3D compact geometry inversion for gravity data

Abstract Gravity inversion quantitatively provides a 3D model of density contrasts, significantly enhancing the information extracted from acquired data. However, the inherent non-uniqueness of inversion poses challenges in precisely determining the boundaries of anomalous bodies. We have developed an iterative algorithm of gravity inversion that reconstructs the geometric features of the anomalous bodies by discretizing the 3D interpretation model with vertical and juxtaposed prism cells. These prisms incorporate sheet-like initial models which are typically derived from prior information or imaging results. This study proposed a new parameter, the Thickness Factor (TF), which is determined by the thickness of the prism cells under the assumption of homogeneous anomalous bodies. The TF establishes an approximate linear relationship between the source geometry and gravity anomalies, enabling the reconstruction of the source geometry to be formulated as a linear optimization problem. The approach demonstrates the potential for target inversion in the presence of multiple causative sources in synthetic cases and shows insensitivity to noise signals and reliability in reconstructing the geometry of complex sources. The proposed method is then applied to real data from the Galinge iron ore deposit in Northwest China and the drilling data is used as prior information. The inversion results are consistent with previous drilling interpretations and allow a rough estimation of the volume of the ore bodies.