Clouds have a complex and pivotal role in the climate system, both at local and global scales, and many aspects of their behavior are not well understood. Because cloud formation results from a concert of atmospheric processes -- e.g., moisture convergence, upward motion, radiative cooling, and surface evaporation -- numerically simulating and predicting clouds is a challenge. This is painfully apparent in comparisons of results from general circulation models (GCMs) (e.g., Cess et al, 1990; and Zhang et al, 1994). Parameterization schemes for processes involved in cloud formation vary widely among these models, yielding global cloud fractions that differ by a factor of nearly 2. This result underscores the need for improved cloud-radiation interactions in GCMs. Not only are clouds difficult to model well, but they are also a challenge to observe from the surface, space, and aircraft. The Earth Radiation Budget Experiment (ERBE; Ramanathan et al, 1989), for example, has contributed much to our understanding of cloud radiative forcing over the globe, but under certain conditions where cloud detection with only two broad channels is difficult, ERBE results are uncertain. This is particularly true in snow- and ice-covered regions.
Cloud radiative forcing (CF) -- defined as the difference in net radiation, either longwave (CFL) or shortwave (CFS), between clear-sky and cloudy conditions (Ramanathan et al., 1989) -- is a useful concept for evaluating the effect of clouds on the surface and top-of-the-atmosphere (TOA) radiation budget and the cloud-radiation feedback. The work presented here attempts to shed light on which cloud variables are most important in determining CF, and thus which should be the focus of efforts to improve modeling and observing methods. Using a forward radiative transfer model, I examine the dependence of CF at the surface and TOA on cloud properties (height, phase, optical depth) and on polar surface conditions (albedo, snow type, ice type). Results demonstrate that just knowing cloud amount and how it might change is not enough: the bulk cloud properties are also required to simulate accurately the role of cloud-radiation interactions in a changing climate