Global River Forecasting with a Topology-Informed AI Foundation Model

2026-02-25

AI Summary

This study addresses the challenge of basin-wide hydrological modeling in data-scarce river systems by proposing GraphRiverCast (GRC), a novel AI foundation model that explicitly encodes river network topology into its architecture. GRC introduces a topology-aware, physics-aligned neural operator capable of ColdStart inference—eliminating the need for historical state initialization and thereby circumventing error accumulation inherent in traditional autoregressive approaches. Leveraging a pretraining-finetuning paradigm, GRC enables end-to-end, cross-scale simulation of multivariate hydrodynamic processes. Experimental results demonstrate that GRC-ColdStart achieves a global 7-day pseudo-forecast Nash–Sutcliffe Efficiency (NSE) of 0.82, significantly outperforming both physical models and local AI baselines across both gauged and ungauged river segments.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Reasoning under Uncertainty: Graphical Models
• Humans and AI: Other Foundations of Human Computation & AI

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

River systems operate as inherently interconnected continuous networks, meaning river hydrodynamic simulation ought to be a systemic process. However, widespread hydrology data scarcity often restricts data-driven forecasting to isolated predictions. To achieve systemic simulation and reduce reliance on river observations, we present GraphRiverCast (GRC), a topology-informed AI foundation model designed to simulate multivariate river hydrodynamics in global river systems. GRC is capable of operating in a "ColdStart" mode, generating predictions without relying on historical river states for initialization. In 7-day global pseudo-hindcasts, GRC-ColdStart functions as a robust standalone simulator, achieving a Nash-Sutcliffe Efficiency (NSE) of approximately 0.82 without exhibiting the significant error accumulation typical of autoregressive paradigms. Ablation studies reveal that topological encoding serves as indispensable structural information in the absence of historical states, explicitly guiding hydraulic connectivity and network-scale mass redistribution to reconstruct flow dynamics. Furthermore, when adapted locally via a pre-training and fine-tuning strategy, GRC consistently outperforms physics-based and locally-trained AI baselines. Crucially, this superiority extends from gauged reaches to full river networks, underscoring the necessity of topology encoding and physics-based pre-training. Built on a physics-aligned neural operator architecture, GRC enables rapid and cross-scale adaptive simulation, establishing a collaborative paradigm bridging global hydrodynamic knowledge with local hydrological reality.

Problem

• river forecasting
• data scarcity
• systemic simulation
• hydrodynamic modeling
• topology encoding

Innovation

• topology-informed AI
• foundation model
• ColdStart forecasting
• neural operator
• global river simulation