An Air Quality Chemical Postprocessor for Lagrangian Dispersion Models

Many urban areas across the U.S. suffer from high levels of urban/regional ozone, PM10, and PM2.5. The control of ozone and PM2.5 is difficult because a natural background of these pollutants exist. Furthermore, many urban areas that suffer from elevated pollutant levels are located in coastal areas where significant uncertainty exists in the prediction of pollutant dispersion associated with land/sea interfaces. To reduce this uncertainty a need exists to develop more effective predictive tools for development of effective strategies to reduce air pollution in coastal urban areas.

Traditional approaches to assess the impact of emissions in complex environments have included the application of statistical, Gaussian dispersion, single chemical box, Lagrangian, and Eulerian models. Each of these models has advantages and disadvantages. As part of a current SERDP supported study, we are developing a meteorological air quality model with in-line chemistry that combines the advantages of both Eulerian and Lagrangian models.

We are modeling several dates in the summer of 2003 with high ozone levels that have been identified by airborne measurements. The measurements where made over the port of San Diego and down wind (eastward) from 1200 to 1700 LST to understand the contribution of the urban area and the port of San Diego on downwind air quality. Our modeling domain includes the area of the flights. MM5 output is used to drive the DRI LPM (Lagrangian Particle Model) that produces hourly output of particle positions with a pointer to each specific particle. We also track the chemical species from each source for each particle and calculate the concentration in an Eulerian grid by the RACM box model, so we can fallow the hourly chemical concentration change of each particle. We can thus analyze in a more detailed way the contribution and danger from small sources to small areas. Preliminary results show agreement in the temporal and spatial distributions of modeled and measured ozone areas, although concentrations were not corrected matched.