Deep Learning Meets Teleconnections: Improving S2S Predictions for European Winter Weather

2025-04-10

AI Summary

This study addresses the challenge of improving subseasonal-to-seasonal (S2S) predictability of European winter weather at lead times of 2–8 weeks, focusing on the synergistic modulation by stratospheric polar vortex (SPV) and Madden–Julian oscillation (MJO) teleconnections. We propose a novel ViT-LSTM hybrid deep learning framework that, for the first time, directly encodes physically interpretable stratospheric zonal wind and tropical outgoing longwave radiation (OLR) fields via Vision Transformers to enable dynamical attribution of remote teleconnection processes. The model integrates reanalysis data, ECMWF hindcasts, and SPV/MJO indices. Experiments demonstrate that ViT-LSTM surpasses the operational ECMWF S2S system in forecast skill during weeks 5–6; specifically, prediction skill for Scandinavian blocking and the Atlantic ridge improves by over 15%. High-confidence forecasts are robustly associated with SPV anomalies and MJO phase, thereby empirically validating and extending classical teleconnection theory.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Large Vision Models
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Predictions on subseasonal-to-seasonal (S2S) timescales--ranging from two weeks to two month--are crucial for early warning systems but remain challenging owing to chaos in the climate system. Teleconnections, such as the stratospheric polar vortex (SPV) and Madden-Julian Oscillation (MJO), offer windows of enhanced predictability, however, their complex interactions remain underutilized in operational forecasting. Here, we developed and evaluated deep learning architectures to predict North Atlantic-European (NAE) weather regimes, systematically assessing the role of remote drivers in improving S2S forecast skill of deep learning models. We implemented (1) a Long Short-term Memory (LSTM) network predicting the NAE regimes of the next six weeks based on previous regimes, (2) an Index-LSTM incorporating SPV and MJO indices, and (3) a ViT-LSTM using a Vision Transformer to directly encode stratospheric wind and tropical outgoing longwave radiation fields. These models are compared with operational hindcasts as well as other AI models. Our results show that leveraging teleconnection information enhances skill at longer lead times. Notably, the ViT-LSTM outperforms ECMWF's subseasonal hindcasts beyond week 4 by improving Scandinavian Blocking (SB) and Atlantic Ridge (AR) predictions. Analysis of high-confidence predictions reveals that NAO-, SB, and AR opportunity forecasts can be associated with SPV variability and MJO phase patterns aligning with established pathways, also indicating new patterns. Overall, our work demonstrates that encoding physically meaningful climate fields can enhance S2S prediction skill, advancing AI-driven subseasonal forecast. Moreover, the experiments highlight the potential of deep learning methods as investigative tools, providing new insights into atmospheric dynamics and predictability.

Problem

• Improving subseasonal-to-seasonal (S2S) weather predictions for Europe
• Utilizing teleconnections like SPV and MJO to enhance forecast accuracy
• Developing deep learning models to analyze atmospheric dynamics and predictability

Innovation

• LSTM network predicts NAE regimes
• Index-LSTM incorporates SPV and MJO
• ViT-LSTM encodes stratospheric wind fields