Characteristic of the leaf functional traits in different populations of L. ruthenicum
In this study, we measured 14 leaf functional traits (Table 2). Among them, leaf water physiology traits TWC, RWC, SLA, SLV, LT, LDMC, Suc, LD were 79.35-88.37%, 70.41-137.35%, 5-8cm2▪g-1, 5.36-12.80 cm3▪g-1, 1.02-1.62 mm, 125.0-197.9 mg▪g-1, 0.80-1.38 g▪cm-2, 0.08-0.19 g▪cm-3, respectively. Leaf ecological stoichiometry traits, namely C, N, and P contents were 307.39-351.78, 8.09-17.82, 0.62-5.77 mg▪g-1, respectively, and C:N, C:P and N:P ratios were 20.28-37.97, 56.85-415.44, 2.79-17.70, respectively.
We compared the differences between L. ruthenicum functional traits at eight different moisture and salinity sites (Table 1) and found that greater leaf thickness appeared in slightly saline site VIII which was significantly different from non-saline Gobi sites I and VI (Table 2). In addition, the largest SLV, Suc, TWC and RWC traits were also observed to appear at slightly saline sites VIII. Conversely, LDMC, LD, and N contents had the lowest values in heavy saline sites. Leaf N concentration was the least variable between different regions, but it still showed the effects of heavy saline stress on L. ruthenicum via significantly lower leaf C content. Statistical analysis showed that the ratios of N:P were more stable than the ratios of C:N and C:P in adaptation to drought-salt stress. Moreover, we found that there was no significant difference in the SLA trait values between the eight different habitats, showing that intra-specific variation in SLA at our finer ecological scale was minimal or non-existent.
Correlation between leaf functional traits of L. ruthenicum in different habitats
Correlation coefficients between 14 leaf traits of L. ruthenicum showed that LT was significantly positively correlated with Suc (r=0.60, P<0.05), but significantly negatively correlated with C content (r=-0.75, P<0.001). SLV was highly positively correlated with SLA (r=0.74, P<0.001) and both were significantly negatively (r=-0.95, P<0.001 and r=-0.68, P<0.01 respectively) correlated with LD and significantly positively correlated with TWC (r=0.78, P<0.001 and r=0.56, P<0.05, respectively). LDMC was significantly positively correlated with LD (r=0.87, P<0.001), and both were significantly negatively correlated with TWC (r=-0.94, P<0.001 and r=-0.81, P<0.001, respectively). Suc was significantly positively correlated with TWC and RWC, but was significantly negatively correlated with C content (r=-0.86, P<0.001). TWC was significantly positively correlated with P content (r=0.56, P<0.05), while P content was significantly negatively correlated with N:P and C:P ratios (r=-0.77, r=-0.87,P<0.001). N:P and C:P ratios were significantly positively correlated with each other (r=0.95,P<0.001), while RWC was highly negatively correlated with N:P and C:P ratios (r=-0.65, r=-0.67, P<0.01).
RDA restriction ordering of functional traits in soil water and salinity gradients
Two RDA maps of different soil layers showed the distribution pattern of traits along the salinity gradients. From non-saline to slightly saline gradients, populations had higher C:N ratios, lower N content, and lower N:P ratios (see RDA vertical axis direction). In the horizontal axis, populations growing in high salinity soils had lower C content than populations growing in lower salinity soils (Fig.2, Table 2), while the distribution of other leaf traits didn’t change much with environmental gradients. 0–40cm and 40-80cm soil properties respectively explained 70.99% and 71.09% of leaf traits variation (the sum of the first two axes explained). Permutation tests for all canonical axes were not significant (0-40cm RDA, Df=10, F=1.53, Pr(>F)=0.31; 40-80cm RDA, Df=10, F=1.56, Pr(>F)=0.29, Fig.2). In general, the spatial distribution of the eight community types might be driven by variation in soil chemical characteristics. Populations I, II, III, IV, and VI were close to each other due to their similar soil chemistry, as were populations V and VII. However, population VIII was located away from the other populations, so its soil properties likely differed from the other locations.
Relative importance of soil factors to leaf trait variation
We were not only interested in the effects of total soil salinity on leaf functional traits, but also in the exploration of which salt ions affect plant functional trait formation and variation the most. In general, moisture, salinity, and eight major ions corresponded to leaf character variation in different amplitudes. In the 0-40cm soil layer, leaf traits patterns were mainly influenced by SWC, HCO3- and CO32-, and their relative importance values for the fourteen leaf traits are shown in Fig.3. The relative contribution of 0-40cm layer SWC to all but the LT trait was more than 17%, which might be because soil moisture had no direct effect on leaf thickness. SWC had a significant effect on C:P ratios, with an importance of 34%. HCO3- was more than 13% important for all traits except SLV and N content. CO32-was less important for traits than SWC and HCO3-. Soil salinity and other ions contributed relatively little to leaf properties. In the 40-80cm layer, HCO3- and SO42- were two main drivers for trait differentiation. The relative importance of HCO3- for all trait patterns was higher than 20%, and its influence on P content was up to 52%. The influence of SO42- on traits was above 12%, except for LDMC, LD, and N content, which were below 10% (values below 10 are not shown in the Fig.3).