Combining convection-permitting reanalysis with satellite overshooting top detections for investigating hailstorm environments over south-central Europe

Among severe weather events generated by deep convection, hail represents one of the most hazardous perils for agriculture, buildings, and properties. Moreover, in response to global warming, hailstorm frequency and severity are expected to increase in Europe. Therefore, a better understanding of hail hazards and risks is becoming increasingly crucial. However, major hurdles are posed by direct observations, by being heterogeneous, temporally limited, and scarce, as well as by numerical simulations, which generally lack sufficient detail to represent hailstorm dynamics properly. As a possible solution, hail likelihood can be indirectly assessed by combining multiple data sources, such as remote-sensing detections and ambient numerical predictors.

The new high-resolution reanalysis SPHERA (High rEsolution ReAnalysis over Italy), developed at ARPAE, is considered to describe hail-favoring environments over Italy and nearby countries. SPHERA is dynamically downscaled from ERA5 and driven by the model COSMO at the convection-permitting resolution of 2.2 km. A set of predictors is selected and combined with Overshooting cloud Top (OT) satellite detections, constituting a reliable proxy for hail. OTs are automatically detected with a probabilistic algorithm (NASA) from geostationary Meteosat Second Generation SEVIRI infrared images. However, not all OTs are linked to hail. Hence, a filter based on their surrounding ambient conditions is developed to retain only OT occurrences with the potential for hailstorm formation. For this purpose, ESWD (European Severe Weather Database) crowdsourced hail reports are coupled with SPHERA proxies to characterize hailstorm environments.

The analysis is performed over 2016-2020. Hence, the primary intent is to present the methodology rather than a comprehensive hail frequency characterization. More than a quarter of non-hailing OTs are removed, mainly over the Mediterranean sea and complex-topography areas. Maximum hail likelihood characterizes pre-Alpine regions and the northern Adriatic sea around 15 UTC in June-July, agreeing with previous European hail climatologies. The hit rate of ESWD reports with OTs exceeds 60%, i.e., ~20% more than the previous coupling of non-probabilistic OT detections with ERA-Interim over Europe. Different ambient characteristics are revealed by separating hit/miss reports for small/large hail events. Most hits present hailstones with diameters ≥3 cm, suggesting a better suitability of the method in case of severe hailstorms. Further, missed small-hail reports show peculiar environmental signatures characterized by lower instability, less wind shear, and colder atmospheric profiles.

These results suggest a promising avenue to enhance hailstorm events identification. In addition, further extension of the analysis should shed more light on the climatology of hailstorms.