ARROW: An Adaptive Rollout and Routing Method for Global Weather Forecasting

2025-10-10

AI Summary

Existing data-driven weather forecasting methods struggle to capture the multiscale spatiotemporal dependencies inherent in the global atmospheric system, while also failing to simultaneously mitigate error accumulation during long-horizon autoregressive rollout and resolve fine-grained dynamic processes. To address these challenges, we propose an adaptive rollout framework with multiscale routing. Our approach introduces a multi-interval prediction architecture, integrates a reinforcement learning–driven dynamic scheduler for adaptive horizon selection, employs a shared-private mixture-of-experts (MoE) design to enhance parameter efficiency, and adopts ring positional encoding (Ring PE) to better model periodic spatiotemporal patterns. Evaluated across multiple global meteorological forecasting benchmarks, our method achieves state-of-the-art performance, significantly improving forecast accuracy and stability for 1–7 day predictions. Extensive experiments demonstrate its effectiveness and robustness in high-resolution, long-sequence weather modeling.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Machine Learning: Mixture of Experts (MoE)
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Weather forecasting is a fundamental task in spatiotemporal data analysis, with broad applications across a wide range of domains. Existing data-driven forecasting methods typically model atmospheric dynamics over a fixed short time interval (e.g., 6 hours) and rely on naive autoregression-based rollout for long-term forecasting (e.g., 138 hours). However, this paradigm suffers from two key limitations: (1) it often inadequately models the spatial and multi-scale temporal dependencies inherent in global weather systems, and (2) the rollout strategy struggles to balance error accumulation with the capture of fine-grained atmospheric variations. In this study, we propose ARROW, an Adaptive-Rollout Multi-scale temporal Routing method for Global Weather Forecasting. To contend with the first limitation, we construct a multi-interval forecasting model that forecasts weather across different time intervals. Within the model, the Shared-Private Mixture-of-Experts captures both shared patterns and specific characteristics of atmospheric dynamics across different time scales, while Ring Positional Encoding accurately encodes the circular latitude structure of the Earth when representing spatial information. For the second limitation, we develop an adaptive rollout scheduler based on reinforcement learning, which selects the most suitable time interval to forecast according to the current weather state. Experimental results demonstrate that ARROW achieves state-of-the-art performance in global weather forecasting, establishing a promising paradigm in this field.

Problem

• Modeling spatial and multi-scale temporal dependencies in global weather systems
• Balancing error accumulation with fine-grained atmospheric variation capture
• Improving long-term global weather forecasting through adaptive rollout strategies

Innovation

• Multi-interval forecasting model with shared-private experts
• Ring Positional Encoding for Earth's circular latitude structure
• Reinforcement learning-based adaptive rollout scheduler