An Improved System to Support Airborne Dispersion Modeling and Emergency Response at the Savannah River Site

Abstract:

The Atmospheric Technologies Group (ATG) at the Savannah River Site (SRS) is charged with supporting the site’s Emergency Response Organization (ERO) in the aftermath of an airborne contaminant release. Following such an event, site emergency managers will require immediate estimates and projections of downwind concentrations. To meet this need, the ATG currently uses a Gaussian diffusion model (‘Puff-Plume’), which uses as input observed wind and turbulence data collected on site, complemented with forecast data from a mesoscale model - the Regional Atmospheric Modeling System (RAMS). This coupled model can provide a map of downwind airborne concentrations and is therefore useful for informing the decision-making process.

Two improvements have been developed to exploit new modeling algorithms and the large quantity of boundary-layer turbulence data collected on site.

First, a genetic algorithm has been implemented to create an ensemble of RAMS forecasts to provide input to the Puff-Plume model. This entails varying the model parameterizations among the different members, scoring the resulting forecasts according to the observed meteorology, and selecting the best-performing members to create a new generation. The parameters primarily dictate the way the model simulates turbulence and interactions with the surface, two of the most difficult processes to simulate in a mesoscale model. This optimizes the RAMS model and allows us to characterize uncertainty in the transport forecasts needed to provide emergency managers with guidance on the range of possible outcomes.

Second, the Puff-Plume model has been revised to implement Taylor’s theory of dispersion. Currently, the model parameterizes dispersion as a function of atmospheric stability and the high-frequency crosswise deviations in wind speed. Taylor’s theory incorporates the autocorrelation of deviations in wind speed, and is used to more explicitly determine the increase in plume width with downwind distance.

These changes should improve both parts of the coupled modeling system and allow ATG to provide more useful information to the ERO during an airborne release.