Electrical resistance tomography during in-situ trichloroethylene remediation at the Savannah River Site

Abstract

Electrical resistance tomography was used to monitor in-situ remediation processes for removal of volatile organic compounds from subsurface water and soil at the Savannah River Site near Aiken, South Carolina. This work was designed to test the feasibility of injecting a weak mixture of methane in air as a metabolic carbon source for natural microbial populations which are capable of trichloroethylene degradation. Electrical resistance tomograms were constructed of the subsurface during the test to provide detailed images of the process. These images were made using an iterative reconstruction algorithm based on a finite element forward model and Newton-type least-squares minimization. Changes in the subsurface resistivity distribution were imaged by a pixel-by-pixel subtraction of images taken before and during the process. This differential tomography removed all static features of formation resistivity but clearly delineated dynamic features induced by remediation processes. The air-methane mixture was injected into the saturated zone and the intrained air migration paths were tomographically imaged by the increased resistivity of the path as air displaced formation water. We found the flow paths to be confined to a complex three-dimensional network of channels, some of which extended as far as 30 m from the injection well. These channels were not entirely stable over a period of months since new channels appeared to form with time. Also, the resistivity of the air injection paths increased with time. In another series of tests, resistivity images of water infiltration from the surface support similar conclusions about the preferential permeability paths in the vadose zone. In this case, the water infiltration front is confined to narrow channels which have a three-dimensional structure. Here, similar to air injection in the saturated zone, the water flow is controlled by local variations in formation permeability. However, temporal changes in these channels are minor, indicating that the permeable paths do not seem to be modified by continued infiltration.