Sensitivity of simulated chemical concentrations and aerosol-meteorology interactions to aerosol treatments and biogenic organic emissions in WRF/Chem

Coupled air quality and climate models can predict aerosol concentrations and properties, as well as aerosol direct and indirect effects that depend on aerosol chemistry and microphysics treatments. In this study, Weather Research and Forecasting with Chemistry (WRF/Chem) simulations are conducted over continental U.S. (CONUS) for January and July 2001 with the same gas-phase mechanism (CB05) but three aerosol modules (Modal Aerosol Dynamics Model for Europe/Secondary Organic Aerosol Model (MADE/SORGAM), Model for Simulating Aerosol Interactions and Chemistry (MOSAIC), and Model of Aerosol Dynamics, Reaction, Ionization and Dissolution (MADRID)) to examine the impacts of aerosol treatments on predictions of aerosols and their effects on cloud properties and radiation. The simulations with the three aerosol modules give similar domain mean predictions of surface PM2.5 concentrations but exhibit a strong spatial variation in magnitudes with large differences in eastern U.S. Large discrepancies are found in the predicted concentrations of sulfate and organicmatter due to different treatments in secondary inorganic and secondary organic aerosol (SOA) formation. In particular, the nucleation calculation in MADE/SORGAM causes mass buildup of sulfate which results in much higher sulfate concentrations that those predicted by WRF/Chem with the other two aerosol modules. Different PM mass concentrations and size representations lead to differences in the predicted aerosol number concentrations. The above differences in PM concentrations lead to large differences in simulated condensation nuclei (CCN) and cloud properties in both months. The simulated ranges of domain mean are (1.9-14.3) x 10(9) m(-3) and (1.4-5.4) x 10(9) m(-3) for PM2.5 number concentration, (1.6-3.9) x 10(8) cm(-2) and (1.9-3.9) x 10(8) cm(-2) for CCN, 102.9-208.2 cm(-3) and 143.7-202.2 cm(-3) for column cloud droplet number concentration (CDNC), and 4.5-6.4 and 3.6-6.7 for cloud optical depths (COT) in January and July, respectively. The sensitivity simulation for July 2001 using online biogenic emissions increases isoprene concentrations but decreases terpene concentrations, leading to a domain mean increase in O-3 (1.5 ppb) and a decrease in biogenic SOA (-0.07 mu g m(-3)) and PM2.5 (-0.2 mu g m(-3)). Anthropogenic emissions contribute to O-3, biogenic SOA (BSOA), and PM2.5 concentrations by 38.0%, 44.2%, and 53.6% domain mean and by up to 78.5%, 89.7%, and 96.3%, respectively, indicating that a large fraction of BSOA is controllable through controlling atmospheric oxidant levels in CONUS. Anthropogenic emissions also contribute to a decrease in downward shortwave flux at ground surface (-5.8 W m(-2)), temperature at 2 m (-0.05 degrees C), wind speed at 10m (-0.02 m s(-1)), planetary boundary layer height (-6.6m), and precipitation (-0.08mm), as well as an increase in CCN (+5.7 x 10(-7) cm(-2)), in-cloud CDNC (+40.4 cm(-3)), and COT (+0.6). This work indicates the need for an accurate representation of several aerosol processes such as SOA formation and aerosol-cloud interactions in simulating aerosol direct and indirect effects in the online-coupled models.