Small Processes Make a Big Difference in Model Outcomes
Cole, J. N., University of British Columbia
General Circulation and Single Column Models/Parameterizations
Cloud Modeling
Cole, J. N. S., H. W. Barker, D. A. Randall, M. F. Khairoutdinov, and E. E. Clothiaux (2005), Global consequences of interactions between clouds and radiation at scales unresolved by global climate models, Geophys. Res. Lett., 32, L06703, doi:10.1029/2004GL020945.
To address small scale atmospheric processes, cloud system resolving models divide a typical GCM grid cell into 64 columns, each approximately 4 km wide.
Typical climate models are based on a grid pattern that blankets the globe, with each square of the grid covering an area of about 300 km by 300 km. Each square then extends vertically into the sky, resulting in approximately 8,192 "columns" of atmospheric properties. This large-scale grid is used because of the extensive computational time needed for smaller grids. However, because cloud and radiative effects occur on small scales, their processes are parameterized. This means the effects are expressed as a set of simple approximate relationships, rather than through explicit equations that simulate the actual physical processes in the climate model. In a recent study published in Geophysical Research Letters (March 2005), scientists sponsored by the Department of Energy's Atmospheric Radiation Measurement Program found that these small-scale processes may have important impacts on the climate and climate change simulated by global climate models (GCMs).
In their study, the researchers replaced the conventional GCM cloud parameterization in each of the GCM's 8,192 columns with a 2D cloud system resolving model (CSRM) which used 64 columns, each with a horizontal grid length of 4 km. They then calculated the radiative heating rates independently for each of the 64 CSRM columns.
In addition to the control, they carried out three model experiments using the nested model system, also known as the Multi-scale Modeling Framework, to explore the sensitivity of the model to the coupling between radiative heating and model dynamics. Model experiment runs were conducted for a single season due to computational constraints.
Results of the comparison study showed that experiments using the independently calculated heating rate profiles produced a similar large-scale model climate state. Experiment runs using a single averaged heating rate profile produced a different model climate state. This suggests that small-scale variability in cloud scale properties feeds up and modifies the large-scale average states.
