Reconciling the difference of aerosol effects between global climate models and satellite observations

Abstract

Aerosol effects are considered a major source of uncertainty in global climate models and direct and indirect radiative forcings differ from model to model. These forcings are routinely evaluated (and calibrated) against satellite observations for their near-global coverage. However, the forcings calculated from model output are not directly comparable with those computed from satellite retrievals since sampling and algorithmic differences between models and observations are not typically accounted for. The conventional model validation procedures comparing satellite retrievals directly to model simulations leads to apple/orange inter comparisons that can mislead model development and introduce biases. Hence, we have developed an aerosol lidar simulator for global climate models that uses the same algorithms as those used to produce the CALIPSO data products. The simulator allows us to sample and derive aerosol fields in the model and real atmosphere in identical ways. Using the ACME model simulations, we have found that the simulator-retrieved aerosol distribution and aerosol-cloud interactions are significantly different from those computed from conventional approaches, and that the model is much closer to satellite estimates than previously recognized.