Long-term Trend Correction of Seasonal Forecast of East Asian Air Temperature in APCC MME

APCC MME 동아시아 기온 계절예측 온난화 보정

Recent global warming is apparent in observations and seasonal forecasts as well. In the East Asian region, however, the observed warming signal is not as strong as that of the globe in the recent decade, although climate forecast models tend to predict above-normal temperatures, similar to those of the global mean temperature. The marked discrepancies between seasonal forecasts and observations of 2-m temperature (T2m) in East Asia during 2013~2022 are corrected using a linear trend scaling method in this study. Trends of individual models are scaled with consideration of the observations, and T2m forecasts are corrected. Then, multi-model ensemble (MME) forecasts are generated from the temperature outputs of the corrected individual models. As a result, monthly long-term averaged corrected air temperature forecasts better reflect real-world conditions. Furthermore, monthly long-term averaged verification skill scores show significant improvement and greater stability compared to the original forecasts. The temperature correction method presented in this study accounts for the real-time operation of seasonal forecasts by carefully selecting the training and testing periods; therefore, it is applicable to the production of real-time seasonal forecasts.