Joint inversion of Surface Nuclear Magnetic Resonance and Vertical Electrical Sounding

Abstract

The method of Surface Nuclear Magnetic Resonance (SNMR) provides a very new technology to directly determine subsurface water distribution. The microscopic magnetization of water molecules is used to derive water content and pore size information from SNMR soundings. The observed similarity and agreement between interpreted aquifer structure from SNMR and resistivity distribution from Vertical Electrical Sounding (VES) has led to our objective to jointly invert both data sets using a generalized petrophysical model based on Archie's Law. To perform inversion of both methods, the Simulated Annealing (SA) technique was applied. Since a very fast numerical solution is available for both geophysical methods, this kind of guided random search algorithm promises better performance than least square methods. The developed inversion algorithm has been applied on a number of different synthetic data to study its properties and prove its reliability. Investigations on well-known test sites where both methods were conducted finally proved the effectiveness of the joint inversion on real data. The interpretation of the subsurface model could be optimized beyond an enhanced spatial resolution to a quantitative interpretation of the ratio of mobile and adhesive water contents, leading to prediction of hydrological parameters from geophysical investigations.