[A-CG43] Multi-scale ocean-atmosphere interaction in the tropics

[ACG43-P06] Quantifying the teleconnections between East Asian winter precipitation and tropical sea surface temperatures using causal analysis

*TIANJIAO MA^1,2, Wen Chen¹, Xiadong An¹, Yu Kosaka² (1.Yunnan University, Kunming, China, 2.RCAST, UTokyo)

Keywords:East Asian winter precipitation, ENSO teleconnection, Causal inference, Peter-Clark Momentary Conditional Independent (PCMCI) algorithm

It was previously revealed that winter precipitation in East Asia is linked to tropical Sea Surface Temperature (SST) anomalies in the Indian and Pacific Oceans. While several teleconnection pathways have been identified, accurately measuring these connections and determining the relative contributions of SSTs remains a challenge. We employed a novel causal inference approach to quantify the effects of SSTs on East Asian winter precipitation, focusing on its leading mode (denoted as Pr). Using the PCMCI method, we first derived a causal graph between the Pr and SSTs in the tropical eastern/western Pacific and Indian Ocean (denoted as EP, WP, and IO respectively). Based on the causal graph, we then applied a quantitative analysis using Wright’s path method. The results indicated that WP-SST had the most substantial impact on the Pr, with a 1 Standard Deviation (SD) increase leading to a 0.24 SD decrease in Pr at a 1-month lag and 0.17 SD decrease at a 2-month lag. The winter monsoon winds and subtropical jet stream are identified as critical mediators of this effect. Additionally, EP-SST had a notable positive influence on the Pr, with a 1 SD warming leading to a 0.12 SD increase in Pr at a 2-month lag, while IO-SST exhibits a minor effect. It is essential to note that, traditional regression analysis indicates that EP-SST explained the largest portion of the Pr variance among the three SSTs, but the causal analysis revealed that WP-SST exerts a more significant role in driving Pr variations, aligning with numerical experiments.