Using ERA-5 reanalysis to characterize extreme rainfall in Italy

Heavy precipitation events are likely to become more frequent in most parts of Europe; yet, records of hourly precipitation are often insufficient to study trends and changes in heavy rainfall. Atmospheric reanalyses are an important source of long-term meteorological data, often considered as a solution to overcome the unavailability of direct measurements. The reanalysis procedure makes use of a large amount of heterogeneous historical observations, both sensed and remotely measured (in situ, satellite, etc), assimilated within a dynamical model to reconstruct the state of the atmosphere, land surface and oceans in the past. Among the available reanalyses, the ERA5 dataset released by the ECMWF, can be considered one of the state-of-the-art products. Atmospheric and surface variables are provided hourly, from 1950 to almost real time, with a horizontal resolution of 31 km. The land model of the ERA5, driven by the downscaled meteorological forcing from the lowest ERA5 model level, and with an elevation correction for the thermodynamic near-surface state, is also used to derive the ERA5-land dataset, characterized by a higher spatial resolution (9 km) and finer precipitation distribution details.

In this paper, data from the ERA5-land reanalysis dataset were used to characterize the 1-hour maximum yearly rainfall values in Italy. Specifically, 3215 grid series of 1-hour rainfall for the period 1950-2020 have been first extracted. Then, for each grid series the 71 1-hour maximum yearly rainfall values have been evaluated. Moreover, the time frame 1950-2020 has been divided into several intervals, and for each one, the frequency distribution of the months recording the annual maxima was calculated. Finally, a cluster analysis has been performed to evaluate the area with a similar monthly distribution of these values. Results showed that, considering the data over the whole of Italy, the monthly distribution of occurrences of annual maxima of 1-hr rainfall is characterized by a peak in September occurring in all the time windows considered. Furthermore, clustering cells with a similar distribution of annual hourly rainfall maxima, using k-means, allowed to identify three groups characterised by different months with the highest frequency of occurrence of the maximum.

This work was funded by the Next Generation EU - Italian NRRP, Mission 4, Component 2, Investment 1.5, call for the creation and strengthening of 'Innovation Ecosystems', building 'Territorial R&D Leaders' (Directorial Decree n. 2021/3277) - project Tech4You - Technologies for climate change adaptation and quality of life improvement, n. ECS0000009. This work reflects only the authors’ views and opinions, neither the Ministry for University and Research nor the European Commission can be considered responsible for them.