Evaluation of Cloud Description in General Circulation Models Using A-Train Observations

Abstract

The representation of clouds and cloud feedbacks remain very different from one model to another, and they still constitute a major limitation to the reliability of climate change projections. It is therefore imperative to improve the representation of cloud processes in models. Until recently, the evaluation of several fundamental aspects of the cloudiness as the three-dimensional distribution of the cloud cover has been largely indirect, based on passive remote sensing satellites (e.g. ERBE, Scarab, ISCCP) which measure the TOA radiative fluxes. The A-train observations constitute exceptional tools to characterize the cloud properties. We will show that due to errors compensations, the climate models produce correct top-of the atmosphere fluxes. The A-train observations (CALIPSO, PARASOL, CERES) are used to unravel the errors compensations and to evaluate quantitatively the clouds description in various climate models using the COSP (CFMIP Observation Simulator Package) within the CFMIP (Cloud Feedback Model Intercomparison Program). A process oriented evaluation is conducted by analysing statistically the co-located A-train observations at high spatial resolution to built pictures of the cloud properties at the scale of the cloud process. This multi-instrument dataset at high spatial resolution is then used to assess the cloud parametrization in a climate model.