Problem
Research questions and friction points this paper is trying to address.
subseasonal-to-seasonal forecasting
predictability desert
AI weather models
long-range weather prediction
S2S-JEPA: Predicting the Predictable at Subseasonal-to-Seasonal Timescales
🤖 AI Summary
This study addresses the accuracy degradation in subseasonal-to-seasonal (S2S) meteorological forecasting caused by model overfitting to unpredictable noise. To this end, we introduce the Joint Embedding Predictive Architecture (JEPA) into S2S forecasting for the first time. Methodologically, latent space representation learning is employed to capture slowly evolving predictable components while discarding high-frequency noise details. By integrating advanced AI weather modeling components, our approach transcends the limitations of conventional end-to-end prediction, enabling physically consistent long-range forecasts. Experimental results demonstrate that the proposed method achieves overall forecast skill comparable to the ECMWF standard ensemble prediction system. Notably, it surpasses the baseline across multiple metrics during weeks five and six, significantly enhancing medium-range forecasting capabilities.
📝 Abstract
The subseasonal-to-seasonal (S2S) timescale, roughly from two weeks to two months ahead, is a critical forecast window for sectors such as agriculture, energy, and water management. Yet, it is widely known as the `predictability desert'. Recent AI weather models excel up to two weeks ahead but deteriorate beyond, largely because they are trained to predict fine-scale details that are neither predictable nor essential at S2S timescales. We argue that a more physically grounded objective is to forecast only the slowly varying components that remain predictable. Computer vision reached the same conclusion with the Joint-Embedding Predictive Architecture (JEPA), which predicts in latent space, discarding unpredictable details. In this work, we introduce S2S-JEPA, which brings the JEPA paradigm to S2S forecasting. It is tailored to this task through design elements from state-of-the-art AI weather models. S2S-JEPA achieves comparable skill to the gold-standard ECMWF physics-based ensemble and surpasses it on multiple metrics at weeks 5 to 6.
Innovation
Methods, ideas, or system contributions that make the work stand out.
JEPA
Subseasonal-to-Seasonal Forecasting
Latent Space Prediction
Slowly Varying Components
AI Weather Models
🔎 Similar Papers
No similar papers found.
