Modeling Secondary Organic Aerosol and Its Climatic Impacts using WRF/Chem over Continental U.S.: Sensitivity to Gas-Phase Mechanisms and Secondary Organic Aerosol Treatments

Modeling secondary organic aerosols (SOA) is challenging due to a variety of organic compounds present in the atmosphere with varying chemical characteristics, volatilities, and reactivities. In addition, the mechanism of SOA formation in the atmosphere is not as well understood as the mechanisms involved in inorganic aerosol formation. The volatility basis set (VBS) approach provides an improved SOA treatments than those used in a majority of 3-D air quality models due to its ability to represent a volatility distribution of organic vapors in the atmosphere as well as the successive oxidation of organic compounds. This study explores the impact of using different gas-phase mechanisms associated with the VBS-SOA treatment in the Weather Research and Forecasting model with Chemistry (WRF/Chem) v3.6.1 coupled with the 2005 Carbon Bond gas-phase mechanism (CB05), the Version 6 of the Carbon Bond mechanism (CB6) and the Statewide Air Pollution Research Centre 2007 (SAPRC07). The objectives are to identify the best combination of the gas-phase mechanism and SOA treatments that produce results in the closest agreement with SOA and organic compounds (OC) observations over the continental U.S., and to analyze simulated climatic impacts using different mechanism combinations on climatic impacts. Simulations with a 36-km horizontal resolution are conducted over a 15-day period from May 15 to 31, 2010 during which SOA observations are available over Pasadena and Bakersfield in California from the CALNEX field campaign. A set of simulations comparing the default Modal for Aerosol Dynamics in Europe/Secondary Organic Aerosol Model (MADE/SORGAM) with the VBS are conducted. The VBS treatment significantly improves the SOA predictions over Pasadena and Bakersfield as compared to those with the MADE/SORGAM treatment although OC concentrations are largely underpredicted for both sets of simulations. Another set of simulations are conducted comparing CB05-VBS, SAPRC07-VBS and CB6-VBS. All simulations show slight underpredictions against IMPROVE and STN for OC over continental U.S., with the best using CB05-VBS (NMB of -2% against IMPROVE) and the worst using CB6-VBS (NMB of -10% against IMPROVE). The results from this study will illustrate the importance of accurately simulating SOA and OC concentrations for climate predictions.