Improved seasonal predictions of the Western Pacific subtropical high by incorporating tropical Atlantic influences

Abstract The Western Pacific Subtropical High (WPSH) significantly influences East Asian weather and air pollution. Previous studies have indicated that tropical sea surface temperature anomalies (SSTAs) are important indicators of WPSH intensity. While current seasonal predictions rely on spring predictors from the Indo-Pacific and North Atlantic, the growing influence of the tropical Atlantic (TA) has been overlooked. This study introduces springtime TA SSTA as a new predictor into a Long Short-Term Memory (LSTM) model, alongside established predictors, for seasonal prediction of WPSH. Across the WPSH prediction models using 8-predictor combinations, incorporating TA SSTA consistently improved WPSH prediction accuracy. The optimal prediction model, which combines TA SSTA with three established predictors, exhibits stable and high predictive skill and achieves an 18.0% increase in explained variance (R 2 ). Our results highlight the critical role of springtime TA SSTA for predicting summertime WPSH intensity, establishing a promising way for extreme weather and pollution event predictions.