Three-Dimensional Joint Inversion Of Gravity,Magnetic, And Magnetotelluric Data Based On The Cross-Gradient Constrains

We developed the 3D joint inversion of gravity, magnetic and magnetotelluric(MT)algorithm based on the cross-gradient constrains. We introduce the cross-gradient to the joint inversion of three kinds of physical properties for the very first time in order to complement advantages and characteristics of different geophysical data. We did some basic research of3 D inversion of MT, gravity and magnetic respectively: on MT, we studied the data-space 3D MT inversion algorithm, which can reduce the memory requirements and improve the computational efficiency. On gravity and magnetic inversion, we studied the algorithm using the smoothest model constrain, to which the depth weighing and the logarithmic barrier method are also introduced, for the sake of stabler model perturbation, more reasonable property range, and acceleration of the convergence. Numerical experiments with synthetic data were carried on to test the single inversion algorithms, and we did detailed analysis on the single inversion results.We also test the above-mentioned joint inversion algorithm with synthetic data, the results suggest that using cross-gradient to constrain resistivity, density and susceptibility does not rely on the quantitative relationship between different physical properties, and can constrain each other during the process of the joint inversion. The results also suggest that models obtained from joint inversion are more reliable than those from single inversion,which indicates the effectiveness and applicability of the method.