INFLUENCE OF GRID SIZE AND TERRAIN RESOLUTION ON WIND-FIELD PREDICTIONS FROM AN OPERATIONAL MESOSCALE MODEL

One of the activities of the National Oceanic and Atmospheric Administ ration's Air Resources Laboratory is to predict the consequences of at mospheric releases of radioactivity and other potentially harmful mate rials. This paper describes the application of the Regional Atmospheri c Modeling System (RAMS) to support air quality forecasting. The utili ty of using RAMS for real time prediction of local-scab flows and for detailed postevent analysis is examined for a Nuclear Regulatory Commi ssion exercise at the Susquehanna nuclear power plant in Pennsylvania. During the exercise (10 December 1992) a strong East Coast low pressu re system created complex interactions between the regional-seal and l ocal topographical features of the Susquehanna River valley. Results f rom a series of sensitivity experiments indicated significant topograp hical forcing and vertical decoupling although the synoptic forcing wa s quite strong in this relatively wide and shallow valley. The best ag reement between the RAMS predictions and observations was obtained wit h horizontal and vertical resolutions of 2.5 km and at 12 m above grou nd level for the first vertical wind level, respectively. Therefore, i t would have been very difficult to configure RAMS to predict the loca l circulations in real time, given the very high resolution requiremen ts. The vertical resolution needed to properly resolve terrain forcing s in the Susquehanna Valley was similar to vertical resolution used by other researchers over steeper and narrower valleys. However, the hor izontal resolution requirements were not as critical: about 10 times c oarser than in more complex terrain. The degree of topographical smoot hing was also found to have a significant effect upon the predictions. Experiments performed by assimilating all available surface-level win ds in the model domain with various degrees of nudging slightly improv ed the simulation of the low-level winds. Subsequent analyses indicate d that pressure-driven channeling and downward momentum mixing were th e primary physical mechanisms for this case.