Important global and regional differences in aerosol cloud-albedo effect estimates between simulations with and without prognostic aerosol microphysics

Abstract

Aerosol-cloud interactions are among the most uncertain climate forcings, in part due to the strong sensitivity of cloud droplet number concentration (CDNC) to changes in the size distribution of potential cloud condensation nuclei. Despite this sensitivity of simulated aerosol-cloud interactions to variations in size-resolved aerosol concentrations being well established, many chemistry-climate and chemical-transport models do not include explicit treatment of the aerosol size distribution. We use a global chemical-transport model to estimate the aerosol cloud-albedo effect (CAE) due to anthropogenic emissions with prognostic sectional aerosol microphysics and compare this to the CAE calculated when the simulated aerosol mass of each species is remapped onto a prescribed size distribution. We find that although both the prescribed and prognostic methods compare similarly well with present-day size-distribution observations, there are substantial differences in the relative CDNC and CAE due to anthropogenic emissions. When using the prognostic size-distribution method, anthropogenic emissions yield a 25-75% larger increase in CDNC over most land masses but a 50-75% smaller increase in some remote-marine regions than in the prescribed size-distribution methods. Simulations using the prognostic scheme yield a global mean anthropogenic CAE of -0.87 W m-2, while the simulations with the prescribed scheme predict -0.66 W m-2. In South America and South Asia, differences in the CAE exceed 3.0 W m-2. These differences suggest that simulations with prescribed size-distribution mapping are unable to capture regional and temporal variability in size-resolved aerosol number and thus may lead to biases in estimates of the CAE.