Development of a cirrus parameterization scheme for use in GCMs: Comparisons with the observation in the tropics

Cirrus cloud cover and the amount of ice water content (IWC) in them are the two most important properties of cirrus. However, in GCMs their treatment is inadequate. For example in the standard NCAR CCM3 IWC is prescribed as a function of column-integrated water vapor and height. As a result cirrus IWC is rather uniform with values corresponding to what is normally regarded as thin cirrus. The in-situ observations in the tropics indicate that there is a significant amount of thick cirrus having the IWC an order of magnitude larger than what is prescribed in the model. The comparison with the ISCCP satellite data indicate that the modeled cover is on the average 40-50% larger that the observed one. An abundance of thin cirrus at high altitudes detected by HIRS but not by ISCCP, indicates the importance of representing both thin and thick cirrus in a GCM.

In order to better understand the impact that tropical cirrus have on the climate, we are developing an alternative cirrus scheme that takes into account both model large scale dynamics and basic microphysical properties of cirrus. Accordingly, the rates of large-scale water vapor convergence and convective detrainment are assumed to be largely responsible for determining the actual IWC. Homogeneous nucleation is expected in clear sky, while heterogeneous nucleation is expected otherwise. These two assumptions determine the relative humidity thresholds used for cirrus formation. The mean crystal effective radius and the partitioning between small and large ice crystals is evaluated using expressions for the ice crystal size spectra retrieved from the in-situ observations in the tropics. The IWC and the corresponding effective radius are then used in radiative transfer calculation.

The effects of employing the new cirrus scheme in NCAR CCM3 have been evaluated by a comparison between the monthly means of cirrus cover and IWC for the spring 1993 in the tropical Western Pacific. Preliminary results show a decrease from around 90% cirrus cover created in the warm pool region by the standard CCM3 to around 45% at approximately 11km altitude; a decrease by 10-20% at the level of 13km, and no significant change at the level around 14.5km. At the cirrus top level, usually around 16.5km, only subvisible cirrus, with visual optical thickness less than 0.1 are created, compared to approximately 40% coverage produced by the standard CCM3 cirrus treatment. The mean IWC is increased by around 50% compared with the standard version. If the mean ice precipitation is added to IWC the total condensate is three times higher on average and compares much better with the in-situ observations during CEPEX. These preliminary results indicate that the use of a more sophisticated cirrus scheme will change the frequency and microphysical properties of modeled cirrus. We plan to evaluate the impact of these results on the relative humidity fields and the outgoing long-wave radiation fields in this region as well.