Protocol Optimization for Determining Spatial Extent of Areas in Need of Remediation in Compliance with 10 CFR 40 Appendix A Criterion 6(6) - 23384

Areas affected by uranium mill tailings often have spatially heterogeneous concentrations of both gamma and non-gamma-emitting radionuclides of concern (ROCs) which present unique challenges for remedial decision-making. In these areas, Ra-226 is most often the limiting ROC for compliance due to its relatively high radiotoxicity compared to other ROCs. However, in the US, 10 CFR 40, Appendix A Criterion 6(6) requires that all ROCs, not just Ra-226, meet a sum of fractions (SOF) criterion when averaged over any 100 square meters (m²). The SOF criterion requires that the sum of the ratio of each ROC, after background subtraction, to its surface soil concentration limit (SSCL) does not exceed one (unity). Areas that exceed the Criterion 6(6) standard require remediation. Historically in the US, soil decommissioning work at uranium mills has favored in situ gamma measurements (a proxy for Ra-226) as the sole basis for remedial decisions. These measurements were typically collected through dynamic gamma surveys across a potentially impacted area subdivided into 10 meters by 10 meters (10 x 10 m) grids. However, the historical approach may result in unacceptably high type I and type II error rates during final status survey sampling, most often due to spatial heterogeneity of ROCs in uranium mill tailing-affected areas. The purpose of this study is to improve the reliability of remedial decision-making by comparing the type I and type II error rates of five statistical methods (one parametric and four non-parametric) used to identify areas that require remediation. The parametric method consisted of selecting grids where the average gamma counts per second (cps) in a 10x10 m grid is above an assigned gamma guideline level (GGL) which represents a threshold above which an area must be remediated. The non-parametric methods consisted of selecting grids in which some fixed number (n) of individual gamma measurements per grid exceed the assigned GGL. This study tested four n values using a non-parametric method (n = 30, 25, 20, 15). In addition, each parametric and non-parametric method was tested using two potential GGLs: 62 and 70 cps. These GGLs represent potential action levels solely for the purpose of this study. Following grid identification, this study calculated the number of areas remaining larger than 100 m² in exceedance of the study GGL. These are areas which would not be chosen for remediation using the method, though remediation would be necessary based on the GGL and area size (false negative or type I error). Within the confines of this study, the results indicate that using a non-parametric method correctly identifies a higher fraction of areas that require remediation than a parametric method. Additionally, an n value of 15 correctly identifies the greatest fraction of areas that require remediation according to the assigned GGL and leaves the lowest fraction of areas remaining larger than 100 m². This method can be used to inform soil sampling decisions regarding environmental remediation of areas affected by uranium mill tailings. (authors)