Application of co-occurrence and trend analysis methods for identification of long-term groundwater quality monitoring strategies at a USDOE facility

V.M. Vyas, M. Ouyang, S. Tanwar, P.G. Georgopoulos (EOHSI, UMDNJ - R.W. Johnson Medical School and Rutgers University); W. Strawderman (Department of Statistics, Rutgers University); D.S. Kosson (Department of Civil and Environmental Engineering, Vanderbilt University)

Current regulations designed to safeguard human health and environmental quality do not explicitly specify requirement for long-term monitoring strategies. However, monitoring strategies based on a holistic analysis of existing data could lead to early warning of future non-compliance. Such analyses would also identify existing gaps and redundancies in the data to highlight deficiencies in the existing monitoring programs, and suggest additional monitoring leading to improved spatial and temporal resolution of contaminant concentrations. Identification of redundancies will suggest changes in existing environmental monitoring practices, and lead to more cost efficient monitoring programs that do not compromise the information content of the data. This study conducts a statistical evaluation of existing groundwater monitoring data for the US Department of Energy (USDOE) facility at Savannah River Site (SRS) in South Carolina, on the South Carolina-Georgia border. Sixteen Constituents of Concern (COC) were identified prior to the commencement of the study, on the basis of their health risk and also their detected on-site levels: aluminum, arsenic, cesium-137, cobalt, iron, manganese, mercury, nitrate as nitrogen, nitrite as nitrogen, pH, selenium, thallium, tin, tritium, uranium-238, and vanadium. Cesium-137, and tritium were measured as radionuclides, while the metallic constituents were measured as dissolved metals. Data for the study were obtained from the Geochemical Information Management System (GIMS) maintained by SRS, for the period from October 1992 to September 1998. The analysis was restricted to the wells screened in the water table aquifer. Measurements for each COC were subjected to geostatistical spatial declustering tests to identify redundant wells in a region, and to variogram analysis to identify spatial and temporal correlation within measurements of a particular COC, as well as correlations between different pairs of COC. Additional tests such as Principal Component Analysis (PCA) were carried out to identify the parameters that were the most critical contributors to variability in measurements. The results of PCA were used to identify the most critical inputs for a monitoring strategy. The results from this study are designed to be used as inputs to identify optimal environmental monitoring strategies that maximize the information content of collected data.

ACKNOWLEDGEMENTS: Our research has been partially 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.