Graph Transformer-Based Flood Susceptibility Mapping: Application to the French Riviera and Railway Infrastructure Under Climate Change

2025-03-31

AI Summary

Facing escalating flood frequency and intensity due to climate change—posing severe risks to critical infrastructure—conventional machine learning models fail to adequately capture spatial dependencies and sensitivity boundaries. This study proposes, for the first time, a graph transformer (GT)-based flood susceptibility mapping framework that integrates watershed topological structure with multi-source geoenvironmental data, applied to the French Riviera region and its railway network. We innovatively introduce Laplacian positional encoding to explicitly embed spatial topology into the GT architecture and leverage self-attention mechanisms to enhance boundary delineation and cluster identification. The model achieves an AUC of 0.9739 and Moran’s I of 0.6119 (p < 0.0001), significantly outperforming baseline methods. Under the RCP 8.5 scenario, projections indicate that the area classified as “very high susceptibility” will increase to 17.46% by 2050, while exposed railway length will rise to 54%.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Computer Vision: Remote Sensing / Geospatial AI
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Increasing flood frequency and severity due to climate change threatens infrastructure and demands improved susceptibility mapping techniques. While traditional machine learning (ML) approaches are widely used, they struggle to capture spatial dependencies and poor boundary delineation between susceptibility classes. This study introduces the first application of a graph transformer (GT) architecture for flood susceptibility mapping to the flood-prone French Riviera (e.g., 2020 Storm Alex) using topography, hydrology, geography, and environmental data. GT incorporates watershed topology using Laplacian positional encoders (PEs) and attention mechanisms. The developed GT model has an AUC-ROC (0.9739), slightly lower than XGBoost (0.9853). However, the GT model demonstrated better clustering and delineation with a higher Moran's I value (0.6119) compared to the random forest (0.5775) and XGBoost (0.5311) with p-value lower than 0.0001. Feature importance revealed a striking consistency across models, with elevation, slope, distance to channel, and convergence index being the critical factors. Dimensionality reduction on Laplacian PEs revealed partial clusters, indicating they could capture spatial information; however, their importance was lower than flood factors. Since climate and land use changes aggravate flood risk, susceptibility maps are developed for the 2050 year under different Representative Concentration Pathways (RCPs) and railway track vulnerability is assessed. All RCP scenarios revealed increased area across susceptibility classes, except for the very low category. RCP 8.5 projections indicate that 17.46% of the watershed area and 54% of railway length fall within very-high susceptible zones, compared to 6.19% and 35.61%, respectively, under current conditions. The developed maps can be integrated into a multi-hazard framework.

Problem

• Improving flood susceptibility mapping using graph transformers
• Addressing spatial dependency challenges in traditional ML methods
• Assessing future flood risks under climate change scenarios

Innovation

• Graph transformer architecture for flood mapping
• Laplacian positional encoders capture spatial dependencies
• Climate scenarios assess future flood susceptibility