Development of a MCNP Library in Support of a Broad-Area Search Detection System

In this paper, we describe our efforts to develop a library of radiation transport modeling data to support the broad-area search for radioactive materials. Our purpose is two- fold: first, the challenges plaguing radiation detection and localization decisions at extended source-detector distances stem largely from inaccurate simplifications applied to the measurement model. Second, the computational time required to model a given wide-area search-space calls for the development of a series of estimates prior to a crisis or trigger event occurring. By leveraging the ability of the Monte Carlo Neutral Transport (MCNP) code to accurately estimate parameters of interest based on the phenomena associated with radiation transport, we demonstrate the ability to overcome the limitations induced by applying gross approximations to the radiation transport model. For example, the 1/r2 approximation is increasingly inaccurate at ranges beyond 50 m. In support of our work to develop the Broad-Area Search Bayesian Processor, we show how the development of a library of environmental models executed using MCNP significantly impacts the range at which detection and localization decisions are made. Future work will focus on the development of MCNP models with sufficient fidelity to replicate additional environmental factors affecting detector response.