• We developed a new hybrid prediction model this year, which now serves as one of the reference models in KMA’s operational S2S prediction system since Nov 2025. 

• This page provides the two-month winter temperature forecasts from the model and compares them with the KMA’s operational GloSea6 system. 

So, the hybrid model approach, correcting teleconnection patterns in the model with observations. 

The hybrid model significantly improves the prediction skill of East Asian surface climate in both hindcasts and real-time forecasts (Park and Son, 2025). The hybrid v1.1 model has been integrated into the KMA operational forecasting system (Since Nov. 2025).

2. two-month real-time forecast (initialized in October)

For this December, the GloSea6 dynamical prediction system expects temperatures to remain near normal, with only a slight (and not statistically significant) tendency toward warmer-than-normal conditions.
Similarly, the hybrid prediction model expects temperatures to remain near normal, with only a slight (and not statistically significant) tendency toward cooler-than-normal conditions (without linear trend).

Refs. 

1. Improving subseasonal-to-seasonal prediction of ENSO–East Asia teleconnection: a dynamical-statistical model approach (2025) Park, C.-H. and S.-W. Son*.Quarterly Journal of the Royal Meteorological Society, e70031, https://doi.org/10.1002/qj.70031.

2. Subseasonal Variability of ENSO–East Asia Teleconnections Driven by Tropical Convection over the Indian Ocean and Maritime Continent  (2024) Park, C.-H. and S.-W. Son*. Geophysical Research Letters, 51, e2023GL108062, https://doi.org/10.1029/2023GL108062.

3. Sub-seasonal variability of ENSO teleconnections in western North America and its prediction skill (2023) Park, C.-H., J. Choi*, S.-W. Son*, D. Kim, S.-W. Yeh, and J.-S. Kug. Journal of Geophysical Research: Atmospheres, 128, e2022JD037985 https://doi.org/10.1029/2022JD037985