Impact of gas‐phase mechanisms on Weather Research Forecasting Model with Chemistry (WRF/Chem) predictions: Mechanism implementation and comparative evaluation
Journal Article
Overview
abstract
Gas‐phase mechanisms provide important oxidant and gaseous precursors for secondary aerosol formation. Different gas‐phase mechanisms may lead to different predictions of gases, aerosols, and aerosol direct and indirect effects. In this study, WRF/Chem‐MADRID simulations are conducted over the continental United States for July 2001, with three different gas‐phase mechanisms, a default one (i.e., CBM‐Z) and two newly implemented ones (i.e., CB05 and SAPRC‐99). Simulation results are evaluated against available surface observations, satellite data, and reanalysis data. The model with these three gas‐phase mechanisms gives similar predictions of most meteorological variables in terms of spatial distribution and statistics, but large differences exist in shortwave radiation and temperature and relative humidity at 2 m at individual sites under cloudy conditions, indicating the importance of aerosol semi‐direct and indirect effects on these variables. Large biases exist in the simulated wind speed at 10 m, cloud water path, cloud optical thickness, and precipitation, due to uncertainties in current cloud microphysics and surface layer parameterizations. Simulations with all three gas‐phase mechanisms well reproduce surface concentrations of O3, CO, NO2, and PM2.5, and column NO2. Larger biases exist in the surface concentrations of nitrate and organic matter (OM) and in the spatial distribution of column CO, tropospheric ozone residual, and aerosol optical depth, due to uncertainties in primary OM emissions, limitations in model representations of chemical transport, and radiative processes. Different gas‐phase mechanisms lead to different predictions of mass concentrations of O3 (up to 5 ppb), PM2.5 (up to 0.5 μg m−3), secondary inorganic PM2.5 species (up to 1.1 μg m−3), organic PM (up to 1.8 μg m−3), and number concentration of PM2.5 (up to 2 × 104 cm−3). Differences in aerosol mass and number concentrations further lead to sizeable differences in simulated cloud condensation nuclei (CCN) and cloud droplet number concentration (CDNC) due to the feedback mechanisms among H2SO4 vapor, PM2.5 number, CCN, and CDNC through gas‐phase chemistry, new particle formation via homogeneous nucleation, aerosol growth, and aerosol activation by cloud droplets. This study illustrates the important impact of gas‐phase mechanisms on chemical and aerosol predictions, their subsequent effects on meteorological predictions, and a need for an accurate representation of such feedbacks through various atmospheric processes in the model. The online‐coupled models that simulate feedbacks between meteorological variables and chemical species may provide more accurate representations of the real atmosphere for regulatory applications and can be applied to simulate chemistry‐climate feedbacks over a longer period of time.
