Hybrid Seasonal Prediction of Meridional Temperature Gradient Associated with "Warm Arctic-Cold Eurasia"

The meridional gradient of surface air temperature associated with "Warm Arctic-Cold Eurasia" (GradT(AE)) is closely related to climate anomalies and weather extremes in the mid-low latitudes. However, the Climate Forecast System Version 2 (CFSv2) shows poor capability for GradT(AE) prediction. Based on the year-to-year increment approach, analysis using a hybrid seasonal prediction model for GradT(AE) in winter (HMAE) is conducted with observed September sea ice over the Barents-Kara Sea, October sea surface temperature over the North Atlantic, September soil moisture in southern North America, and CFSv2 forecasted winter sea ice over the Baffin Bay, Davis Strait, and Labrador Sea. HMAE demonstrates good capability for predicting GradT(AE) with a significant correlation coefficient of 0.84, and the percentage of the same sign is 88% in cross-validation during 1983-2015. HMAE also maintains high accuracy and robustness during independent predictions of 2016-20. Meanwhile, HMAE can predict the GradT(AE) in 2021 well as an experiment of routine operation. Moreover, well-predicted GradT(AE) is useful in the prediction of the large-scale pattern of "Warm Arctic-Cold Eurasia" and has potential to enhance the skill of surface air temperature occurrences in the east of China.