Two-dimensional Joint Inversion Of Magnetic And CSAMT Methods

Joint inversion is an approach which inverses data collected via multiple geophysical methods to mutually constrain models according to the relationship between identical or distinct physical property. Joint inversion has been a major trending of geophysics nowadays because of its superiority of reducing ambiguity of individual method, and suppressing interfering signal. In this paper, we apply CSAMT and magnetic method to explore joint inversion with magnetic by the research of international major geophysical joint inversion method.CSAMT is one of electrical prospecting method. We develop a 2D forward and inverse algorithm considering the artificial source, which can inverse the CSAMT data in full-zone (near zone, intermediate zone and far zone). To verify if the effect of magnetic anomalous body can be ignored during CSAMT inversion, we complete a 2D forward and inverse program in consideration of resistivity and magnetic suseptibility simultaneously for the first time, and by trial calculation of theoretic models we verify that the result of inversion is affected significantly by strongly magnetic anomalies at CSAMT operation frequencies. However, the observed data is so insensitive to magnetic suseptibility that during the inversion of both parameters, susceptibility model could not change significantly; as a result, we finally introduce magnetic inversion to solve this problem.As for magnetic 2D forward and inverse algorithm, choosing logarithm form of magnetic suseptibility as model parameter, which leaves logarithmic barrier method away from objective function; utilizing data space Occam's inversion that is generally used in electrical data processing to complete magnetic inversion program, which cuts down the computation cost on solving linear equations during each iterative, and saves computer's memory; in the case of this algorithm, applying depth weighting on model covariance matrix, we introduce depth weighing matrix in a different way from general method and it worked well.Last, for tiny association between resistivity and magnetic suseptibility, we apply cross-gradient function to constrain these two physical properties and completed joint inversion algorithm. By testing strongly magnetic models in different geologic configurations, the correctness of joint inversion program is verified and we draw a conclusion that it not only removes the effect of strongly magnetic suseptibility on CSAMT result, but also makes great improvement in magnetic inversion for complicated models.