Multi-Objective Gravity Data Inversion for Basement Depth Estimation in Sedimentary Basins

<title>Abstract</title> <p>Accurate estimation of basement depth is significant in the energy resource exploration and mineral prospecting. Basement depth is inherently controlled by a combination of sedimentation, tectonic subsidence, and geological processes that collectively shape the subsurface architecture. The reconstruction of basement depth and its density variations represents a valuable and well-established application of gravity anomaly inversion. Conventional approaches to generating geologically realistic basement models typically rely on regularization techniques, frequency-domain filtering methods, and smoothness constraints within optimization frameworks to stabilize the inversion process. This study investigates the application of the adaptive multi-objective optimization algorithm (AMALGAM) for accurate characterization of sedimentary basin geometry. Within this optimization framework, gravity data inversion is formulated to simultaneously minimize two competing objective functions, namely, the data misfit term and the model smoothness constraint, thereby eliminating the need for post-inversion smoothing or spectral filtering of model parameters. The multi-objective optimization ensures that one objective does not over-dominate without applying weighting and regularization schemes. The result is an ensemble of non-dominated optimal solutions called the Pareto-optimal set. The robustness and reliability of the proposed algorithm are systematically evaluated and validated using both noise-free and noise-added synthetic gravity datasets. The algorithm’s effectiveness is further demonstrated through its application to gravity anomaly data acquired over two real sedimentary basins. The results confirm that the proposed approach constitutes a promising, accurate, and computationally efficient method for subsurface imaging in geological and geophysical exploration.</p>