Transport and dispersion in coastal regions were studied in the Model Validation Program (MVP) conducted by U.S. Air Force, National Oceanic and Atmospheric Administration, Kamada Science and Design, and Aerospace Corporation at Cape Canaveral Air Force Station (CCAFS), FL, and Vandenberg Air Force Base, CA. During three different two-week long sessions from 1995-97, 336 short puffs of sulfur hexaflouride (SF6), and 143 ground-level and 56 elevated plumes, each of 2-3 hrs duration, were released. Plume source strength was ~20 kg/hr. Fast-response gas analyzers mounted on six mobile vans and one or two aircraft were used for plume sampling, while multiple infrared cameras were deployed to monitor the puffs. Atmospheric stabilities ranged from strongly convective to mildly stable conditions. Extensive data on meteorology, surface energy fluxes, and tracer concentrations were collected and stored in an excellent, but thus far underused MVP data base.

The MVP data are unique in the range of release and sampling heights of up to 1.2 times the atmospheric boundary layer (ABL) depth H, as shown by a plot of scaled plume release height (h/H) versus atmospheric stability (H/L), where L is Monin-Obukhov length. Thus, the potential exists for comparing MVP results with the laboratory data of Willis and Deardorff (Atmos.Environ.12, 1978, 1305-1311). Several rawinsonde-based ABL depth algorithms are compared with reported estimates by analysts. Scatterplot comparisons and figures of merit are shown. A bulk Richardson (RiB) number-based algorithm gives the best results at CCAFS. Downwind distances to maximum ground-level concentration (MGC) are found to be less than linearly proportional to release height, as shown by plots of distance to MGC as a function of scaled release height (h/H). Plume behaviors within a seabreeze, cumulus-convergence zone are also presented. Elevated sampling shows consistent over-water subsidence and wider plumes than are seen for surface sampling. These characteristics may be typical for plume dispersion in coastal regions of the eastern U.S., involving cumulus convergence and inland waterways.

In work sponsored by the U.S. Defense Threat Reduction Agency (DTRA), the Hazard Prediction Advisory Capability (HPAC) model, normally driven by stationary diagnostic winds, was driven by high-resolution fields from the Weather Research and Forecasting (WRF) model. The latter utilized North American Regional Reanalysis (NARR) data from the National Center for Environmental Prediction (NCEP) archive at National Center for Atmospheric Research, as well as local data from the MVP. The WRF/HPAC model simulations are evaluated using the MVP data base. The first study case, MVP Test 209, a tracer release from 20 m height at CCAFS, occurred on 6 November 1995, from 21:22 - 23:56 GMT. WRF was initialized shortly after local sunrise, with 3-D fields, surface data, and fluxes from the NARR archive. The mobile plume sampling data are averaged to stationary grid points and compared directly to WRF/HPAC simulation results.

The project's ultimate objective is to test and evaluate the potential of the WRF/HPAC system for enabling quicker and more accurate real-time dispersion forecasts of toxic agents in coastal and littoral regions, and to show how local wind data and/or in situ plume measurements can be used to refine and update sophisticated dispersion modeling results within an interactive, real-time response system.