Multi-Source Dynamic Graph Learning for Compound-Flood Forecasting in Managed Coastal Systems

2026-08-03

AI Summary

This study addresses the limitations of existing flood forecasting models, which struggle to accurately represent the prolonged high-water plateau phases critical for early warnings and fail to capture the systemic dynamics of compound flooding using single-station data. To overcome these challenges, this work proposes an anchored multi-source dynamic graph neural network framework that integrates hydrological, meteorological, and reservoir operation data from multiple stations. By employing a state- and lead-time-dependent bounded residual correction mechanism, the model adaptively captures inter-station relationships and enables temporally aligned evaluation. The approach maintains high predictive accuracy under typical hydrological conditions while significantly improving forecast reliability during high-water plateau periods, thereby offering robust support for early warning of coastal compound floods and water resource management.

Abstract

Compound flooding in managed coastal systems is influenced by hydrological conditions and water-management activity observed across multiple monitoring stations. Current forecasting models can capture temporal dependencies with low average errors, but global error metrics may conceal poor reproduction of prolonged high-water plateaus that are relevant to flood early warning. Because hydrometeorological and operational signals are distributed across heterogeneous gages, single-site records do not fully represent high-water dynamics. Nevertheless, unconstrained fusion of cross-site signals can degrade the stability of local temporal forecasts. This work proposes an anchored forecasting framework that incorporates cross-site information through state- and lead-dependent bounded residual corrections. A multi-source regime representation constructed from hydrometeorological and operational observations adaptively calibrates inter-site relationships and correction scales, enabling targeted cross-site adjustment while preserving the local temporal forecast as a stable anchor. Beyond conventional global error statistics, we evaluate event-scale high-water characteristics through the temporal alignment of forecasted and observed high-water processes. Experiments demonstrate that selectively integrating multi-station dynamic conditions improves the prediction reliability of sustained high-water plateaus while maintaining high accuracy during routine hydrological conditions, supporting flood early warning and water-management decision support.

Problem

• compound flooding
• multi-source data
• high-water plateaus
• forecast stability
• coastal water management

Innovation

• multi-source dynamic graph learning
• anchored forecasting
• bounded residual correction
• compound-flood forecasting
• high-water plateau prediction