Final Report for Statistical Methods and Tools for UXO Site Characterization on Final Simulated Site

Abstract : This report details the demonstration of statistical tools developed by Pacific Northwest National Laboratory (PNNL) and Sandia National Laboratory to identify preferred geophysical transect designs for sites suspected of containing UXO and to analyze the resulting geophysical anomaly data. The analysis tools developed provide the Visual Sample Plan (VSP) user the ability to identify potential target areas, delineate boundaries of the target areas wherein UXO is most likely, and estimate the anomaly density within each identified and delineated target area. The methods are also useful for providing justification for no further investigation in very large portions of investigation sites. The PNNL/Sandia characterization design team received information on a simulated munitions site from Mitretek Systems where 4.2 mortars, 155 mm munitions, and differing size precision bombs were used. Mitretek provided varying degrees of information for each impact area in the conceptual site model and the characterization team used the available information to develop geophysical transect survey designs to identify potential targets areas located in each impact area. The resulting anomaly data from the transect designs were analyzed and potential target areas were identified and delineated. Once the potential target areas were identified, delineated and anomaly densities estimated, the truth data was provided by Mitretek and an evaluation of the performance of the statistical methods was conducted. This report summarizes the performance of the statistical tools in identifying and delineating target areas and estimating anomaly density. The simulated site consisted of 5 major areas (ARA-1, ARA-2A, ARA-2B, SRA-1, and BGR-1) ranging from 6200 acres to 74650 acres in size. Three of the areas were used as artillery ranges, one site was used as a small arms practice range, and the one was a precision bombing range.