A case study of a frontal system simulated by a climate model: Clouds and radiation

Abstract

A case study is carried out to evaluate the capability of the Community Atmosphere Model (CAM3) in simulating frontal clouds over the Southern Great Plains (SGP). It focuses on a midlatitude cyclonic storm system observed during the March 2000 Intensive Observation Period (IOP) of the Atmospheric Radiation Measurement Program (ARM). Cloud biases are found to be consistent with climate model biases in this region: overestimation of optically thick clouds and shortwave cooling at the top-of-the-atmosphere (TOA), underestimation of optically thin middle clouds. Satellite observations, NOAA ETA analysis and a suite of ARM measurements are used to analyze the model cloud biases. We found two independent causes of model errors during two stages of the evolution of the cyclone. In the first stage, the biases are from cloud microphysical properties. The model significantly overpredicted cloud liquid water path, while it underpredicted cloud ice water path. As a result, it overestimated the magnitude of shortwave cloud forcing. In the second stage, the model cloud biases are primarily caused by a faster eastward propagation of the 500 mbar ridge behind the cyclone, which dissipated the high- and middle-level clouds but favored low clouds, leading to biases in cloud forcing. Averaged over the cyclone domain and period, the model simulated shortwave and longwave TOA cloud forcing of -113 W/ m 2 and 30 W/m 2, respectively, while the corresponding observations are -69 W/m 2 and 38 W/m 2. Our results illustrated the feasibility of using synoptic cases to understand and eventually eliminate systematic cloud biases in climate models. Copyright 2007 by the American Geophysical Union.