Enhanced Seasonal Prediction Skill of the Western Pacific Subtropical High Over the Past Century

Abstract The Western Pacific Subtropical High (WPSH) is crucial in shaping summer climate of East Asia. However, existing studies on WPSH seasonal prediction have primarily focused on the recent 20–30 years, limiting our ability to assess decadal variations in prediction skill. Here, based on CSF‐20C seasonal hindcast (1901–2010), we find the prediction skill of WPSH improved post‐1960s due to increased signal‐to‐noise ratio in model, which is driven by strengthened predictable signals rather than noise reduction. Through comparative analysis of high‐skill (1961–2010) and low‐skill (1901–1950) periods, we identify three persistent oceanic predictors of WPSH variability: sea surface temperature (SST) anomalies in tropical central‐eastern Pacific, tropical North Atlantic, and Indo‐Pacific. Notably, the Indo‐Pacific region's relative contribution to prediction skill increased substantially from 40% to 59%, establishing it as the dominant driver of enhanced predictability. The expansion of observational data and intensification of quasi‐biennial ENSO variability since the 1960s may be two primary factors contributing to the enhanced predictable components of WPSH. Beyond predictable signals, we also examine the unpredictable noises in both observations and CSF‐20C. For observations, unpredictable component of SST across three key oceanic regions constitutes the primary source of WPSH‐related unpredictable noises during low‐skill period. In contrast, during high‐skill period, the main contributors to unpredictable noises shift to the SST‐independent atmospheric teleconnections in mid‐high latitudes (e.g., the East‐Atlantic pattern and summer North Atlantic Oscillation), which influence WPSH by modulating the East Asian westerly jet. Our findings provide novel insights into the mechanisms governing seasonal prediction of WPSH.