Tropical Cyclone Genesis Potential Indices in a New High-Resolution Climate Models Ensemble: Limitations and Way Forward
The eiLab participates in a study in collaboration with colleagues from the Fondazione Centro Euro-Mediterraneo sui Cambiamenti Climatici, to address the questions of whether Genesis Potential Indices (GPIs) are still relevant in the era of Tropical Cyclones-permitting climate model ensembles, and whether they have sufficient predictive skills.
Tropical Cyclones (TC) are extreme weather events which cause large damage when they hit inhabited areas. Understanding future cyclone activity in advance is very important to limit the damage. Using cyclone data from a new set of climate model projections, results show that genesis potential indices are not always accurate in explaining future cyclone activity. The contribution discuss how future research could help us improve the understanding of the drivers of tropical cyclone activity, and therefore the accuracy of genesis potential indices.
In the study published in Geophysical Research Letters, the authors suggest to reinforce statistical methodologies based on Machine Learning (ML) techniques to alleviate some of the mentioned problems (e.g., by applying regularization techniques or transfer learning). Likewise, the results provide a systematic and robust benchmark of current GPIs skills, against which newly developed ML-enhanced TC genesis indices can be effectively tested. On the other hand, progress in the theoretical understanding of tropical cyclone genesis would provide more robust indicators of TC activity.
Read more here: Cavicchia L., Scoccimarro E., Ascenso G., Castelletti A., Giuliani M., Gualdi S. (2023). Tropical cyclone genesis potential indices in a new high-resolution climate models ensemble: Limitations and way forward, Geophysical Research Letters, 50(11), e2023GL103001. https://doi.org/10.1029/2023GL103001
