PNNL

Abstract

Model estimates of aerosol impacts on climate frequently exceed observational estimates. Conventional evaluation procedures generally neglect sampling and retrieval differences that confound interpretation of discrepancies. Here we show that satellite simulators can reconcile differences between modeled and retrieved estimates of aerosol climate impacts. A-­train satellite simulators overflying the model atmosphere matching satellite retrieval procedures moderately underestimate the model’s total aerosol direct radiative effect and significantly underestimate anthropogenic aerosol indirect effects (by a factor of 3). Model cloud droplet size and precipitation probability retrieved from simulators indicate significantly less sensitivity to aerosols than the real model sensitivity, exhibiting much better agreement with satellite estimates. While climate models are still likely to have substantial deficiencies in representing aerosol­climate interactions, inconsistent comparisons between modeled and observed estimates can lead to inappropriate model calibration, development, and interpretation. Simulators are a necessary tool for comparing models and satellite estimates of aerosol climate effects.