The effective signal reconstruction of CSEM data with strong interference using inter-station transfer function

Abstract

The impact of spatial and temporal superposition with various noises in strong interference areas is significant, leading to severe and complex influences for electromagnetic exploration. Previous methods for signal processing of Controlled-Source Electromagnetic Method (CSEM) data have primarily focused on single-channel analysis, thereby neglecting the correlation between the channels and resulting in unnecessary errors. Based on the synchronous observed data and remote reference theory of Magnetotelluric (MT), we put forward a data processing method using the inter-station transfer function to reconstruct the effective signal of CSEM data. In contrast with the conventional CSEM denoising approach, the seasonal-trend decomposition based on Loess (STL) is first used to remove the long-period nonlinear baseline drift and large-scale step noise of the strongly interfering signal. Secondly, we need to select a high-quality reference signal from the synchronous data and data fragments without noise of local signal by the ratio of variance (ROV), followed by a stable inter-station transfer function is calculated. This function and the synchronous reference signal are employed to achieve the high-precision reconstruction of the effective signal from the strongly interfering local signal, thereby enhancing the signal-to-noise ratio (SNR) of the observed data. The validity and practicability are verified by the simulated and measured data of Wide Field Electromagnetic Method (WFEM). The results demonstrate that the method we proposed not only considers the correlation between synchronously multi-channel data but also effectively suppresses strong electromagnetic interference and the extra field work is not necessary, which has a broad range of potential applications in the context of strong interference areas.