University of Colorado
Boulder, CO 80309-0216
The importance of clouds in the climate system is well-known but poorly understood. Modeling and observational studies have suggested that there may be positive feedbacks associated with certain cloud processes, but it is not known how strong these feedbacks are in the context of the overall system. Examples include ice microphysics feedback, as shown by Liou's model, and the relationship between SST and cloud cover in the tropics (Hanson, 1991), which is the focus of this research.
The traditional description of clouds in climate models has been deterministic; the resolved-scale variables in the model are used to diagnose or predict specific cloud distributions. An alternative approach would be to avoid the issue by inferring radiative properties for nondeterministic "fuzzy" cloud populations (nondeterministic, in the sense that there are typically not enough grid box variables to uniquely determine the nature of the cloud). In this approach, the radiative properties of the cloud field are obtained by averaging over probability distributions of clouds with different cloud fractions.
Also, the traditional focus on clouds had proven to be distracting from the more general problem of the radiative properties of water substance (in all three phases) on scales relevant to climate models. Adopting a "fuzzy concept" approach to the problem of clouds in climate models could avoid this distraction and, potentially, lead to increased understanding of how the climate works. In some ways, these approaches to the cloud/climate problem are complementary to observational methods that rely on multifractal-based pattern recognition; in this case, the fuzzy logic is applied from the cloud process perspective.
The combination of these approaches is attempted here for clouds in the subtropical marine boundary layer, with encouraging, albeit highly preliminary, results. Boundary layer models generally require either a specification of, or a closure for, cloud cover. However, in the former mode they provide physically realistic solutions which satisfy external requirements for a limited range of cloud covers. It could therefore be assumed that solutions within this range have an equal probability of occurrence and that the relevant results (e.g. radiative properties) could be obtained by averaging over the range. This is illustrated using a very simple, empirically-based parameterization of cloud transmittance from the Derr et al. paper, in a variety of cloud populations. (Figures 7.1 and 7.2)
Treating water in the atmosphere in a unified fashion is a somewhat more daunting concept, but the mixing line approach of Betts offers one possible point of attack. A model due to Betts and Ridgeway (basically a single Hadley cell; figure 7.3) is used here to illustrate how this could be done. Comparisons between this model and the simple, empirical model mentioned above, while imperfect, suggest consistency and the need for follow-up studies. Also, the positive feedback between SST and cloud cover (for subtropical marine boundary layer clouds) is shown to be consistent with these model approaches.
References
Hanson, H. P., 1991: Marine Stratocumulus Climatologies. Int. J. Climatology, 11,147-164.
Derr, V. E., R. S. Stone, L. S. Fedor, and H. P. Hanson, 1990: Parameterization for the shortwave transmissivity of stratiform water clouds based on empirical data and radiative transfer theory. J. Atmos. Sci., 47, 2774- 2783.
