Many atmospheric transport and dispersion models now exist to provide consequence assessment during emergency response to near-field releases. One way of estimating the uncertainty for a given forecast is to statistically analyze an ensemble of results from several models. This paper discusses the involvement of the Savannah River Technology Center (SRTC) in a European Union program, ENSEMBLE. This program utilizes an internet-based system to ingest transport results from numerous modeling agencies. SRTC uses a prognostic atmospheric numerical model to provide three-dimensional and time-varying meteorology as input to a stochastic Lagrangian particle model. This modeling system provides better resolution than is typically available using standard large-scale weather products and Gaussian plume models.

The model design is discussed, as well as results from other modeling agencies during a series of exercises conducted over a two-year period. The exercises assumed hypothetical releases from various locations throughout Europe, with each agency delivering results as quickly as possible. The design used by SRTC is shown to successfully provide model results for each of the events in a relatively short time period, and to be versatile enough to handle assumed sources of varying geometry, strength and duration. Plots showing agreement among models for integrated concentration above a certain threshold value for the exercises indicate that the SRTC modeling system qualitatively captures similar plume characteristics provided by the other modeling agencies.