Inversion of surface nuclear magnetic resonance data by an adapted Monte Carlo method applied to water resource characterization

Abstract

Inversion of surface nuclear magnetic resonance (SNMR) provides important information about aquifers, such as their depths, thickness, pore size and water content. Different methods (inverse quadratic, linear programming) have been applied to the problem of SNMR data inversion, but there has not yet been any attempt to explore model space. We propose that an adaptation of the Monte Carlo method presented here is suitable for exploring the set of solutions consistent with SNMR data. We also demonstrate the capability of this method applied to the interpretation of SNMR data with examples of both synthetic and field data inversions. We also show that the method can be used to obtain various results, such as a posteriori water distribution with depth and a posteriori pore-size distribution with depth.