Abstract: Understanding tree water use strategies under altered precipitation is vital for forest ecosystem management. However, most studies have focused on tree water use strategies at seasonal or monthly scales, while neglecting those following individual rainfall events. Here, we employed stable isotope (δD and δ¹⁸O) and Bayesian mixed model (MixSIAR) to quantify water use patterns of dominant trees in Quercus aliena, Pinus armandii pure forests, and Q. aliena-P. armandii mixed forests across light, moderate, and heavy rainfall events (7.9, 23.8, and 47.3 mm) in the northern subtropical-warm temperate transition zone of China. Furthermore, correlation analysis and structural equation modeling (SEM) were used to identify key regulatory factors. Our results showed that tree mixing exerted species-specific and precipitation-dependent effects on water use. Under light and moderate rain, tree mixing significantly increased the absorption proportion from 0-20 cm soil water (18.53-42.93% vs. 48.97-86.53%) by Q. aliena while reducing its reliance on subsoil water. Tree mixing only enhanced the absorption proportion from 0-20 cm soil layer by P. armandii under light rain (16.07% vs. 25.47%). SEM revealed that soil nutrients (total nitrogen and total phosphorus) served as the core drivers regulating tree water use patterns across pure and mixed forests. This study highlights the vital nutrient regulation of tree water use strategies, challenging the traditional view that tree water use controlled by soil water distribution. It provides theoretical support for stand structure optimization and adaptive forest management in climate transition zones.

Key words: Water use strategy, δD and δ¹⁸O, Natural secondary forest, Precipitation events, Tree mixing