THE USE OF MODEL-DERIVED AND OBSERVED PRECIPITATION IN LONG-TERM SULFUR CONCENTRATION AND DEPOSITION MODELING

As a continuation of a previous study for the winter of 1987, the Hybr id Single-Particle Lagrangian Integrated Trajectory Model (HY-SPLIT) w ith Atmospheric Chemistry Including Deposition (ACID) is run for each season of 1989. Two simulations for each season are run; one using for ecast precipitation data from the Nested Grid Model (NGM), and the oth er using gridded measured precipitation data from observing stations o f the National Weather Service (NWS). Modeled seasonal and annual sulf ur dioxide (SO2) and sulfate (SO42-) concentrations in air and SO42- w et deposition over the eastern United States are compared with observe d values. The model does well predicting the seasonal and annual spati al pattern of SO2 and SO42- concentrations in air, and of SO42- wet de position when using NWS precipitation. In all but the spring season, t he model tends to overpredict SO2 concentration in air, but is well wi thin a factor of 2; mean relative errors (MREs) are less than 20%. SO4 2- concentration in air is slightly overpredicted during the fall and winter and underpredicted during the spring and summer; MREs are less than 12%. SO42- wet deposition is underpredicted for all seasons; MREs are less than 30%. The model-predicted and measured values of the pre vious study are well within the scatter of the data in this study. Usi ng NGM precipitation does not significantly affect the results for the 1989 study period. Although the NGM does overpredict the precipitatio n during the summer months along the Gulf states, thereby increasing s lightly the SO42- wet deposition and further underpredicting the model ed SO42- air concentration, the differences are small enough to sugges t that the use of model-forecast precipitation can be an adequate subs titute for precipitation observations over a seasonal or longer period .