A Deep U-Net Framework for Flood Hazard Mapping Using Hydraulic Simulations of the Wupper Catchment

arXiv cs.LG / 4/24/2026

📰 NewsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper proposes a deep-learning surrogate model to rapidly predict flood hazard outcomes that traditionally require computationally expensive hydraulic simulations.
It builds and optimizes a U-Net-based architecture, focusing on patch generation and data handling to approximate a hydraulic model’s predictions of maximum water levels over a grid.
The approach was evaluated using hydraulic simulations of Germany’s Wupper catchment, with results reported as comparable to the traditional simulation outputs.
The work suggests deep learning can provide a computationally efficient alternative to conventional hydraulic modeling for flood prediction and hazard mapping workflows.

Abstract

The increasing frequency and severity of global flood events highlights the need for the development of rapid and reliable flood prediction tools. This process traditionally relies on computationally expensive hydraulic simulations. This research presents a prediction tool by developing a deep-learning based surrogate model to accurately and efficiently predict the maximum water level across a grid. This was achieved by conducting a series of experiments to optimize a U-Net architecture, patch generation, and data handling for approximating a hydraulic model. This research demonstrates that a deep learning surrogate model can serve as a computationally efficient alternative to traditional hydraulic simulations. The framework was tested using hydraulic simulations of the Wupper catchment in the North-Rhein Westphalia region (Germany), obtaining comparable results.