Application of GIS and geostatistical analysis tools for characterizing contamination at the USDOE Savannah River Site in South Carolina
V.M. Vyas, Ming Ouyang, Panos Georgopoulos (EOHSI, UMDNJ - R.W. Johnson
Medical School and Rutgers University);
William Strawderman (Rutgers University)
David Kosson, (Vanderbilt University)
Identification of background levels and trends of Constituents of Concern (COC) is an essential prerequisite for characterizing contamination at a site. Background characterization helps in identifying the true extent of contamination from a particular source, and in setting reasonable remediation endpoints in cases where background levels may be higher than state and federally mandated Maximum Contaminant Levels. Background levels are defined here as concentrations arising from natural geological characteristics or diffuse off-site contaminant sources.
In this case study, we identify background groundwater concentrations of certain COC at the USDOE's Savannah River Site facility. Background groundwater concentrations of the COC at a contaminated site are generally established by monitoring the groundwater in wells sufficiently upgradient of the contamination. In this study, however, the distinction between impacted and unimpacted areas had to consider several complicated features of the nature of contamination such as intermingling of sources; lack of recognized unimpacted areas; varying sampling, analysis and reporting procedures; and a large proportion of non-detects for several critical COC. Furthermore, the background analysis methodology had to sift through a large volume of data to identify patterns that indicate contamination - approximately 90,000 observations from 869 water table monitoring wells were used.
GIs, statistical, and geostatistical analysis software were used to develop, apply, and evaluate alternative methodologies for defining background levels. GIs methods were used to map the extent and distribution of contamination, and for identifying potentially impacted areas. Alternative sets of mathematical filters were used to identify potentially contaminated areas and screen out potentially contaminated wells. The filters consisted of screening steps based on geostatistical estimation, trend analysis, regression, cluster analysis, and regression-tree methods. The final outcomes are lists of background wells and probability distributions of background levels for each COC; and maps of locations for background monitoring wells. Salient features of different methodologies and filtering schemes, and selected results highlighting the sensitivity of the methodologies to alternative filters are presented.
ACKNOWLEDGMENTS: Our research has been supported by a grant to the Institute for Responsible Management, Consortium for Risk Evaluation with Stakeholder Participation, from the US Department of Energy, Instrument DE-FG26-00NT 40938. The viewpoints expressed in this report are solely the responsibility of the authors and do not necessarily reflect the views of the US Department of Energy or its contractors.