Improving Water Level Predictions in 2D Hydraulic Models: Optimizing Low-Resolution Meshes

Accurately predicting river water levels is essential for effective environmental and water resource management, particularly in flood mitigation, drought forecasting, and infrastructure planning. Two-dimensional (2D) hydraulic models are widely used for simulating water levels, but achieving high accuracy remains a challenge due to uncertainties related to input data, model structures, technical configurations like mesh design, and parameters such as roughness. Mesh configuration plays a pivotal role in shaping bathymetry, influencing discharge capacity, and determining water levels. While high-resolution meshes deliver greater accuracy, they often come at the cost of longer computational times. In contrast, low-resolution meshes are computationally efficient but can introduce significant errors, requiring complex calibrations that may struggle to handle extreme flow conditions effectively.

This study adapts a novel developed algorithm, first for hypothetical river systems, and then to real-world applications. The method adjusts the elevation of individual mesh nodes, ensuring that the flow volume in low-resolution meshes aligns with high-resolution riverbed data. By improving mesh accuracy while maintaining computational efficiency, this innovative approach addresses mesh-related errors and enhances model reliability. The modified low-resolution mesh was tested through hydraulic simulations and validated against real-world measurements.

Results demonstrate that the modified low-resolution mesh produces water level predictions up to 50% closer to observed measurements compared to the original low-resolution mesh. This significant improvement underscores the potential of the algorithm to enhance prediction accuracy. The findings contribute to advancing hydraulic modeling by optimizing mesh configurations and hold broader implications for flood management and water resource planning. By improving the reliability of water level simulations, this research supports more informed and effective environmental management strategies.