Evaluation of the RAMS model for estimating turbulent fluxes over the Chesapeake Bay

This study has examined the ability of a mesoscale model to compute pollutant deposition velocity and surface fluxes over water with the accuracy needed for air pollution studies. The Regional Atmospheric Modeling System (RAMS) was evaluated against buoy data over the Chesapeake Bay. Turbulence and surface layer variables (e.g., surface fluxes, friction velocity) and ultimately deposition velocities were primarily examined against the buoy data for 23 cases in 1994. Sensible heat fluxes and temperatures compared well with observations when the model horizontal resolution was reduced to 5 km or less and the first model level were 12 m. Other turbulent variables compared poorly especially under stable stratifications. Overall, latent heat fluxes and deposition velocities were significantly under predicted even with these finer resolutions. Results degraded when the effects of clouds on short- and long-wave radiation were incorporated. Results were also insensitive to increase in the initial sea surface temperature (SST) over the observed daytime range (1°C), and to the use of an improved roughness length parameterization over water. Errors were much larger during stably stratified conditions, therefore, substantial under prediction of pollutants to the bay would be expected by RAMS during stable periods. However, RAMS would be a reliable predictor of fluxes and deposition velocity during unstable conditions. Further experiments were done to help explain the poor predictions during the daytime for stable conditions. Fluxes over water were strongly sensitive to the soil moisture content of the surrounding land masses for stable atmospheric conditions and improved when geostrophic forcing terms were added to the momentum equations. However, the improved results given with geostrophic forcing were either representative of synoptic forcing or a decoupled layer from the surface that could not be resolved.