资源环境科技发展态势分析平台

Skillfully predicting persistent extreme temperature anomalies more than 10days in advance remains a challenge although it is of great value to the society. Here the two leading modes of subseasonal variability of surface air temperature over the extratropical Northern Hemisphere in boreal winter are identified with pentad (5days) averaged data. They are well separated geographically, dominating temperature variability in North America and Eurasia, respectively. There exists a two-pentad lagged correlation between these two modes, implying an intercontinental link of temperature variability. Forecast skill of these two modes is evaluated based on three operational subseasonal prediction models. The results show that useful forecasts of the Eurasian mode (EOF2) can be achieved four pentads in advance, which is more skillful than the North American mode (EOF1). EOF2 is found to benefit from the Madden-Julian Oscillation signal in the initial condition.

Plain Language Summary In this study, the two leading modes of temperature variability in wintertime Northern Hemisphere are identified. They are well separated to represent dominant subseasonal variability in North America and Eurasia, respectively. They account for more than 60% of the local variance. Also, it is found that there exists an intercontinental connection of temperature variability as these two modes are correlated with a time lag of about 10days. Sources of variability for these two modes are found to be located in the high-latitude regions. Three state-of-the-art operational subseasonal-to-seasonal prediction systems are evaluated for their performance on predicting these two leading temperature variability modes. The Eurasian mode, which is closely associated with cold surges in East Asia, is found to be more predictable than the North American mode. Contributions of several important sources of predictability are explored such as air-sea coupling and the Madden-Julian Oscillation. The results are of great value for future development of subseasonal predictions.