Distinct Future Changes and Biases of Tropical Cyclone Activities Between Fully-Coupled and Atmospheric-Only Models in the CMIP6-HighResMIP

Abstract

Most models project a decreasing trend in global tropical cyclone (TC) frequency under future climate scenarios. However, discrepancies in TC activity projection between atmosphere-only and full-coupled models remain unclear. Here, we utilize the CMIP6-HighResMIP data to investigate future changes in TC activity, including TC genesis frequency (TCGF) and track density (TCTD), in response to climate warming and El Niño-Southern Oscillation (ENSO). Results reveal significant variability in TC activity projections between the two modeling approaches, reflecting substantial uncertainties. Fully-coupled models projected an El Niño-like sea surface temperature (SST) warming over the equatorial Pacific, while atmosphere-only models demonstrated a La Niña-like warming pattern. These contrasting SST responses drive diverse atmospheric circulation anomalies, eventually contributing to distinct TCGF and TCTD changes under climatology and during ENSO phases for future projections. Further detailed analysis reveals that the El Niño-like or La Niña-like warming patterns belong to two extreme conditions. It shows a strong spread of simulated SST biases across different models. The accurate projection and simulation of SST magnitudes and ENSO patterns are key to reducing the diverse behaviors of TC activities among these models. We detected three crucial regions, the central Pacific, the subtropical eastern North Pacific, and the equatorial eastern Pacific, that could modulate TC activity in response to ENSO in numerical models. These findings provide critical insights into how global warming and ENSO modulate TC activities, highlighting the importance of model configuration in predicting future TC behavior.