Following the completion of this project, Pacific Northwest National Laboratory (PNNL) now offers a public-domain software tool which helps munitions contractors to plan a site survey and then validate Munitions of Environmental Concern (MEC) site remediation. The Visual Sample Plan Software (VSP) contains modules to calculate site footprint reduction and remediation verification of sites that may have or did have MEC. VSP software may be downloaded from http://vsp.pnnl.gov/.

Post-remediation verification (PRV) methods of VSP during an actual remediation verification process on the South Tract (~157 acres) of the Navy/Denver Research Institute (DRI) Site were performed. The PRV demonstrated that the UXO Estimator software tool and the VSP-RI module for estimating TOI/acre and confidently showing that it is less than some value using an uninformative Bayesian prior both require exactly the same amount of survey acreage to be performed. However, the VSP-RI module using reasonable and defensible informed priors can result in significantly reduced required survey acreage while maintaining the desired confidence or significantly increased confidence using the same survey acreage. 

Post-remediation verification (PRV) methods of VSP during an actual remediation verification process on the South Tract (~157 acres) of the Navy/Denver Research Institute (DRI) Site were performed. The PRV demonstrated that the UXO Estimator software tool and the VSP-RI module for estimating TOI/acre and confidently showing that it is less than some value using an uninformative Bayesian prior both require exactly the same amount of survey acreage to be performed. However, the VSP-RI module using reasonable and defensible informed priors can result in significantly reduced required survey acreage while maintaining the desired confidence or significantly increased confidence using the same survey acreage. 

At the San Luis Obispo, CA demonstration site, VSP’s discrimination technologies and classification analysis routines generated a metric which reflects the degree of belief that the anomalies are TOI. TOI’s are defined as MEC anomalies or anomalies which exhibit MEC properties. These metrics are then used to determine the anomaly digging order. The metrics are unable to exactly classify each anomaly as TOI/Not-TOI due to measurement error (including machine noise and environmental nuisance factors). PNNL has developed a viable statistically-defensible algorithm, Bayesian Dig Stop (BDS), that can be used to support early dig stopping recommendations. It allows one to state with X% confidence that there is no more than a Y% chance that MEC remains on the site. The methods can be used after an initial set of digs or sequentially as digging proceeds. It also provides a universal metric to compare the performance of various anomaly classification algorithms.