Impact of a new scheme for optical properties of ice crystals on climates of two GCMs

A new parameterization scheme for the optical properties of nonspherical ice crystals is implemented in two global climate models. The scheme is based on polynomial fits to the results from a more elaborate scheme recently developed by the third author. The scheme assumes a temperature dependency of ice crystal size and is able to handle five different crystal shapes. We first carry out sensitivity experiments highlighting the direct ("noninteractive") forcing effect of the new scheme for different assumptions on ice crystal shape and size. It is found that the effects of crystal shape and size are of comparable importance in determining the radiative fluxes at the top of the atmosphere. We have further conducted multiyear simulations with full dynamic coupling. In these experiments we have chosen planar polycrystals as our basic crystal shape. Compared to control integrations with columns in one case and 30 mu m spheres in the other, we find a reduction in the cold biases of the models in the upper tropical troposphere. This is shown to be mainly due to enhanced longwave radiative warming there, caused by the presence of smaller ice crystals. The upper tropospheric warming affects tropical convection in such a way that moisture and ice water distributions are modified. Consequently, relative humidity biases are somewhat reduced in both models. In one of the models there is a significant improvement in the longwave radiative budget at the top of the atmosphere, while in the other, the radiative budget remains in fairly close agreement with observations.