Evaluating the impact of parameterization choice on WRF-chemistry simulations

The accurate prediction of air quality using numerical models involves correctly simulating both the meteorology and chemical processes. With the recent increases in computing power, complex three-dimensional air quality models have become a cost effective tool to study physical processes and their interaction with air chemistry, as well as different model formulations to treat the formation and evolution of aerosols. Some of these models are also used now to forecast air quality on an operational or semi-operational basis. One such tool is the Weather Research and Forecast (WRF)/Chemistry model (WRF-Chem). This model is unusual in that the transport and transformation of all chemical and aerosol components are calculated online, or in lock-step with the meteorological and thermodynamic calculations.

Among the physical processes that are most critical for air pollution modeling are the parameterization of the planetary boundary layer (PBL) and cloud processes. Several fundamentally different PBL parameterizations – also used for air quality applications – are available within the WRF framework. In addition, the simulated PBL growth and behavior is dependent upon the parameterization of the land surface, clouds, and shortwave radiation, and how they interact with the PBL parameterization.

The presentation will cover the results from the approach taken to determine the optimal model configuration through evaluating the influence that each parameterization or module has on the meteorological and air quality forecasts, using a systematic evaluation against data collected during the ICARTT/NEAQS-2004 field study (http://www.al.noaa.gov/2004). With so many possible combinations of physical and chemical modules, it was necessary to first determine which meteorological model configurations best reproduce the observed meteorological conditions for a select 2-week period from the NEAQS2004 field experiment.