Coupled IMPACT aerosol and NCAR CAM3 model: Evaluation of predicted aerosol number and size distribution

[1] Simulated aerosol fields from a coupled aerosol/atmospheric circulation model that includes prediction of both sulfate aerosol size and number are evaluated. Sensitivity tests are used to evaluate uncertainties due to the inclusion of primary emitted particulate sulfate as a means of representing nucleation of particles in subgrid-scale plumes, the use of two boundary layer aerosol nucleation mechanisms, and a three-mode sulfate aerosol representation. Simulated annual and global aerosol budgets are comparable to other model studies with the exception of carbonaceous aerosols and fine mode dust, where smaller mass concentrations are simulated. The model underestimates the accumulation mode aerosol number in the marine boundary layer over middle and low latitudes, which is consistent with an underestimate of fine mode sea salt mass in these locations. Primary emitted particulate sulfate contributes significantly to aerosol number at sites located in the boundary layer over Europe, but the absence of constraints on the number of such particles from either observations or fine-resolution models makes this treatment undesirable. Boundary layer nucleation mechanisms improve the comparison of simulated aerosol number concentrations with observations in the marine boundary layer, suggesting that a treatment of boundary layer nucleation is needed in global aerosol models, although more studies are needed to quantify how different nucleation mechanisms and condensable gases other than sulfuric acid affect aerosol number. The three-mode representation of sulfate aerosol simulates the observed increase in accumulation mode number concentration with altitude in the upper troposphere and improves the simulated Aitken mode aerosol number concentration there. This indicates the importance of a separate representation of freshly nucleated particles when nucleation is an important source of particle number concentrations.