Evaluating CMIP6 Models in Simulating Southwest Indian Ocean Tropical Cyclone Activity (1979–2014): Comparison with ERA-5 Reanalysis

Abstract

Tropical cyclones are environmentally and socio-economically destructive phenomena in sub-Saharan. Yet there is still a lack of early warning signs and minimal studies on the simulation of tropical cyclones which attribute more venerability to the societies in the sub-continent. This study investigates the performance of ten global climate models in simulating the characteristics of Southwest Indian Ocean (SWIO) tropical cyclone activities. The model’s simulation was performed for 35 years (1979-2014) and validated against the ERA-5 reanalysis data. The study also considers the influence of large-scale environmental conditions through a calculation of the genesis potential index (GPI). The strength of global climate models was tested through the simulation of the inter-annual variability of tropical cyclones in the SWIO. The results show each CMIP6 model displaying a different signature on simulations. The ten analysed CMIP6 models are capable of reproducing the observed annual mean of genesis potential index (GPI) and its associated terms (environmental conditions) in a given region, but most CMIP6 have a lower genesis potential index (GPI) (including terms) than the reanalysis. The selected CMIP6 analysis indicates that all models produce realistic representations of tropical cyclones over the SWIO, while some underestimate and overestimate tropical cyclone activities such as lifetime maximum intensity (LMI), mean intensity, and genesis. However, models that struggled to resemble ERA-5 reanalysis on tropical cyclone activities were able to simulate the inter-annual variability well relative to the reanalysis. The present study illustrates how different models of horizontal resolution simulate tropical cyclone activities and offers insights on how to improve future projections of tropical cyclone activity over the SWIO region.