Publications
Qualitative evaluation of the accuracy of maps for release of hazardous materials
The LinguisticBelief%C2%A9 software tool developed by Sandia National Laboratories was applied to provide a qualitative evaluation of the accuracy of various maps that provide information on releases of hazardous material, especially radionuclides. The methodology, %E2%80%9CUncertainty for Qualitative Assessments,%E2%80%9D includes uncertainty in the evaluation. The software tool uses the mathematics of fuzzy sets, approximate reasoning, and the belief/ plausibility measure of uncertainty. SNL worked cooperatively with the Remote Sensing Laboratory (RSL) and the National Atmospheric Release Advisory Center (NARAC) at Lawrence Livermore National Laboratory (LLNL) to develop models for three types of maps for use in this study. SNL and RSL developed the maps for %E2%80%9CAccuracy Plot for Area%E2%80%9D and %E2%80%9CAerial Monitoring System (AMS) Product Confidence%E2%80%9D. SNL and LLNL developed the %E2%80%9CLLNL Model%E2%80%9D. For each of the three maps, experts from RSL and LLNL created a model in the LinguisticBelief software. This report documents the three models and provides evaluations of maps associated with the models, using example data. Future applications will involve applying the models to actual graphs to provide a qualitative evaluation of the accuracy of the maps, including uncertainty, for use by decision makers. A %E2%80%9CQuality Thermometer%E2%80%9D technique was developed to rank-order the quality of a set of maps of a given type. A technique for pooling expert option from different experts was provided using the PoolEvidence%C2%A9 software.