Projected 21st century changes in the length of the tropical storm season in downscaled simulations and HIRAM

We investigate projected changes in the length of the tropical cyclone season due to greenhouse gas increases in two models that capture the relevant features of the observed annual cycle of tropical cyclones. Both models, a downscaling technique in which incipient vortices are randomly seeded into the large-scale atmospheric conditions and a high resolution atmospheric model (HIRAM), use output from the global climate models of the CMIP3 or CMIP5 suites, and differ in their projections for future tropical cyclone season length. The downscaling model on CMIP5 data projects a longer season in most basins, while the downscaling model on CMIP3 data and HIRAM on both CMIP3 and CMIP5 project a shorter tropical cyclone season in most basins. Season length is measured by the number of consecutive days that the data is greater than a fixed threshold, but other metrics give consistent results. There is also good agreement between the projected decrease in season length for tropical cyclones and a genesis index of tropical cyclones for HIRAM with CMIP3 data. While the annual mean frequency change accounts for much of the change in season length, the non-annual mean changes also contribute, demonstrating the importance of changes during the transition months. Finally, using a Fourier metric for seasonality, which doesn't depend on the annual mean, we find a shift of storms to later in the year in many basins for the downscaling model, but less consistent timing changes for HIRAM.