Elucidating the water utilization strategy of trees during forest succession is a prerequisite for predicting the direction of forest succession. However, the water utilization characteristics of trees in forests across a successional gradient remain unclear. Here, we utilized the hydrogen and oxygen stable isotopes combined with the Bayesian mixed model (MixSIAR) to analyze the water utilization of dominant trees (Pinus massoniana, Castanea henryi, and Schima superba) in forests along a successional gradient in the Dinghushan Biosphere Reserve of China. Furthermore, we determined the primary factor affecting the water utilization of various trees based on variation partitioning analysis and a random forest model. Our results illustrated that in the early-successional forest, the water utilization ratios from shallow soil layers by P. massoniana were significantly lower than that in the mid-successional forest (51.3%–61.7% vs. 75.3%–81.4%), while its water utilization ratios from deep soil layers exhibited the opposite pattern (26.1%–30.1% vs. 9.0%–15.0%). Similarly, the ratios of water utilization from shallow soil layers by C. henryi (18.9%–29.5% vs. 32.4%–45.9%) and S. superba (10.0%–25.7% vs. 29.2%–66.4%) in the mid-successional forest were relatively lower than in the late-successional forest, whereas their water utilization ratios from deep soil layers showed the contrary tendency. Moreover, our results demonstrated that the diverse water utilization of each tree in different successional forests was mainly attributed to their distinct plant properties. Our findings highlight the increased percentage of water utilization of trees from shallow soil layers with forest succession, providing new insights for predicting the direction of forest succession under changing environments.
Read full abstract