An event-oriented database of meteorological droughts in Europe based on spatio-temporal clustering

Carmelo Cammalleri*, Juan Camilo Acosta Navarro, Davide Bavera, Vitali Diaz, Chiara Di Ciollo, Willem Maetens, Diego Magni, Dario Masante, Jonathan Spinoni, Andrea Toreti

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

9 Downloads (Pure)


Droughts evolve in space and time without following borders or pre-determined temporal constraints. Here, we present a new database of drought events built with a three-dimensional density-based clustering algorithm. The chosen approach is able to identify and characterize the spatio-temporal evolution of drought events, and it was tuned with a supervised approach against a set of past global droughts characterized independently by multiple drought experts. About 200 events were detected over Europein the period 1981-2020 using SPI-3 (3-month cumulated Standardized Precipitation Index) maps derived from the ECMWF (European Centre for Medium-range Weather Forecasts) 5th generation reanalysis (ERA5) precipitation. The largest European meteorological droughts during this period occurred in 1996, 2003, 2002 and 2018. A general agreement between the major events identified by the algorithm and drought impact records was found, as well as with previous datasets based on pre-defined regions.

Original languageEnglish
Article number3145
Number of pages9
JournalScientific Reports
Issue number1
Publication statusPublished - 2023

Cite this