From 1970 to 2001, the global average annual number of tropical cyclones was 94, with a standard deviation of 10. The stability of the annual mean would seem to be at odds with the fact that genesis is a complex process with low predictability. We hypothesize that the stability of global tropical cyclone activity results from a sensitive dependence of genesis rates on potential intensity (loosely speaking, SST) coupled with a corresponding sensitivity of global SST to global tropical cyclone activity. If correct, this hypothesis would explain both the stability of global tropical cyclone activity and the observed stability of tropical SST over most of the geological record, including such periods as the early Eocene, when tropical SSTs were only a few degrees warmer than today even while the poles were as much as 20 C warmer.

As a test of the first part of this hypothesis, we perform a statistical analysis of the relationship between spatial and temporal variability of observed genesis rates with a limited number of predictors, including shear and potential intensity. We find that the genesis rate varies as the cube of the potential intensity.

To explain the sensitive dependence of genesis rates on potential intensity (SST), we postulate that tropical cyclones are the result of a subcritical instability of the tropical atmosphere-ocean system. If this is the case, then tropical cyclones should form spontaneously in a warm enough world. To test this proposition, we run a cloud-resolving model (CRM) in a doubly periodic domain into a statistical state of radiative-convective equilibrium, with fixed, specified SST. We observe that for typical tropical SSTs, the radiative-convective equilibrium state is stable, and tropical cyclones do not form spontaneously. But when the SST is greater than about 35 C, tropical cyclones do develop spontaneously, demonstrating that they do indeed constitute a subcritical instability at 30 C in the model. This in turn suggests that increasing SSTs in the presence of constant weather noise will result in increased frequency of tropical cyclones.