Gravimetry, HVSR and seismic geostatistical–geophysical inversion for improved density modelling in mining industry

Accurate density modelling is critical in mining industry, as density directly controls ore tonnage estimates and economic viability. Conventional approaches commonly rely on average values or simple linear regressions, leading to substantial uncertainty in resource estimation. This study proposes the integration of geophysical data through three geophysical–geostatistical inversion methodologies: gravimetry, horizontal-to-vertical spectral ratio (HVSR), and seismic reflection. The approaches were evaluated using a synthetic iron ore deposit characterised by sparse density sampling, three-times more grade data, and exhaustive geophysical information. The inversion results were compared with a pre-defined best-performing purely geostatistical methodology. Gravimetric inversion showed the fastest convergence and effectively reproduced major density anomalies. Although requiring more iterations, HVSR and seismic inversions also generated density models closely matching the reference, with seismic inversion achieving the best performance in reproducing the full density range with minimal bias. All three inversion methodologies significantly outperformed the purely geostatistical approach, reducing uncertainty and better honouring the spatial variability of the deposit. Model validation using variogram reproduction, accuracy plots, and multidimensional scaling confirmed the robustness of the proposed workflows. Practical considerations related to acquisition time and cost indicate that gravimetry offers the most cost-effective solution, whereas seismic inversion provides higher-resolution results at increased cost.