S2S-JEPA: Predicting the Predictable at Subseasonal-to-Seasonal Timescales
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.
Related
- Predicting birth weights in West Africa using subseasonal to seasonal climate and socioeconomic variables shows potential for an early warning system
- Marine heat waves in vulnerable coral reef regions are skillfully predictable on multi-year timescales
- Distinct intraseasonal pathways driving rapidly escalating heat extremes in North China and their subseasonal predictability
- Stress-Testing Dynamical and Generative Downscaling Using Subseasonal Extreme Precipitation Forecasts
- Skillful Data-Driven Subseasonal Soil Moisture Forecasting: Prospects and Limits for Flash Drought Prediction
BibTeX
@misc{dong2026s2sjepa,
title = {{S2S-JEPA: Predicting the Predictable at Subseasonal-to-Seasonal Timescales}},
author = {Chenyu Dong and Gianmarco Mengaldo},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.03106},
eprint = {2610.03106},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.03106},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - S2S-JEPA: Predicting the Predictable at Subseasonal-to-Seasonal Timescales
AU - Chenyu Dong
AU - Gianmarco Mengaldo
PY - 2026
DA - 2026/10/02
PB - arXiv
DO - 10.48550/arxiv.2610.03106
UR - https://arxiv.org/abs/2610.03106
AB - 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.
ER -