Tropopause transition layer cooling and tropical cyclones

Over the past 25 years or so, there has been a strong increase in almost every metric of tropical cyclone activity in the North Atlantic, and in some metrics of intense cyclones elsewhere around the globe. These have exceeded expectations based on simply theory applied to observed increases in sea surface temperature over the period, leading to questions about whether the observed increase in activity is related to the observed warming of the ocean. In this study, we apply a recently developed tropical cyclone downscaling technique to several atmospheric general circulation models (AGCMs) driven only by observed sea surface temperatures over the last 100 years or so. We show that, for the tropical Atlantic, such downscalings capture the observed variability of tropical cyclones up to about 1970, but largely fail to capture the recent large upswing in activity. We show that this is directly attributable to the failure of the ACGMs to capture the large increase in potential intensity present in reanalyses, in spite of the fact the observed increase in sea surface temperature is prescribed in the models. Finally, we show that the large increase in potential intensity over the North Atlantic over the past quarter century is attributable both to the increase in sea surface temperature and to the decline in temperatures near the tropical tropopause, which together lead to a precipitous decline in hurricane outflow temperatures. We speculate on reasons behind the failure of AGCMs to capture the near-tropopause cooling.