A Downscaling Modeling Framework for Evaluating the CAM5 Physics Suite within WRF-Chem

The Community Atmosphere Model Version 5.1 (CAM5), the atmospheric component of the Community Earth System Model (CESM), is typically run using a grid-spacing of 1 or 2 degrees. Future climate models will be run at much higher resolutions (e.g. grid spacings of 10-20 km), but the behavior of the CAM5 physical parameterizations at these scales is not well known and current computing capabilities generally do not allow global models to be run routinely at mesoscale resolutions. Hence, we developed a framework to explore CAM5 process representations with WRF-Chem (Weather Research and Forecasting model with chemistry) using identical emissions, consistent initial and boundary conditions for meteorology, trace gases, and aerosol particles, and the same parameterizations as used in CAM5, including the treatment of cloud microphysics, fractional clouds, aerosols, shallow and deep convection, and turbulent mixing. This framework allows us to test CAM5 physics over a range of scales within WRF-Chem, and apply our Aerosol Modeling Testbed tools to directly compare WRF-Chem simulations with field campaign measurements that contain large temporal and spatial variability in cloud and aerosol properties. In this study, we explore the resolution dependency of CAM5 physics by comparing WRF-Chem simulations (with CAM5 physics) run at various horizontal grid-spacings (10, 20, 40, 80, and 160 km). The results are evaluated against the Indirect and Semi-Direct Aerosol Campaign (ISDAC) field campaign data and compared with simulations from WRF-Chem with typical mesoscale parameterizations. We find that CAM5 physics produces higher ice and liquid cloud condensate amount and higher aerosol concentration with increasing resolution. However, the CAM5 physics suite shows less scale dependence than the typical mesoscale parameterizations because it considers sub-grid variability of clouds (i.e., allowing fractional clouds). We also find that mesoscale eddies are better resolved in high-resolution simulations, producing filaments of concentrated aerosol plumes that are responsible for higher aerosol loading episodes observed over Barrow, Alaska associated with "Arctic haze". The spatial variation of aerosol and clouds from the high-resolution simulations provide a means of quantifying sub-grid scale variability in the coarse-resolution global models, facilitating the reformulation of current parameterizations toward improved scale-aware parameterizations.