UREG/CR-6672: PROBABILITY DISTRIBUTIONS DEVELOPED FOR IMPORTANT PARAMETERS EMPLOYED IN THE RADTRAN 5/LHS INTERFACE

The construction of cumulative distributions of values for RADTRAN variables that take on a wide range of values in the real world is described. Where adequate data existed, probability distribution functions were constructed for important RADTRAN input parameters. Values obtained by structured Monte Carlo sampling (Latin Hypercube Sampling) from these distributions were used to construct RADTRAN input files. Among the distributions of important variables that could be developed were: route characteristics (population densities, length, and rural, suburban and urban fractions), truck and rail accident rates, truck stop time, post-accident evacuation time, atmospheric stability category, dose rate at one meter, and highway traffic density and vehicle occupancy As an example, methods and derivations for rail-mode data, and the RADTRAN calculations performed with these input values, will be presented. The number of LHS samples needed to obtain a representative, stable, random sample for use in risk calculations is also discussed.