EGU22-4102
https://doi.org/10.5194/egusphere-egu22-4102
EGU General Assembly 2022
© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

Modelling hail probability over Italy using a machine learning approach

Riccardo Hénin1, Veronica Torralba1, Antonio Cantelli2, Enrico Scoccimarro1, Stefano Materia1, and Silvio Gualdi1
Riccardo Hénin et al.
  • 1Fondazione Centro Euro-Mediterraneo sui Cambiamenti Climatici (CMCC), Bologna, Italy
  • 2Generali Italia S.p.A, Mogliano Veneto (TV), Italy

Hail is a meteorological phenomenon with adverse impacts affecting multiple socio-economic sectors such as agriculture, renewable energy and insurance (e.g. Púčik et al., 2019; Martius et al., 2018; Macdonald et al., 2016). The mitigation of the hail-related risk in particularly sensitive regions such as Italy has fostered hail research, aiming at a deeper understanding of the favorable environmental conditions for hail formation and the improvement of hail forecasting skills (Mohr and Kunz, 2013). Nevertheless, one of the major limitations for the study of long-term hail variability is the inherent difficulty in measuring all the hail occurrences and the consequent scarce temporal and spatial coverage of hail observations (Mohr et al., 2015). Therefore, in this study, the Probability Density Functions (PDFs) of several large-scale meteorological variables and convective indices from the ERA5 reanalysis are considered instead, with the aim of describing a conditioned hail probability, following the statistical method by Prein and Holland (2018). Then, the best set of variables to be used as predictors in the hail model are selected with a machine learning approach, based on a genetic algorithm. The model output is an estimation of the hail probability over Italy in the 1979-2020 period, on a 30x30 km grid. The model is validated over the Friuli-Venezia-Giulia region through an independent dataset based on hail pads. The estimated hail probability has been used to characterize the seasonality, long-term variability and trends of the hail frequency and to investigate the potential large-scale drivers of hailstorms over Italy. 

 

REFERENCES:

Púčik, T., Castellano, C., Groenemeijer, P., Kühne, T., Rädler, A. T., Antonescu, B., & Faust, E. (2019). Large hail incidence and its economic and societal impacts across Europe. Monthly Weather Review, 147(11), 3901-3916. doi: 10.1175/MWR-D-19-0204.1.

Martius, O., Hering, A., Kunz, M., Manzato, A., Mohr, S., Nisi, L., & Trefalt, S. (2018). Challenges and recent advances in hail research. Bulletin of the American Meteorological Society, 99(3), ES51-ES54. doi: 10.1175/BAMS-D-17-0207.1.

Macdonald, H., Infield, D., Nash, D. H., & Stack, M. M. (2016). Mapping hail meteorological observations for prediction of erosion in wind turbines. Wind Energy, 19(4), 777-784. doi: 10.1002/we.1854.

Mohr, S., Kunz, M., & Geyer, B. (2015). Hail potential in Europe based on a regional climate model hindcast. Geophysical Research Letters, 42(24), 10-904. doi:10.1002/2015GL067118.

Prein, A. F., & Holland, G. J. (2018). Global estimates of damaging hail hazard. Weather and Climate Extremes, 22, 10-23. doi: 10.1016/j.wace.2018.10.004.

 

How to cite: Hénin, R., Torralba, V., Cantelli, A., Scoccimarro, E., Materia, S., and Gualdi, S.: Modelling hail probability over Italy using a machine learning approach, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4102, https://doi.org/10.5194/egusphere-egu22-4102, 2022.