Lessons on tropical cyclone genesis from paleoclimate simulations

There are many questions about how the climatology of tropical cyclones may change with climate, but the observational record is ill-equipped to address them: it is restricted by a short length, subjective storm intensity estimates, and questions about its completeness. Analysis of 21st century model simulations and theoretical developments have provided some guidance, but it will be many more years before sufficient data exist to finally confirm or refute their predictions of the future. But there is a wealth of data emerging from a variety of paleoclimate approaches, and here we present some avenues for learning from model simulations of the late Pleistocene and Holocene. Atmospheric composition remained fairly steady between the mid-Holocene (6ka) and the preindustrial era, but the latitudinal distribution of incoming solar radiation changed substantially as the perihelion of Earth's orbit progressed from September (6ka) to January (today). The magnitude of the differences at 6ka fades over the late Holocene. We examined the simulations of the paleoclimate intercomparison modeling project (PMIP2) to investigate how the dynamic and thermodynamic factors known to be important for tropical cyclogenesis change in response to cold climates (last glacial maximum) and solar radiation changes (6ka). We find that despite the increase in Northern Hemisphere storm season solar radiation at 6ka, there is a robust decline in the favorability of thermodynamic variables important for genesis (e.g., maximum potential intensity and measures of tropospheric entropy deficits). These changes appear principally related to the differential heating of the ocean surface and free troposphere, but the geographic distribution of changes is not uniform, and model variability is high in the central Pacific. In addition to the genesis factors, we are examining the characteristics of model storm tracks in the latest CESM paleoclimate runs. This high frequency, high resolution output also feeds various downscaling techniques that we use to study model track and intensity changes. We compare our results to the available proxy record from sedimentary cores and to ENSO proxies, which have been used previously as a proxy for Atlantic wind shear.