Operational flood forecasting still relies on high-fidelity two-dimensional hydraulic solvers, but their runtime can be prohibitive for rapid decision support on large urban floodplains. In parallel, AI-based surrogate models have shown strong potential in several areas of computational physics for accelerating otherwise expensive high-fidelity simulations. We address this issue on the lower Têt River (France), starting from a production-grade Telemac2D model defined on a high-resolution unstructured finite-element mesh with more than 4×105 nodes. From this setup, we build a learning-ready database of synthetic but operationally grounded flood events covering several representative hydrograph families and peak discharges. On top of this database, we develop a graph-neural surrogate based on projected meshes and multimesh connectivity. The projected-mesh strategy keeps training tractable while preserving high-fidelity supervision from the original Telemac simulations, and the multimesh construction enlarges the effective spatial receptive field without increasing network depth. We further study the effect of an explicit discharge feature Q(t) and of pushforward training for long autoregressive rollouts. The experiments show that conditioning on Q(t) is essential in this boundary-driven setting, that multimesh connectivity brings additional gains once the model is properly conditioned, and that pushforward further improves rollout stability. Among the tested configurations, the combination of Q(t), multimesh connectivity, and pushforward provides the best overall results. These gains are observed both on hydraulic variables over the surrogate mesh and on inundation maps interpolated onto a common 25m regular grid and compared against the original high-resolution Telemac solution. On the studied case, the learned surrogate produces 6-hour predictions in about 0.4s on a single NVIDIA A100 GPU, compared with about 180min on 56 CPU cores for the reference simulation. These results support graph-based surrogates as practical complements to industrial hydraulic solvers for operational flood mapping.
Figures & tables
Figure 1: Overview of the dataset design: (a) spatial mesh-density strategy and (b) synthetic hydrograph families used to populate the database.
Graph
Nodes
Directed edges
Mean edge (m)
r10 median (m)
r10 mean (m)
r10 max (m)
Original mesh
412,844
2,464,330
15.7
67
156
733
×8 mesh
16,222
96,984
81.0
479
809
5,468
×8 + multimesh
16,222
130,597
110.4
596
994
11,571
Table 1: Indicative 10-hop physical scale for different graph constructions. We report summary statistics across nodes of r10(i)=10ℓi , where ℓi is the mean outgoing edge length at node i (all distances in metres).
Name
Connectivity
Q(t)
PF
E1
standard
no
no
E2
standard
yes
no
E3
standard
yes
yes
E4
multimesh
no
no
E5
multimesh
yes
no
E6
multimesh
yes
yes
Table 2: Main experimental configurations. PF denotes pushforward training.
AI-driven flood digital twins demand fast hydrodynamic surrogates for ensemble forecasting and observation assimilation. Yet even GPU-accelerated two-dimensional shallow water equation (SWE) solvers still require ∼55 minutes per 96-hour run on a ∼4.2-million-active-cell metropolitan basin (the DesPlaines River basin at 30m resolution), making such workloads prohibitive at native resolution. We present the Conditional Latent Dynamics Network (CLDNet): a low-dimensional latent neural ODE driven by rainfall, paired with a coordinate-based decoder conditioned on static terrain (elevation, slope, Manning roughness) that reconstructs depth and discharge at arbitrary query points. Pointwise decoding decouples memory from grid size and handles irregular watersheds natively, enabling metropolitan-scale training on a single compute node and direct queries at exact gauge coordinates without raster snapping. We evaluate CLDNet on a synthetic 250,000-cell Texas benchmark and on a new DesPlaines case study of 114 real-rainfall StageIV storms whose reference simulator we validate against United States Geological Survey (USGS) gauges at the April2013 flood-of-record (Nash--Sutcliffe efficiency 0.57--0.94 on mean-recentered water-surface elevation). CLDNet roughly halves the relative root-mean-squared error of an unconditional baseline, outperforms regular-grid VAE--ConvLSTM and FNO baselines on the Texas benchmark (both presuppose a Cartesian grid and do not apply to the irregular Des~Plaines watershed), reaches a critical success index of ≈86% at the 0.5m inundation threshold, and produces a full 96-hour basin-wide forecast in ∼29 seconds -- a ∼115× speedup.
Phillip Si, Yuan Qiu, Omar Sallam +4
School of Computational Science and Engineering, Georgia Institute of Technology, 756 West Peachtree Street Northwest, Atlanta, 30308, GA, USA · Environmental Science Division, Argonne National Laboratory, 9700 S Cass Ave, Lemont, 60439, IL, USA
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.
Christian Lammers, Fernando Arévalo, Leonie Märker-Neuhaus +3
Institute for Computing and Information Sciences Radboud University Nijmegen, Netherlands · Wupper Association (Wupperverband) Wuppertal, Germany
Near-real-time flood depth prediction demands surrogate models that are accurate, fast, and transferable across watersheds. Supervised surrogates can match physics-based simulators in accuracy but need millions of training rows per watershed and cannot extrapolate beyond their original mesh. We propose a domain-aware coreset construction pipeline that conditions a tabular foundation model at inference time. The pipeline stratifies storms by return period and most-affected watershed, then samples hexagons with a target-aware spatial selector. With 0.7% of the per-watershed training pool, the model attains a mean R2 of 0.663 across nine Houston-area watersheds, within 98.5% of the supervised reference (R2 = 0.673). It transfers to held-out watersheds without task-specific retraining, staying ahead of a coreset-trained supervised baseline. On real storms it exceeds the supervised reference on a far out-of-distribution case and trails it on a mostly in-distribution one. Domain-aware coreset construction lets tabular foundation models deliver data-efficient, watershed-transferable flood predictions without per-watershed training.
Lipai Huang, Adithi Srinath, Manas Singh +2
Urban Resilience.AI Lab, Zachry Department of Civil and Environmental Engineering, Texas A&M University, College Station, TX, USA · Department of Computer Science and Engineering, Texas A&M University, College Station, TX, USA · Resilitix Intelligence LLC, Houston, TX, USA +1