Projections of tropical cyclones and their formation regions in CMIP6 climate models

Abstract This study evaluates changes in tropical cyclone (TC) characteristics under different future emissions scenarios in a large set of climate models from the Coupled Model Intercomparison Project Phase 6 (CMIP6). The Okubo–Weiss–Zeta parameter (OWZP) TC detection method is used to identify large-scale conditions conducive to TC formation. By comparing TC detections from CMIP6 historical simulations (1984–2014) with projections for the end of the twenty-first century (2070–2100) under three shared socioeconomic pathways (SSP1-2.6, SSP3-7.0, and SSP5-8.5), we assessed the changes in TC frequency and associated formation regions. TC formation regions during the mid-summer cyclone season of each TC basin were also analysed across the three pathways. Among the 19 CMIP6 models assessed, 12 models are found to simulate TC frequency ( i.e. , the mean annual totals within ± 50% of actual observations) and geographical distribution reasonably well. The remaining seven models captured only 19–27% of the observed TC numbers. For all the emission scenarios, most of the models with reliable TC climatology show a decline in global TC frequency. For SSP5-8.5 scenario, all 12 models project a global decrease with reductions ranging from 6 to 19%. In the SSP3-7.0 scenario, eight of the 14 models show global declines ranging from 1% to 42%, while nine out of 12 models under SSP1-2.6 scenario project global reductions from 2% to 12%. Notable variability in the magnitude and direction of changes is observed among models and basins, especially in the Northern Hemisphere. Analysis of TC formation regions indicates slight poleward expansion in the northwestern Pacific and a longitudinal expansion of the eastern North Pacific basin toward the central Pacific. There is also potential for latitudinal narrowing of TC formation regions in the South Indian and southwest Pacific basins across all scenarios. These results are consistent with those from earlier-generation climate models, improving the robustness of projections for a global reduction in TC frequency.