Basin dependencies of tropical cyclone genesis environment and possible future changes revealed by machine learning methods.

Tropical cyclone (TC) genesis mechanisms remain debated, complicating predictions of climate change impacts. This study uses principal-component analysis (PCA), confidence ellipses, and correlation circles to analyze TC genesis environments across ocean basins. Results show that TC genesis is basin dependent, except in the North Atlantic (NA), where absolute vorticity primarily drives differences in genesis locations. Ocean basins are categorized into three groups based on PCA, and three MaxEnt machine learning (ML) models are developed to predict TC genesis under future scenarios. The ML models consistently project robust basin-specific TC genesis trends, demonstrating their utility in such studies. A multivariate environmental similarity analysis indicates significant climate change impacts on TC genesis environments globally, with the weakest changes in the NA. These findings underscore the critical role of absolute vorticity in TC genesis and highlight basin-specific differences in future environmental changes.