Assessment of Two Aerosol Modules of CAM5

The Community Atmosphere Model (CAM) is widely employed in research of climate simulation and climate change. The latest version 5.0, provides two modules to simulate atmosphere aerosol, named MAM3 and MOZART, respectively. Several main atmosphere aerosols are simulated by these two modules, and the simulated surface concentrations of these aerosols are examined by Interagency Monitoring of Protected Visual Environments Program (IMPROVE) and European Monitoring and Evaluation Program (EMEP). The simulated global distributions of aerosol column concentration, as well as aerosol global budgets are compared with median model results on AeroCom website.Both MAM3 and MOZART modules can capture the seasonal distribution of sulfate aerosol; the simulated surface concentrations are in reasonable agreement with observations, although the values in summer are usually high. The correlation coefficients between models and observations for two modules are both around 0.89. Also, both MAM3 and MOZART can capture spatial and temporal distribution of black carbon aerosol. However, these two modules both underestimate surface concentration of black carbon by a factor of 2—3. The correlation coefficients between models and observations for two modules are both around 0.62, which are believed to be caused by smaller emission fluxes and higher rates of wet removal. The two modules have large difference in simulating organic matter, both having a bias by a factor of 2—3. MAM3 overestimates surface concentrations of organic matter with a normalized mean bias of 92.1%, while MOZART makes an underestimation of 58.1%. It's found that both of these biases usually happen in summer and autumn. A separate analysis demonstrates that the primary organic matter simulated by these two modules are very close, while MAM3 and MOZART have serious differences on simulation of the secondary organic carbon (SOC), which primarily contributes to the bias of total organic matter. Sea salt global budgets by MAM3 and MOZART are close, but the total content of sea salt is larger than median model results on AeroCom. The most likely cause is that lower rates of dry removal and wet removal in the CAM5. With similar mechanism but different emission factor, the two modules perform differently in simulating mineral dust; flux of mineral dust emission in MAM3 is nearly three times as large as that in median model results on AeroCom, and thus overestimates the total content, while MOZART underestimates mineral dust burden, because its emission flux is 40% smaller than that in median model results on AeroCom.According to the comparison in global distribution and global budget, it indicates that CAM5 has a weaker intensity of aerosol translation and diffusion, thus, the removal mechanism should be improved.