EGU24-10635, updated on 08 Mar 2024
EGU General Assembly 2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Heavy-rain Forecasting with the Application of High-density Swarm Network of Optical Rain Sensors and Artificial Intelligence.

Nibesh Shrestha1, Alexander Buddrick1, Benjamin Mewes2, and Henning Oppel2
Nibesh Shrestha et al.
  • 1Nivus GmbH, Business Development, Eppingen, Germany (
  • 2Okeanos Smart Data Solutions GmbH, Bochum, Germany

Heavy rainfall, a prominent consequence of climate change, induces substantial pluvial flooding as the urban drainage systems fail to deal with the water surge. The risks intensify with the cloud-burst rain on a catchment area without any gauge. Especially in topographically complex watersheds, the limitations associated with conventional precipitation monitoring tend to exacerbate. These heavy-rain events, if undetected, pose severe threats, causing extensive damage to the settlements and industries without timely warning.

With a motive to bridge this gap, we present the exemplary development of a cutting-edge AI-supported early warning system and cell detection (now-casting) of heavy rainfall events. Utilizing an IoT-based optical method, we record qualitative rainfall intensity data with a high-density swarm network of rainfall sensors spread across the target region. These data can be immediately used to forecast the path of the rain with the physical optical-flow method. Furthermore, these data are used to train the AI, generating heavy rain forecasts up to 60 minutes before the rain reaches points of interest. This lead time is crucial for citizens and rescue forces to reduce the chaos phase and prepare themselves on time even before the heavy rain cells reach their location and create havoc.

The innovative optical rainfall sensors have been installed and tested in Liederbach am Taunus since the summer of 2022, demonstrating their efficacy and accuracy during the August 2023 heavy rainfall storm event. The system adeptly captured heavy rainfall data, showcasing great potential for early warnings when implemented at a full scale alongside AI applications.

How to cite: Shrestha, N., Buddrick, A., Mewes, B., and Oppel, H.: Heavy-rain Forecasting with the Application of High-density Swarm Network of Optical Rain Sensors and Artificial Intelligence., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10635,, 2024.