A One-Dimensional Study of Possible Cirrus Cloud Feedbacks

Abstract

Feedbacks involving cirrus clouds may Play an important role in modulating the response of the climate system to external forcing. A standard radiative-convective model with high vertical resolution (10-mb grid spacing in the upper troposphere) is used to investigate some of these feedbacks in response to changes in atmospheric carbon dioxide concentration. It is shown that such a high resolution (or, indeed, an even higher resolution) may be necessary to resolve the change in tropopause height following a doubling of carbon dioxide. The cirrus cloud is allowed to adjust its height and ice water content in response to the warming, and experiments are repeated 1) for fixed and moist adiabatic lapse rates, 2) with and without a prescribed relative humidity feedback, 3) at low as well as high vertical resolution, and 4) with the inclusion of a simple correction to account for the uncertain number of small ice crystals in cirrus clouds. In this model the ice water content feedback is found to be positive because changes in cloud temperature have more effect on the thermal infrared emissivity than on the visible albedo. The feedback is substantially reduced upon the inclusion of small ice crystals. The effect of allowing the cloud to relocate as the tropopause height increases is to strongly reduce the ice water feedback by an amount dependent on the assumed lapse rate. Without the tropopause response, and using the moist lapse rate, the ice water feedback without the small crystal correction enhances surface warming by 18%. With the tropopause response, the same feedback is almost zero. For a fixed lapse rate the size of the ice water feedback upon including the tropopause response is also reduced, but not so markedly. In the case of low vertical resolution, closer to those used in GCMS, the ice water feedback remains strong and positive, while the cloud height is either less able to respond, or alternatively, overestimates changes in height. The results imply that models with a coarse vertical revolution may be unable to capture small changes in cloud position that could substantially affect the overall strength of cirrus cloud feedbacks.