This study presents a novel global database, Wikimpacts 1.0, that leverages natural language processing and automated information extraction to systematically capture multi-scale socio-economic impacts of climate extremes over nearly a millennium. By demonstrating strong accuracy and complementary coverage to established datasets, it highlights the transformative potential of AI-driven approaches to expand, refine and democratize climate impact data for research and decision-making.

This paper presents a new Pluvial Flood Index (PFI) for assessing and communicating the hazard from pluvial flash floods caused by intense, localised rainfall. As climate change intensifies heavy rainfall and urbanisation increases surface sealing, pluvial flooding is becoming an increasingly important source of damage and disruption. The PFI combines key hydrological and hydraulic factors into a clear, practical and interpretable measure of pluvial flood hazard. It identifies both the extent and location of potential pluvial flood hazard areas, while remaining accessible for communication with non-specialist audiences. The paper shows how the PFI can support operational flash flood forecasting and warning systems across large regions, as well as broader assessments of pluvial flood susceptibility. It can also help authorities prioritise preparedness and mitigation measures where pluvial flood risk is greatest.