The limitation of curvature gravity gradient tensor for edge detection and a method for overcoming it

• We analyzed the limitation of curvature gravity gradient tensor for edge detection.
• We presented a new improved method to overcome the limitation.
• The new method was tested on the synthetic model and real gravity data.
• It demonstrated that the new method is effective for edge detection.

Eigenvalues of the curvature gravity gradient tensor (CGGT) have been utilized to detect and outline edges of geologic bodies. In this paper, we analyzed and discussed the application scope and the limitation of CGGT using a complex synthetic model. We found that the large eigenvalue only can be used to outline edges of positive density bodies, and the small eigenvalue only can be used to delineate edges of negative density bodies. However, in the actual geological situation, the positive and negative density bodies usually coexist simultaneously. Therefore, it has certain limitations to detect edges using this method directly. In view of the limitations of the method, we combined the gravity anomaly with large eigenvalues, and proposed a new improved method for the eigenvalues of CGGT. The improved method can be utilized to outline edges of causative sources in any case. It was tested on synthetic model data and real field data. All of the results have shown that the new improved method is effective for edge detection.