3D mixed-norm inversion of magnetic data using a log-barrier method

Three-dimensional inversion of magnetic data is highly non-unique because magnetic data have no depth resolution. To alleviate the non-uniqueness of magnetic inversion, a common approach is to inject additional geological or geophysical information into the inversion. Alternatively, one can use mixed-norm inversion, especially in greenfield areas where geological or geophysical information is not available. In this paper, we propose an algorithm for mixed-norm inversion of magnetic data using the log-barrier method. Synthetic tests show that the mixed-norm inversion can produce improved results in the absence of a priori information. The proposed algorithm is also applied to inversion of magnetic data collected over an iron–copper skarn deposit.