Progress In Interpretation Of Transient Electromagnetic Data In Terms Of Subsurface Permeability

Abstract

The relative importance of nonpoint-source (NPS) pollution in the degradation of water quality has increased in the last two decades due to the control of pollutant releases from point sources (Sharpley and Meyer 1994). The most important source of NPS pollutants are agriculture and urban areas, which impact water quality in rivers, lakes, estuaries and groundwaters through the release of eroded sediments, fertilizers, pesticides, and municipal sewage sludge. Because of this, NPS pollution is an important environmental concern at state and national levels. Several transport processes control the dispersal of NPS pollutants, including leaching to groundwater, surface runoff (Pereira and Rostad 1990), and aerial transport and deposition (Glotfelty et al. 1984). Once in groundwater, these contaminants can impact surface water during stream recharge. While the losses of NPS pollutants from agricultural fields or urban areas can be small as a percentage of the total amount released, the cumulative additions to river systems from large drainage areas can be significant. The development of a reliable contaminant transport model for tracing the dispersal of NPS pollutants through a heterogeneous aquifer requires knowledge of the spatial distribution of hydrologic parameters such as hydraulic conductivity or permeability (Zheng and Bennett 1995). In the commonly used stochastic approach (Dagan 1989), these distributions are treated as spatial random functions whose variance and correlation length scale are determined from hydrological information collected by head measurements, pump and tracer tests, and soil core analyses. Point estimates of the hydraulic conductivity or permeability are determined by kriging, a geostatistical interpolation procedure which estimates unknown random functions based on spatial correlations between point observations. Noninvasive geophysical techniques are becoming an increasingly popular component of hydrogeological studies since many geophysical data are sensitive to spatial variations in hydraulic conductivity and permeability. In addition, the cost of surface exploration is only a fraction of the cost of drilling and a wide area1 coverage is readily obtained. Controlled source electromagnetic (CSEM) methods provide maps of electrical conductivity and are an appropriate choice if the depth scale of investigation is on the order of 10 m-l km. We are investigating the possibility of placing electrical constraints on the subsurface permeability as part of a larger, integrated study to determine the fate of agricultural chemicals introduced at a research site near the Brazos River. The research involves a joint analysis of the electrical conductivity structure, the available soil cores, and other hydrologic data. A Bayesian approach is taken, following Copty et al. (1993), in which geophysical data are used to update the variance and correlation length scale of a hydrologicallyderived, random permeability field. In this paper we will describe the underlying theory and demonstrate an example using synthetic CSEM data that have been generated from a simulated geoelectrical section of our study site. The incorporation of CSEM data into a determination of subsurface permeability is made more difficult by the presence of man-made electrical conductors at the site. These include an aluminum pipeline and an overhead power line. Such artifacts are characteristic of the human impact at many environmental sites. We are presently applying the integral equation code of Qian and Boerner (1995) to account for the effect of the aluminum pipeline on the CSEM data.