Analysis of Tropical Cyclone Genesis Indices for Climate Change using the HIRAM model

Here we test the performance of a recently developed tropical cyclone genesis index (TCGI) to represent the behavior of model tropical tropical cyclones in high-resolution climate model simulations of present and future climates. The TCGI was based on a Poisson regression between observed tropical cyclone data and environmental variables from 2 reanalysis products. The high-resolution climate model is the GFDL HIRAM with 50km resolution forced and prescribed sea-surface temperatures (SST). The SST forcings considered are: present day climatological SST, SST anomalies from 8 AR4 models and the multimodel mean SST anomaly of 18 AR4 models. In each of the experiments, model tropical cyclones are detected and tracked, and a model track dataset is constructed. First, the reanalysis-based TCGI is calculated using the model variables, and its ability to predict changes in the number of model tropical cyclones between present day and future simulations tested. Second, to deal with systematic differences between model and reanalysis, we develop a HIRAM-based TCGI using the model simulation of present day tropical cyclones and environmental fields. The HIRAM TCGI is then compared with the reanalysis TCGI, and their ability to predict changes in number of tropical cyclones globally and by region is compared. Diagnostics of the factors responsible for the changes in tropical cyclone frequency in the model are performed using the HIRAM TCGI. Finally, a TCGI is computed for each climate change scenario and compared with the present day HIRAM TCGI. The differences in the regression coefficients among these indices reflect differing sensitivities of model tropical cyclone genesis to the environmental variables, as well as differences among the different climate change forcing scenarios.