A return period-based early warning index for extreme precipitation

Extreme precipitation events often lead to local or downstream flooding and can have devastating impacts from infrastructure damage to loss of life. As climate change progresses, these events have become increasingly frequent and intense, posing significant challenges to societies. While warning systems for fluvial flooding such as the Global Flood Awareness System (GloFAS) exist, localised downpours over impermeable urban areas remain difficult to predict. Even in countries with advanced early warning systems, precipitation intensity is often underestimated, leading to misrepresentations of potential impacts. Further, many localised flood events driven by very intense precipitation are not predicted at all.  

Here, we present a novel warning index that predicts the likelihood of extreme precipitation and targets localised urban and pluvial flooding, thereby addressing a gap in existing warning systems. The presented warning index is based on a novel set of return period forecasts for extreme precipitation, that enable a correction of model biases. The index then estimates a risk through the mapping of likelihood and potential impacts, incorporating a fuzzy neighborhood approach that accounts for displacement errors in the prediction of extreme events as a function of lead time. Through this risk approach, the warning index aims to capture extreme but less probable events to improve the warnings’ reliability. Here, we present results for the 30 activations from the Copernicus Emergency Management System Rapid Mapping (CEMS RM) in 2024. The index shows reliable and actionable warnings for localised flooding events, offering significant advancements in risk management and preparedness for extreme precipitation impacts. 

This work was developed in the context of the Horizon Europe CENTAUR project https://centaur-horizon.eu/