A Cirrus cloud scheme for general circulation models

Abstract

A statistical cloud scheme for non-convective cirrus formed by homogeneous freezing of supercooled aerosols is presented. As large-scale cirrus clouds exhibit metastable thermodynamic states and have long lifetimes, a fully prognostic approach is developed. The scheme is based on separate probability distribution functions of total water representing clear-sky and in-cloud conditions. These distributions are based on in situ observations and have few degrees of freedom so that they can be used in forecast and climate models. Predictive equations are derived for horizontal cloud fraction and mixing ratios of grid-mean water vapour and ice water, in-cloud water vapour and ice crystal number. The scheme allows sub- and supersaturations with respect to ice to occur in cloud-free air and inside cirrus, and produces nucleated ice crystal concentrations and sizes in good agreement with observations, an important prerequisite for an improved treatment of cirrus cloud radiative properties. Implementation in general circulation models, open issues and future research avenues are discussed.