Cross-gradient based 3D sparse joint inversion using smoothness constraint for gravity and magnetic data of Jalalabad Hematite ore

Abstract

Cross-gradient based joint inversion method is one of the very useful methods for reduction of non-uniqueness and elimination of lacks of reconstructed models using separate inversion. This method have also some deficiencies against its advantages. It needs too much time when the dimension of problem is large and sometimes it's not runnable. In this paper data space method, as a fast procedure for inversion of some matrices in inversion process, was represented and cross-gradient based joint inversion algorithm supported by this method was developed in MATLAB software. To investigate the capability of code, a synthetic model include three cubes with various dimensions and depth extend was used. At first the calculated gravity and magnetic anomalies of this synthetic model were separately inverted using data space inversion method. Then these synthetic gravity and magnetic data were jointly inverted using the improved 3D inversion algorithm along with smoothness constraint. The obtained results showed significant improvements in the output model with the same conditions of separate data inversion. These improvements were in delineation of location and depth estimating aspects and also capability of code performance to recover such number of model’s parameters. After validating of results, the proposed routine was applied for 3D joint inversion of the gravity and magnetic data of one Hematite area around Jalalabad iron mine in Zarand city (Kerman province). Comparison of the inversion results with those obtained by drilling, validates correctness and ability of the presented algorithm even when we encounters with deficiency of data.