Effects of Aerosol Representation Complexity On Air Quality and Weather Forecasting Skill Over South America

Abstract

WRF-Chem simulations of regional smoke observed during the 2012 South American Biomass Burning Analysis (SAMBBA) campaign are conducted to determine whether more complex aerosol schemes improve air quality and meteorological forecasting skill over South America. The SAMBBA flights spanned the dry to wet transition season and thus provide useful case studies for examining aerosol radiation and aerosol cloud feedbacks. The WRF-Chem representations of aerosol range in complexity from climatology (meteorology only) to aerosol aware Thompson microphysics (primary smoke aerosol emissions but no gas or aqueous chemistry) to RACM-MADE-VBS-AQCHEM, state of the art gas/aqueous/aerosol chemistry that provides cloud condensation nuclei distributions to the cloud microphysics. Biomass burning emissions and plume rise based on MODIS Fire Radiative Power (FRP) data are used. The simulations are evaluated against ground based and SAMBBA aircraft gas and aerosol, cloud, and meteorology observations, as well as MODIS and AERONET Aerosol Optical Depth (AOD), TRMM precipitation, and radiosoundings. Simulations with aerosol aware microphysics and full chemistry underestimate cloud screened AOD relative to MODIS and AERONET. WRF-Chem simulations provide good agreement with SAMBBA aircraft CO, O3, and NOx mean profiles, except that simulated O3 is underestimated in clean conditions and WRF-Chem NOx is overestimated in smoky conditions. Particulate organic carbon (POC) mean profiles measured by aircraft are well represented by the WRF-Chem simulations, but black carbon (BC) is underestimated. Compared with the TRMM satellite product, simulated hourly mean precipitation covers larger areas and has higher maximum values. Using full chemistry increases the area of resolved scale precipitation, while using Thompson aerosol aware microphysics decreases precipitation area. FIM-Chem global model simulations of the SAMBBA period are evaluated to determine whether more complex chemistry provides benefits for global numerical weather prediction. Finally, simulations using an aerosol aware cumulus parameterization are compared with high resolution simulations and used to quantify regional and seasonal scale aerosol effects.