Complementing drought impact databases with multilingual online sources: A European gap-filling experiment

Abstract. The use of drought impact databases has been increasingly recognized as essential for impact-based drought monitoring and risk assessment. The European Drought Impact Database (EDID) is the first continent-wide effort to systematically compile drought impact information across sectors and countries. However, the integration of existing regional and national datasets has revealed substantial spatial and temporal gaps, particularly for recent drought events and underrepresented countries. To address these deficiencies, we developed and evaluated a semiautomated retrospective gap-filling strategy that is based on multilingual online text searches, automated translation and classification, and expert validation. The workflow covered all official European Union languages, multiple drought events and nine affected sectors. It yielded several hundred new drought impact records and significantly improved coverage in selected countries. Nevertheless, the retrieval efficiency was low, and its performance varied strongly across languages and sectors. Our findings show that multilingual online data mining can complement drought impact databases but cannot replace sustained local monitoring and expert-driven data collection. Robust continental-scale drought impact monitoring, therefore, requires hybrid strategies that combine automated searches with national and regional inventories. The lessons learned from the EDID gap-filling experiment can guide future operational drought impact monitoring and similar efforts in other natural hazard domains.