Evaluation of Cloudiness in the Atmospheric Model Intercomparison Project

Abstract

Estimates of cloudiness in up to twenty-five models participating in the Atmospheric Model Intercomparison Project are compared with the ISCCP C2, Nimbus 7, Meteor and Warren et al. (WH) observations. The root mean square differences between the annual means of the model calculations and the C2 observations, after global means are removed, are all greater than 14%, whereas the difference between the C2 and Meteor observations is less than 8%. Calculations of smoothed seasonal cycles show that many models simulate a seasonal cycle which leads that of the observations by about two months. Other models appear to simulate the proper phase of the seasonal cycle, but have spatial patterns which are in large disagreement with the observations. The global means of model high cloudiness are about 2-6 times greater than that of either of the satellite observations. The large differences with observations and between models is probably primarily related to the lack of a standard definition for what constitutes thin high cloud. The mean low cloud as observed from below in most models is comparable to WH. There is usually good qualitative agreement in the annual mean meridional structure of high and low cloud.