The Aerosol Indirect Effect in Climate Models: Where is it? Where should it be? Where should it not be?

Many climate models participating in the CMIP5 assessment for IPCC AR5 include aerosol indirect effects for the first time. Satellite observations indicate that climate models that include aerosol indirect effects generally estimate the variation of cloud-drop number with aerosol optical depth reasonably well but overestimate the variation of liquid water with aerosol optical depth. Subsequently, aerosol indirect effects could be overestimated. This paper will discuss the methods by which aerosol indirect effects have been included in the GFDL climate model CM3. Although parameterizations for droplet activation for aerosols of specified size distribution are quite robust, important aspects of the dynamics of low-cloud systems which may respond to changing aerosol concentrations are not represented, potentially limiting or even countering the effect on cloud albedo of aerosol-induced changes in cloud drop number. Effects of aerosols on ice processes in deep convection, which could impact longwave cloud forcing associated with upper-tropospheric anvil clouds, are also not included. The presentation will identify key areas where process-based improvements in parameterizations of clouds and convection are required in climate models to better understand the role of the aerosol indirect effect in the climate system.