Borehole Nuclear Magnetic Resonance Estimation of Specific Yield in a Fractured Granite Aquifer

Abstract

In this study, we introduce a novel field-based method to estimate specific yield (S_y) in fractured, low-porosity granite aquifers using borehole nuclear magnetic resonance (bNMR). This method requires collecting a bNMR survey immediately following a pump test, which dewaters the near-borehole fractures. The residual water content measured from bNMR is interpreted as “bound” and represents the specific retention (S_r) while the water drained by the pump is the S_y. The transverse relaxation cutoff time (T_2C) is the length of time that partitions the total porosity measured by bNMR into S_r and S_y. When applying a calibrated T_2C, S_y equals the bNMR total porosity minus S_r; thus, a calibrated T_2C is required to determine S_y directly from NMR results. Based on laboratory experiments on sandstone cores, the default T_2C is 33 ms; however, its applicability to fractured granite aquifers is uncertain. The optimal T_2C based on our pumping test is 110 ± 25 ms. Applying this calibrated T_2C on a saturated, A-type granite at our field site, we estimate the S_y to be 0.012 ± 0.005 m³ m⁻³ which is significantly different from the S_y (0.021 ± 0.005 m³ m⁻³) estimate using the default T_2C of 33 ms. This S_y estimate falls within a range determined using traditional hydraulic testing at the same site. Using the conventional T_2C (33 ms) for fractured granite leads to an inaccurate S_y; therefore, it is essential to calibrate the bNMR T_2C for the local site conditions prior to estimating S_y.