Joint inversion of gravity and gravity gradient data based on Cokriging method with the threshold constraint

Gao XiuHe,Zeng ZhaoFa,Sun SiYuan,Yu Ping
DOI: https://doi.org/10.6038/cjg2019L0622
2019-01-01
Abstract:The conventional Cokriging inversion of gravity and gravity gradient data has the advantages of good anti-noise performance and being easy to add prior information. And this method can identify the position of the anomaly body and to restore its basic shape. But the inversion image of this approach is smooth and its resolution is low which are due to the density covariance matrix estimated by the conventional method are of global divergence and stationary. So the performance of the density covariance matrix has to be improved for getting focusing density distribution by the Cokriging inversion. Firstly, this paper deduces the theoretical formula of density covariance matrix, which proves that the density covariance matrix is non stationary and focused when the theoretical model is compact. In order to break the global stationary and divergence of the conventional covariance matrix, we set a density threshold to deal with the covariance matrix and update the covariance matrix to iteratively perform Cokriging inversion until the focused results are obtained. Then we set two kinds of forward models to testify our methods, showing the recovered models match with the forward models. Finally we use our method to process the gravity and gravity gradient data from the Vinton salt dome, and the inversion results are in good agreement with the known geological structures.
What problem does this paper attempt to address?