Can existing climate models be used to study anthropogenic changes in tropical cyclone climate?

The utility of current generation climate models for studying the influence of greenhouse warming on the tropical storm climatology is examined. A method developed to identify tropical cyclones is applied to a series of model integrations. The global distribution of tropical storms is simulated by these models in a generally realistic manner. While the model resolution is insufficient to reproduce the fine structure of tropical cyclones, the simulated storms become more realistic as resolution is increased. To obtain a preliminary estimate of the response of the tropical cyclone climatology, CO₂ was doubled using models with varying cloud treatments and different horizontal resolutions. In the experiment with prescribed cloudiness, the number of storm‐days, a combined measure of the number and duration of tropical storms, undergoes a statistically significant increase in the doubled‐CO₂ climate. In contrast, a smaller but significant reduction of the number of storm‐days is indicated in the experiment with cloud feedback. In both cases the response is independent of horizontal resolution. While the inconclusive nature of these experimeital results highlights the uncertainties that remain in examining the details of greenhouse gas induced climate change, the ability of the models to qualitatively simulate the tropical storm climatology suggests that they are appropriate tools for this problem.