Causal inference and mediation for summer precipitation over middle and lower reaches of the Yangtze River

Abstract

Abstract The accurate attribution of summer precipitation in the middle and lower reaches of the Yangtze River (MLYR) is essential for operational forecasting and disaster prevention. However, traditional linear correlation methods are insufficient for capturing reliable causal linkages, making causal discovery algorithms a more appropriate solution. Causal effect measures suggest that tropical climate anomalies exert strong driving and mediating influences during boreal summer, while the Asian climate anomalies exhibit greater sensitivity. Causal analysis identifies seven direct drivers of MLYR precipitation: pressure anomalies over northwest Pacific, Northeast Asia, mid-latitude eastern Pacific, Ural Mountains, southwest Pacific, Scandinavia and Greenland. Additionally, we uncovered the further causal pathways linking MLYR precipitation with tropical Pacific and Antarctic Oscillation signals. These results identify the detailed mediations through the direct drivers of MLYR precipitation, which are crucial to capture its remote precursors. Our findings reveal the physical attributions of MLYR precipitation from the global climate, which may improve its operational prediction skills, and even broaden the precursors of East Asian summer monsoon.