Comparison of Two Photochemical Modeling Systems in a Tropical Urban Area

Abstract

Two photochemical smog modeling systems, UAM-V/SAIMM (the Variable-Grid UAM/Systems Applications International Mesoscale Model) and CHIMERE/ECMWF (European Center for Medium Range Weather Forecast), are applied to the same tropical domain (Bangkok Metropolitan Region) and the same episode (January 13–14, 1997) to evaluate their relative performance using the same anthropogenic emission database (emission database available at the Pollution Control Department [PCD] 1997). Ozone (O₃) produced by both models meets U.S. Environment Protection Agency (EPA) suggested prediction criteria of mean normalized bias error and mean normalized gross error on January 14 but none on January 13. Both models are tested with various modified databases of precursors emissions from the PCD original database. Performance of UAM-V is the best when using the modified emission data with volatile organic compound (VOC), NO_x, and CO mobile source emission reduced by 50%, 50%, and 20% from the original database. CHIMERE suggests a similar emission database except for the VOC emission, which is a reduction by 40% from the original PCD mobile source emission. Spatial and temporal variations of O₃, CO, NO_y (total reactive nitrogen), and Ox (NO₂+O₃) predicted by both model systems using the modified emissions are compared with observations. The agreement between the two models is better than their agreement with observations. The difference between the results reflects the gaps in our current knowledge of photochemistry. With the respective modified emission database, CHIMERE produces comparable results with UAM-V for all considered parameters, and O₃ simulated by both models does meet all of the EPA suggested criteria. Both models predict similar results of O₃ productions with various combinations of anthropogenic and biogenic emissions in Bangkok and surrounding provinces. The simulated minumum O₃ and maximum NO_y in the city by both modeling systems with the modified emission database on January 14 are comparable to observed data; therefore, it is considered that the modeling tool can be used for studying the sensitivity of ozone formation to precursor emission reduction for Bangkok. Both model responses to emission reductions show similar trends, which indicate that the O₃ formation in Bangkok is more sensitive to VOC than NO_x emissions.