Akademik Veri Yönetim Sistemi
MATHEMATICAL GEOSCIENCES, cilt.57, ss.523-545, 2025 (SCI-Expanded, Scopus)
Three-dimensional inversion of gravity data consists of an ill-posed and underdetermined system of equations. To overcome this problem, inversions are generally regularized using various stabilizing functionals, and depth weighting is also often implemented to adjust depths of the recovered structures due to the lack of depth resolution of vertical component gravity data. By exploiting the ease of implementation of nonlinear filters such as smoothing operators, we demonstrated rank order smoothing as a nonlinear constraint to regularize the three-dimensional gravity inversion problem and applied a new depth weighting scheme based on the Levy probability density function. Synthetic experiments showed that the rank order smoothing successfully provides boundary information by eliminating low-frequency fluctuations in the recovered models while minimizing the observed-calculated data misfit. The proposed depth weighting approach was found to be able to emphasize depths in which anomalous structures are considered to be more probable. Thereafter, the proposed inversion scheme was demonstrated on field gravity data collected in the Sinanpa & scedil;a Graben in western Turkey to reveal faults and the general basin structure. The results suggest that rank order smoothing can produce models that agree with the geology and are easier to interpret.