Rapid development of impact-based national flood early warning system

Denmark, as a low-lying country, is subject to acute flood risks including urban flash floods (cloudbursts) from intense convective storms, fluvial flooding from heavy rainfall and extreme sea levels and storm surge along more than 7,300 km of coastline. Motivated by the extensive flooding in Denmark in 2020 and the devastating 2021 floods in Central Europe, the Danish Meteorological Institute (DMI) became the national authority for flood forecasting and warning, tasked with developing an operational system to forecast storm surge, pluvial and fluvial, flood risks.

To support anticipatory early actions, the key goals were to provide an operational 24/7 capability to issue timely, accurate and reliable flood forecasts, early warnings and associated flood impacts. With significant pressure to develop and deploy operational tools to support Denmark's emergency authorities within the first 18 months, we adopted pragmatic and simplified modeling approaches, balancing resolution, complexity, data availability and computational efficiency. 

We present the rapid development of operational capabilities to support the local and national emergency services, for storm surges, pluvial and fluvial flood events in Denmark, guided by initial consultations with the emergency services. Within the first year we developed, together with Scalgo, a real time flood mapping service for elevated sea levels and storm surges. This covers the entire Danish coastline, based on hourly water level forecasts, 5 days ahead. This service became operational in October 2022, immediately prior to the 2022 storm surge season. The timely launch allowed us to evaluate the performance of this service against the 100+-year storm that hit the coasts of Denmark in October 2023. A new cloudburst flood mapping service was developed, also together with Scalgo, including a new topography-based flood mapping approach to account, in a computationally efficient way, for effects of infiltration and urban drainage systems. This service became operational for all of Denmark in May 2023, at the beginning of the 2023 cloudburst season. Feedback from meetings with the emergency services during 2024 confirmed the value of this mapping service. Finally, for fluvial flooding, a rule-based warning system was initially developed. This approach uses statistical analysis of river levels and precipitation thresholds and exploits a newly developed national inventory of historical floods. Manual warnings, to the emergency services, based on this approach began in July 2023 focussed on high-risk areas and stations with good quality water level data. Our own evaluations of these new capabilities during the first year of operations were shared, in a series of workshops, with the local and national emergency services. The workshop objectives were to obtain their feedback and to understand their needs for the next development phase. We discuss how this rapid process for operational implementation of a national system was achieved. This includes our initial evaluations, operational challenges and solutions, as well as future end-user involvement and development plans. We are currently developing both machine learning approaches and conceptual hydrological modelling to extend our forecasting capability towards multi-model fluvial forecasting.