Projected changes in ENSO modulation of TC activity in global climate models

ENSO diversity (i.e. the differing patterns associated with eastern vs central Pacific ENSO events) is an important factor in determining its impacts on Tropical Cyclones (TCs). Previous research has shown that model biases in simulating ENSO diversity can reflect onto projected ocean warming patterns in response to external climate forcing. Thus, assessing the relationship between ENSO simulation biases and the ENSO modulation of TC activity in the different ocean basins and across models is needed in order to improve confidence in present-day TC simulation and future TC projections. Here, we analyze TC activity proxies in CMIP models and show that the relationship between ENSO phase and the Tropical Cyclogenesis Index (TCGI) is significantly stronger in models with strong ENSO diversity compared to models with weak ENSO diversity; this relationship is projected to strengthen in response to GHG forcing. In addition to the projected changes in the ENSO-TC relationship, we find that depending on their simulation of ENSO diversity, models project different mean changes in TCGI: the changes in models with strong ENSO diversity are concentrated in the western North Pacific and in the North Atlantic, while models with weak ENSO diversity show more uniform changes mostly concentrated in the central North Pacific. Across models, the degree of projected increase in the western North Pacific is associated with the degree of ENSO diversity. The latter has been shown to be related to the projected warming patterns in response to high GHG emissions. Accordingly, we find that the more "El Niño-like" the projected change in mean surface temperatures, the larger the increase in the average TCGI in the western North Pacific. This work highlights the significance of considering ENSO simulation biases in improving global and regional numerical and empirical TC prediction models and assessing future TC risk.