MARSSIM: Statistical decision methodology

As part of the process of preparing a site for unrestricted release, it must be demonstrated that residual radioactivity in facilities and environmental media has been reduced to acceptable levels. Typically, compliance with radiological criteria for unrestricted release is demonstrated by conducting final status surveys of the site or facility and reporting the survey results to the cognizant regulatory agency for evaluation. At present, the NRC guidance for such surveys is contained in the raft report NUREG/CR-5849, dated June 1992. DOE has similar guidance. However, this survey methodology has limitations that become apparent as residual radioactivity approaches background levels. Among these are the prescription of fixed sampling densities and the assumption that the data are normally (or lognormally) distributed. An alternative statistical approach is being considered for conducting radiological surveys at or near background. Nonparametric statistical techniques are used that do not require the data to have a specific distribution, can be applied to data containing non-detects, and specify sampling densities that depend on the specific data quality objectives of the decision makers. This methodology generally results in final Status survey plans that are flexible and accurate, while being more efficient than previous methods. The method has been successfully applied to example sites using proposed dose-based unrestricted release criteria