Pacific Decadal Oscillation (PDO), one of the main internal decadal climate variabilities, has substantial impacts on climate and environment.
PDO is a combination of multiple processes of different origins, such as stochastic forcing generated by local atmospheric noise, tropical-subtropical interactions via atmospheric and oceanic responses and ocean gyre dynamics. Consequently, it is challenging to model and predict the PDO, especially the PDO phase transition.
Recently, a group led by Prof. MA Zhuguo from the Institute of Atmospheric Physics (IAP) of the Chinese Academy of Sciences, cooperating with researchers from Sun Yat-sen University and Humboldt University of Berlin, proposed a climate network to detect the early warning signal of the PDO phase transition.
The study was published in Geophysical Research Letters on Feb. 11.
The objectively detected early warning signal forewarned all the six PDO phase transitions from the 1890s to 2000s, with an average of 6.5 years in advance, and only one false alarm sounded in the 1950s. These results indicated that the predictive skill is improved and the prediction duration is increased, close to the estimated upper limit of the predictability time of 8-9 years.
This research further discovered that the cooperative behavior of climate network is the reason why the early warning of the PDO phase transition is detected from sea surface temperature successfully. Cooperative behavior is a phenomenon that more and more elements in a dynamical system display similar changes in temporal.
"Due to the effect of cooperative behavior, the slight difference which is hard to be perceived individually can be easily detected macroscopically, which is helpful to reduce the influence of background noise," said Dr. LU Zhenghui from IAP, the lead author of the study.
"As a method driven by big data, climate network has shown a strong potential on climate research. Our research provides a new way for the prediction of the PDO phase transition, which may be applied to other major climate events," said Dr. YUAN Naiming, the corresponding author of the study.
This work was supported by National Key R&D Program of China and National Natural Science Foundation of China.
52 Sanlihe Rd., Xicheng District,
Beijing, China (100864)