Diagnostic analysis of ozone concentrations simulated by two regional-scale air quality models

Since the Community Multiscale Air Quality modeling system (CMAQ) and the Weather Research and Forecasting with Chemistry model (WRF/Chem) use different approaches to simulate the interaction of meteorology and chemistry, this study compares the CMAQ and WRF/Chem air quality simulation results for a month-long retrospective study period (August 2006) over the eastern United States, including comparisons with data from several observation networks. To help improve the comparability of the two models, the 2005 Carbon Bond chemical mechanism (CB05) was implemented into WRF/Chem. In addition, the same emissions, initial and boundary conditions have been used in both models to inter-compare simulated ozone (O3) from the WRF-driven CMAQ and WRF/Chem models. Results reveal that ground-level O3 from both models is biased high, especially in the central South and Ohio River Valley; however, WRF/Chem predicts roughly 10% more O3 aloft (1000–2500 m AGL) than CMAQ. Different model configurations due to the choice of land surface model (LSM), planetary boundary layer (PBL) physics scheme, and convective cloud parameterization contributed to the differences seen in simulated O3, but most important were the different treatments of the radiative effects of clouds by their respective photolysis schemes.

► We compared ozone predictions from two air quality models for a summer month.
► Each model used same chemical mechanism, emissions, initial and boundary conditions.
► Both models overestimate surface ozone, mainly in the South and Ohio River Valley.
► WRF/Chem generally predicts more ozone than WRF-driven CMAQ.
► Boundary layer and photolysis schemes are big factors in modeled ozone differences.