A combining regularization strategy for the inversion of airborne time-domain electromagnetic data

Abstract

Airborne time-domain electromagnetic data are usually over-sampled along the flight line. As the footprint of the ATEM method is larger than the interval between two soundings, inversion results cannot vary sharply in the lateral direction. The results of traditional one-dimensional inversion may be discontinuous. Although the popular lateral and spatial constraint inversion method can obtain a continuous quasi-2D and quasi-3D model respectively, these methods need to consider several soundings simultaneously, and their computational cost is relatively high, in comparison to inversion with only one sounding, for large ATEM data sets.To enhance the smoothness and continuity of the inversion result in the vertical and lateral directions, and simultaneously reduce the computation cost, we present a combining regularization strategy that can invert all ATEM soundings independently. The inversion strategy applies three regularization terms simultaneously. First, the inversion results of the previous sounding are used as a lateral constraint to strengthen continuity in the lateral direction. Second, the vertical model roughness is used to guarantee smoothness in the vertical direction. Finally, the results of conductivity-depth imaging are taken as a reference model to improve the inversion convergence rate.The combining regularization strategy was tested on synthetic data to analyze the influence of the first and second derivatives of vertical model roughness and that of the weight factors of the three regularization terms. The experimental results on synthetic data showed that the smoothness of the inversion result in the vertical direction is better with the second derivative of vertical model roughness than with the first derivative, and the proposed combining regularization strategy can effectively improve the continuity and smoothness of the inversion result in the horizontal and vertical directions, which is better than separately applying each regularization technique. Meanwhile, the combined scheme using conductivity-depth imaging results as reference model can not only obtain the inversion results close to the real distribution of media, but also improve the computational efficiency. In addition, the inversion results on field data acquired from Leach Lake Basin, Fort Irwin, California, verify the practicability of the proposed inversion strategy.