The Future of (Very) Large Hail Globally: Application of the AR-CHaMo models to the CMIP6 Ensemble

The Additive Logistic Regression Models (AR-CHaMo) for large (> 2 cm) and very large hail (> 5 cm) were applied to the CMIP6 multi-model ensemble to assess the future global occurrence of hail (2015-2100) under the high emission Socio-Economic Pathway scenario SSP5-8.5. The analysis focuses on three warming levels: 1.5°C, 2.0°C and 3.0 °C above pre-industrial levels. Under future climate scenarios, the locations of the hail frequency maxima are projected to remain consistent, with Northern Argentina and the South American tri-border region continuing to show the highest activity, followed by the US Great Plains and South Africa. The projected trends of large and very large hail differ considerably on a global scale: hail > 2 cm is expected to decrease in frequency across most regions, whereas hail > 5 cm is projected to increase in most hail-prone areas. Large differences are present across the US Plains and the tri-border region of South America, where hail > 2 cm is expected to decline, while hail > 5 cm exhibits strong positive trends. The presentation will analyse the physical drivers of these trends to understand the origin of the differences with the historical trends that resulted from the application of AR-CHaMo to ERA5 (1950-2023), especially across South America. In Europe and the Middle East, widespread increases in the frequency of hail > 5 cm are projected, while hail > 2 cm is expected to decrease in the southern Mediterranean and across northern Africa. When comparing historical and future trends, only one region shows consistency in the sign of the trends for both hail > 2 cm and > 5 cm: northern Italy and the surrounding Alpine region. This finding fits with the record-breaking hailstorms of 2023 and supports high confidence in the projected increase in hail events in this region.