Optimizing Seasonal‐To‐Decadal Analog Forecasts With a Learned Spatially‐Weighted Mask

• Published in: Geophysical Research Letters
• Main Authors: Jamin K. Rader, Elizabeth A. Barnes
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 2023
• Subjects: interpretable machine learning; analog forecasting; seasonal‐to‐decadal climate prediction; El Niño Southern Oscillation; North Atlantic; sources of predictability; Geophysics. Cosmic physics; QC801-809
• Online Access: https://doi.org/10.1029/2023GL104983; https://doaj.org/article/dea19bb40b7247f685e88cf4ccabb771

Abstract Seasonal‐to‐decadal climate prediction is crucial for decision‐making in a number of industries, but forecasts on these timescales have limited skill. Here, we develop a data‐driven method for selecting optimal analogs for seasonal‐to‐decadal analog forecasting. Using an interpretable neural network, we learn a spatially‐weighted mask that quantifies how important each grid point is for determining whether two climate states will evolve similarly. We show that analogs selected using this weighted mask provide more skillful forecasts than analogs that are selected using traditional spatially‐uniform methods. This method is tested on two prediction problems using the Max Planck Institute for Meteorology Grand Ensemble: multi‐year prediction of North Atlantic sea surface temperatures, and seasonal prediction of El Niño Southern Oscillation. This work demonstrates a methodical approach to selecting analogs that may be useful for improving seasonal‐to‐decadal forecasts and understanding their sources of skill.