Many microclimates can be studied and altered under laboratory conditions. Such control over the essential variables, however, is not obtainable for research into climates and climatic change on a larger scale. Instead scientists try to simulate large-scale climatic conditions by means of elaborate computer programs, called general circulation models. These computer models commonly separate the atmosphere into several layers and break it down into thousands of rectangular grid units. Scientists then use mathematical equations to represent as closely as possible the basic atmospheric conditions that are taking place.

The ocean, clouds, and the numerous biological and chemical processes that are taking place in the environment also profoundly influence the climate. Scientists therefore enter attempted simulations of such factors - and their feedback effects - into the models as well. They can then alter variables such as temperature, atmospheric gases, and absorption of solar radiation to see what effects these changes have on their models.

The better the simulations are of actual environmental situations, however, the longer it takes to run these complex programs on computers. Only the fastest supercomputers can handle advanced atmospheric models, and even they may require thousands of hours to complete a run.