Articles | Volume 26, issue 1
https://doi.org/10.5194/nhess-26-85-2026
https://doi.org/10.5194/nhess-26-85-2026
Research article
 | 
13 Jan 2026
Research article |  | 13 Jan 2026

Enabling real-time high-resolution flood forecasting for the entire state of Berlin through multi-GPU accelerated physics-based modeling

Shahin Khosh Bin Ghomash, Siqi Deng, and Heiko Apel

Data sets

Imperviousness Density 2018 (raster 10 m), Europe, 3-yearly European Environment Agency https://doi.org/10.2909/3bf542bd-eebd-4d73-b53c-a0243f2ed862

Download
Short summary

Urban pluvial flooding is worsening due to climate change and urbanization, requiring faster forecasts. This study presents RIM2D, a multi-graphics processing unit (GPU) 2D flood model, simulating high-resolution events (2–10 m) across Berlin (891.8 km²) with up to 8 GPUs. Simulations of real and synthetic floods show multi-GPU use is vital for fine-scale, timely forecasts. RIM2D proves operationally viable for urban-scale early warning using modern GPU hardware.

Share
Altmetrics
Final-revised paper
Preprint