Characterization of mesoscale weather prediction errors for dispersion modeling

DTRA uses high-resolution mesoscale forecasts to drive HPAC, the DTRA dispersion modeling tool that generates critical forecasts of dosage resulting from releases of chemical, biological, or radiological agents. HPAC is designed to provide probablistic information based on estimates of the uncertainty of the input forecast fields of the meteorological variables (wind, temperature, etc.). We will develop and implement more accurate, reliable, and efficient means of specifying the forecast uncertainty. For speed and efficiency, which is especially needed in real-time applications of HPAC, we will use a combination of statistical methods and the feature calibration and alignment (FCA) technique to provide 5d (x,y,z,t,variable) estimates of the uncertainty of the input weather forecasts. In the future this information may be provided by using ensembles of forecasts. Our implementation of FCA uses a variational algorithm to partition errors into phase errors, bias (or amplification) errors, and residual small scale errors. We will derive error statistics for mesoscale forecasts in terms of the components of FCA, and use these statistics to define the variability parameters used as input to HPAC.