Hydraulic Conductivity from Nuclear Magnetic Resonance Logs in Sediments with Elevated Magnetic Susceptibilities

This study examined the application of slim-hole nuclear magnetic resonance (NMR) tools to estimate hydraulic conductivity (K_NMR ) in an unconsolidated aquifer that contains a range of grain sizes (silt to gravel) and high and variable magnetic susceptibilities (MS) (10^-4 to 10^-2 SI). A K calibration dataset was acquired at 1-m intervals in three fully screened wells, and compared to K_NMR estimates using the Schlumberger-Doll research (SDR) equation with published empirical constants developed from previous studies in unconsolidated sediments. While K_NMR using published constants was within an order of magnitude of K, the agreement, overprediction, or underprediction of K_NMR varied with the MS distribution in each well. An examination of the effects of MS on NMR data and site-specific empirical constants indicated that the exponent on T_2ML (n-value in the SDR equation, representing the diffusion regime) was found to have the greatest influence on K_NMR estimation accuracy, while NMR porosity did not improve the prediction of K. K_NMR was further improved by integrating an MS log into the NMR analyses. A first approach detrended T_2ML for the effects of MS prior to calculating K_NMR , and a second approach introduced an MS term into the SDR equation. Both were found to produce similar refinements of K_NMR in intervals of elevated MS. This study found that low frequency NMR logging with short echo times shows promise for sites with moderate to elevated MS levels, and recommends a workflow that examines parameter relationships and integrates MS logs into the estimation of K_NMR .