Results
Under a common environment of the tree plantations, the seven pine species exhibited relatively large variations in trunk radial growth rate. Across a cambial age sequence of 15 years, the cumulative basal area of all the seven pine species exhibited linear relationships with age but the rate of increment showed relatively large interspecific differences (Fig. 2). The cumulative basal area at cambial age of 15 years (CBA15) varied more than two folds among the studied species, i.e. ranged from 83.9 cm2 in P. tabuliformis to 209.1 cm2 in P. sylverstriformis . Similarly, the standardized tree-ring width chronologies of the seven studied species showed that there were relatively large variations in climate sensitivity of the annual tree-ring growth with the mean sensitivity ranged from 0.21 in P. koraiensis to 0.33 in P. tabuliformis (Table 1). Moreover, our tree-ring analyses showed that radial growth of the seven species showed substantial interspecific differences in responding to the extreme drought event happened in 2015, i.e. with values of the drought resistance index ranged from 0.54 in P. densiflora to 1.33 inP. koraiensis and the drought resilience index ranged from 0.51 in P. densiflora to 1.30 in P. koraiensis . All the studied species showed relatively low recovering ability after the extreme drought stress with the recovery index mostly lower than 1.25 (Table 1; Fig. 3).
The leaf mass based maximum photosynthetic rate (A m) ranged from 0.0124 μmol g-1 s-1 in P. koraiensis to 0.0194 μmol g-1 s-1 in P. sylvestriformis (Table S1; Fig. 4a), sapwood-specific hydraulic conductivity (K s) and leaf-specific hydraulic conductivity (K l) ranged from 0.92 kg m-1 s-1 MPa-1 and 1.05×10-4 kg m-1s-1 MPa-1 in P. koraiensis to 1.32 kg m-1 s-1MPa-1 and 3.73×10-4kg m-1 s-1 MPa-1in P. densiflora , respectively (Fig. 4b). There is a significant positive correlation between K s andA m across the seven studied pine species (P < 0.05, Table S2; Pearson correlation). All the studied species showed relatively large hydraulic safety margins (HSM) ranging from 0.88 MPa in P. densiflora var. zhangwuensisto 1.29 MPa in P. koraiensis (Fig. 5a). Interspecific variation in wood density (WD) among the studied species overall showed a consistent order with that of HSM (Fig. 5a, b) that leads to a significant positive correlation between the two parameters (P< 0.05, Table S2; Pearson correlation). The order of interspecific variation in leaf level traits related to drought tolerance, as measured by leaf turgor loss point (π 0), showed different patterns with that of stem traits related to hydraulic tolerance to drought (Fig. 5a-c), which resulted in non-significant correlations across species between leaf and stem traits relevant to drought tolerance (Table S2).
Our results showed that tree drought resistance and resilience indices calculated from tree-ring analyses overall exhibited strong coordination with stem and leaf physiological traits related to drought tolerance (Fig. 6a-f). RT and RS were both positively correlated with WD (P< 0.01; Fig. 6a, d), HSM (P < 0.05, P= 0.115; Fig. 6b, e), and π 0 (P = 0.069 and 0.038; Fig. 6c, f). Contrastingly, both RT and RS showed significant negative correlations with functional traits pertinent to xylem hydraulic efficiency and leaf photosynthetic carbon assimilation (Fig. 7a-d). We found that tree radial growth rate as reflected by CBA15 has a strong negative correlation with growth sensitivity to inter-annual climate variability calculated from tree-ring analysis (P < 0.05; Fig. 8).