Quantitative Assessment of CO2 Leakage Using Time-Lapse Gravity Inversion

We propose an innovative approach for the interpretation of time-lapse gravity data aimed at estimating subsurface mass variations associated with CO₂ injection and storage. The method is based on the Extremely Compact Source (ECS) inversion technique (Maiolino et al., 2024), which is used to isolate the individual contributions of discrete CO₂ mass accumulations in the subsurface and to quantify their associated excess mass.

To date, ECS inversion has been primarily applied as a filtering strategy to remove regional-scale contributions from potential field data or to separate the effects of closely spaced sources. The approach relies on an iterative inversion of potential field observations to derive a subsurface model composed of source distributions with minimal volumetric extent, referred to as atoms, simultaneously ensuring a low data misfit. A key advantage of the method is that it does not require any a priori information about subsurface properties.

Once the ECS model is obtained, the excess mass associated with each source contributing to the observed gravity anomaly can be readily computed. In this study, we demonstrate that the proposed approach enables precise identification of CO₂ accumulations within the reservoir and allows for accurate estimation of net mass variations related to both stored CO₂ and leakage occurring along permeable fault zones.