AI-Informed Model Analogs for Subseasonal-to-Seasonal Prediction

2025-06-16

AI Summary

To address the limited accuracy of subseasonal-to-seasonal (S2S) forecasts, this study proposes an interpretable AI-driven dynamic analog forecasting method: a neural network learns physically interpretable weighted masks to optimize the selection of historical analogs, thereby unifying dynamic adaptability and physical interpretability in analog identification. This is the first systematic application of such an approach at the S2S timescale, supporting classification, regression, and probabilistic prediction of temperature and atmospheric circulation. Evaluated on reanalysis and model output data, the method significantly outperforms conventional analog, climatological, and persistence baselines: temperature classification skill for Weeks 3–4 improves by 12–18%; regional temperature regression correlation coefficients for Month 1 increase by 0.07–0.11; and 200-hPa wind field forecast skill is enhanced. Moreover, the method substantially improves extreme temperature event detection and uncertainty quantification in probabilistic forecasts.

Technology Category

Cognitive Modeling & Cognitive Systems: AnalogyMachine Learning: Calibration & Uncertainty QuantificationSearch and Optimization: Metareasoning and Metaheuristics

Application Category

Search and Retrieval-Augmented AI: Web query analysis, representation and understandingGraph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphsSemantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Subseasonal-to-seasonal forecasting is crucial for public health, disaster preparedness, and agriculture, and yet it remains a particularly challenging timescale to predict. We explore the use of an interpretable AI-informed model analog forecasting approach, previously employed on longer timescales, to improve S2S predictions. Using an artificial neural network, we learn a mask of weights to optimize analog selection and showcase its versatility across three varied prediction tasks: 1) classification of Week 3-4 Southern California summer temperatures; 2) regional regression of Month 1 midwestern U.S. summer temperatures; and 3) classification of Month 1-2 North Atlantic wintertime upper atmospheric winds. The AI-informed analogs outperform traditional analog forecasting approaches, as well as climatology and persistence baselines, for deterministic and probabilistic skill metrics on both climate model and reanalysis data. We find the analog ensembles built using the AI-informed approach also produce better predictions of temperature extremes and improve representation of forecast uncertainty. Finally, by using an interpretable-AI framework, we analyze the learned masks of weights to better understand S2S sources of predictability.

Problem

• Improving subseasonal-to-seasonal forecasting accuracy using AI-informed models
• Enhancing prediction of temperature extremes and atmospheric winds
• Understanding predictability sources via interpretable AI weight analysis

Innovation

• AI-informed model analog forecasting approach
• Neural network optimizes analog selection
• Interpretable-AI framework analyzes predictability sources