Although climate models may have limited application in long-term forecasting, they can be robust tools for understanding some of the basic mechanisms of the climate system. The use of models in prediction is often emphasized in popular coverage of climate science, yet the success of models in research often employs these computational methods to diagnose, rather than predict, aspects of the immensely complex climate system.

One of my current research projects illustrates this use of modeling to address the impacts of climate warming. In response to a global increase in greenhouse warming, the wind patterns and general atmospheric circulation of the atmosphere will be altered. (In particular, warming could lead to a weakening and latitudinal broadening of the trade winds in the tropics.)

To explore this problem, I am using a 3-dimensional model of the atmosphere. The ocean is represented by a 50 meter “slab” (i.e., the ocean can absorb heat, but there is no oceanic circulation) and all terrain is absent. The model also does not differentiate between different atmospheric gases; rather, the effects of increased carbon dioxide are simulated by making the atmosphere thicker or thinner.

I mention this to point out that the model is by no means a realistic recreation of the atmosphere-ocean system. Instead, this simple model is actually an easier way to tease apart some of these basic responses. The results of this study will not be able to provide an accurate prediction for how much the tropical circulation will change, but (hopefully) it will be able to provide a mechanism as to why these changes occur. By treating the problem in a slightly simplified manner and acknowledging the limitations of the technique, climate modeling is an unparalleled approach in uncovering, piece by piece, some of the complex interactions of the atmosphere.

The success of models as diagnostic tools does indeed provide a basis for discussing the impacts of future climate warming, and some of these results may have implications for how the climate system will respond. Diagnostic results can be coupled with observations and perhaps used to predict impacts in the near future, but this is a subtle-yet-important difference from the development of a full-blown climate prediction model. Both tools are used in research, with varying degrees of success, but this important distinction often gets lost in popular and political criticism of scientific climate tools.

All tools have limitations, and therefore all results (models and observations) have inherent limits. While being forthright about these limits is necessary, it is equally crucial that these results not be dismissed outright. We will never have any scientific result about the climate system–or any other science, for that matter–that is free from caveats, yet even with our inherent constraints there is a great wealth of things we can learn.