FlowCast-ODE: Continuous Hourly Weather Forecasting with Dynamic Flow Matching and ODE Integration

2025-09-18

AI Summary

To address rapid error accumulation in autoregressive hourly weather forecasting and temporal discontinuities arising from ERA5’s 12-hour data assimilation cycle, this paper proposes a continuous-time modeling framework integrating dynamic flow matching with ordinary differential equations (ODEs). Methodologically, we design a conditional flow path formulation coupled with a low-rank AdaLN-Zero modulation mechanism, trained via a coarse-to-fine strategy that reduces model parameters by 15% without sacrificing accuracy. Experiments demonstrate significant improvements over strong baselines in RMSE, energy conservation, and fine-grained feature fidelity. The approach effectively mitigates assimilation-induced discontinuities, enhancing short-term forecast stability and temporal coherence. Moreover, it achieves state-of-the-art performance in predicting extreme events—particularly tropical cyclones—surpassing existing methods in both trajectory accuracy and intensity evolution.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Computer Vision: Diffusion Models for Vision
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• 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

Abstract

Accurate hourly weather forecasting is critical for numerous applications. Recent deep learning models have demonstrated strong capability on 6-hour intervals, yet achieving accurate and stable hourly predictions remains a critical challenge. This is primarily due to the rapid accumulation of errors in autoregressive rollouts and temporal discontinuities within the ERA5 data's 12-hour assimilation cycle. To address these issues, we propose FlowCast-ODE, a framework that models atmospheric state evolution as a continuous flow. FlowCast-ODE learns the conditional flow path directly from the previous state, an approach that aligns more naturally with physical dynamic systems and enables efficient computation. A coarse-to-fine strategy is introduced to train the model on 6-hour data using dynamic flow matching and then refined on hourly data that incorporates an Ordinary Differential Equation (ODE) solver to achieve temporally coherent forecasts. In addition, a lightweight low-rank AdaLN-Zero modulation mechanism is proposed and reduces model size by 15% without compromising accuracy. Experiments demonstrate that FlowCast-ODE outperforms strong baselines, yielding lower root mean square error (RMSE) and better energy conservation, which reduces blurring and preserves more fine-scale spatial details. It also shows comparable performance to the state-of-the-art model in forecasting extreme events like typhoons. Furthermore, the model alleviates temporal discontinuities associated with assimilation cycle transitions.

Problem

• Achieving accurate continuous hourly weather forecasting
• Addressing error accumulation in autoregressive weather prediction models
• Overcoming temporal discontinuities in ERA5 data assimilation cycles

Innovation

• Continuous atmospheric flow modeling with ODE integration
• Coarse-to-fine training with dynamic flow matching
• Lightweight low-rank modulation reducing model size