Abstract
Borehole gravity is increasingly used in mineral exploration due to the advent of slim-hole gravimeters. Given the full-tensor gradiometry data available nowadays, joint inversion of surface and borehole data is a logical next step. Here, we base our inversions on cokriging, which is a geostatistical method of estimation where the error variance is minimised by applying cross-correlation between several variables. In this study, the density estimates are derived using gravity-gradient data, borehole gravity and known densities along the borehole as a secondary variable and the density as the primary variable. Cokriging is non-iterative and therefore is computationally efficient. In addition, cokriging inversion provides estimates of the error variance for each model, which allows direct assessment of the inverse model. Examples are shown involving data from a single borehole, from multiple boreholes, and combinations of borehole gravity and gravity-gradient data. The results clearly show that the depth resolution of gravity-gradient inversion can be improved significantly by including borehole data in addition to gravity-gradient data. However, the resolution of borehole data falls off rapidly as the distance between the borehole and the feature of interest increases. In the case where the borehole is far away from the target of interest, the inverted result can be improved by incorporating gravity-gradient data, especially all five independent components for inversion.
We perform 3D joint inversion of surface and borehole data using cokriging. The examples presented demonstrate that the method is remarkable in its ability to include the geological information and physical property. Also, both depth and horizontal resolution of the recovered model can be improved by joint inversion.
Acknowledgements
The authors would like to thank Colin Farquharson for his detailed comments and suggestions, which have helped improve the manuscript significantly. This work was supported by grants from the National Basic Research Program of China (2013CB733203) and China Postdoctoral Science Foundation (2015M580680). It is also supported by the Natural Science Foundation of Hubei province (2015CFA019).