Environmental monitoring of leaks using time-lapsed long electrode electrical resistivity

Abstract

Abstract Highly industrialized areas pose challenges for surface electrical resistivity characterization due to metallic infrastructure. The infrastructure is typically more conductive than the desired targets and will mask the deeper subsurface information. The risk of this occurring may be minimized if steel-cased wells are used as long electrodes in the area near the target. We demonstrate a method of using long electrodes to electrically monitor a simulated leak from an underground storage tank with both synthetic examples and a field demonstration. Although the method of using long electrodes has been proposed by others, no time-lapse resistivity data have been collected, modeled, and analyzed within a nuclear waste tank farm environment. Therefore, the main objective of this work was to test whether the long electrode method using steel-cased wells can be employed to spatially and temporally track simulated leaks in a highly industrialized setting. A secondary objective was to apply a time-lapse regularization procedure in the inverse modeling code, similar to the 4D tomography approach by Kim et al. (2009), and to test the procedure's effect on the quality of the outcome regarding plume intensity and position. For the synthetic examples, a simple target of varying electrical properties was placed beneath different types of layers of low resistivity to simulate the effects of the infrastructure. Both surface and long electrodes were tested on the synthetic domain, and the test cases covered a variety of survey parameters including low and high electrode density, noise, array type, and the explicit location of the wells relative to the target. All data were processed in four dimensions, where the regularization procedure was applied in both the time and space domains. The synthetic test case showed that the long electrode resistivity method could detect relative changes in resistivity that was commensurate with the differing target properties. The surface electrodes, on the other hand, had a more difficult time matching the original target's footprint unless the electrodes were distributed at a greater density on the surface. The simulated tank leak in the field experiment was conducted by injecting a high conductivity fluid in a perforated well within the S tank farm at the Hanford Site, and the resistivity measurements were made before and after the leak test. The field results showed a lowered resistivity feature develops south of the injection site after cessation of the injections. The parameter used in the time-lapsed inversion had a strong influence on the differences in inverted resistivity between the pre- and post-injection datasets, but the interpretation of the target was consistent across all values of the parameter. The long electrode electrical resistivity monitoring (ERM) method may provide a tool for near real-time monitoring of leaking underground storage tanks given a sufficient density of wells.