An iterative laterally constrained inversion scheme for time-domain airborne electromagnetic data

Abstract

The airborne electromagnetic (AEM) methods have been widely used in mineral prospecting, groundwater exploration, oil and gas resource investigation, and for many other purposes. Although the technical advances in 3D inversion are impressive, it is often not feasible to invert the field data from large AEM surveys with the huge amount of data. At present, 1D inversion and the quasi-2D inversion method based on the 1D model with imposed lateral continuity have been applied to invert the field data in AEM industry. In this paper, we propose a laterally constrained inversion scheme for time domain AEM data, based on a lateral correlation which effectively achieves both vertical and lateral smoothness. Firstly, all points of a survey line are inverted independently rather than simultaneously which is computationally very efficient. Secondly, at each iterative computation at each survey point, the inversion results of adjacent survey points at this iteration are used as the reference model to improve lateral continuity and convergence. Finally, the iterative termination condition considers the total data residual of all soundings of a survey line instead of one survey point for improving the overall performance of the inversion results. The experimental results of synthetic and field data show that the proposed inversion scheme can achieve better inversion results with moderate computational cost.