Apparent Density Mapping Algorithm in the Spectral Domain: A Case Study Under Tibet
摘要
We present an algorithm for regional apparent density mapping in the spectral domain. The method formulates the relationship between gravity disturbance and subsurface density distribution through a convolution kernel and applies Tikhonov regularization to stabilize the inversion. Synthetic experiments with variable density layers demonstrate that the algorithm achieves high accuracy, with density differences below 0.2 g/cm3 and gravity prediction errors less than 3 mGal. Robustness tests with Gaussian noise confirm reliable performance even under 20% noise contamination. The algorithm is further validated using satellite gravity observations from the GOCO06s model over the Tibetan Plateau. After removing topographic and Moho contributions, the refined gravity disturbance is inverted to reveal low-density anomalies beneath the plateau and high-density margins in adjacent basins. The results are consistent with the ongoing India–Asia continental collision and associated lithospheric deformation. Gravity forward modeling confirms the accuracy of the recovered density, with residuals approximating a normal distribution and standard deviation below 0.01 mGal.