effectiveness of the proposed algorithm is discussed by comparing it with the robust least squares and conventional remote reference methods, using a case study in east China. The results show that the new scheme is an effective strategy by using the multiple reference channel data, which can improve the quality of MT data processing in noisy environment and provide a reasonable correction for MT “dead band” distortions.