Improving the accuracy of 1D surface nuclear magnetic resonance surveys using the multi-central-loop configuration

Abstract

Surface nuclear magnetic resonance is a near-surface geophysical method for characterizing the spatial distribution of liquid water in the top 100 m of the subsurface. The recovered water content models are obtained through the solution of an ill-posed inverse problem that is a function of acquisition parameters, including location and shape of the transmitter and receiver coils. In this paper, we introduce the multi-central-loop acquisition and inversion strategy where one or several smaller receivers coils are placed in the center of the larger transmitter loop and where all the data sets synchronously recorded through each loop are inverted simultaneously. We investigate the attributes of this acquisition and inversion strategy including the ability to provide improved resolution, accuracy and reduced uncertainty on the estimated subsurface models compared to single channel acquisition methods. Using widely-adopted inversion methods and introducing a new data interpretation technique called Bayesian Evidential Learning 1D imaging, we show that the multi-central-loop configuration provides improved recovery of synthetic models and reduced levels of inverted parameter uncertainty. A field case is also presented where the multi-central-loop results appear to better match the lithologic knowledge of the area compared with single channel configurations, again providing smaller uncertainties.