4.2. Comparisons between P. fruticosa shrub and alpine grassland
A series of water line equations between δ18O and δ2H were established to compare the differences in water interaction from shrub and grassland sites (Figure 5). Soil and plant water samples were well described by linear regressions, resulting of laying at an angle to the LMWL, this is consistent with previous studies (Goldsmith et al., 2012; Evaristo et al., 2015; Che, et al., 2019). The slopes and intercepts of these water lines were determined by the relative evaporation rates of the different water isotopes (Crawford et al., 2014; Benettin et al., 2018; Bowen et al., 2018; Chi et al., 2019), indicating the different magnitudes of evaporative enrichment of isotopes in soil water and plant water. The slope and intercept of soil water at the grassland site (5.46 and -8.66, respectively) were slightly lower than those at the shrub site (5.83 and -4.83, respectively), suggesting soil evaporation was slightly greater at the grassland site than at the shrub site. This was probably because the shrub site had a denser coverage shading the soil surface compared with the grassland site, just as daily evaporation rates were slightly lower for understory vegetation than for grassland during the growing season (Crawford, et al., 2014; Schwärzel et al., 2020). Nevertheless, the slopes and intercepts of δ18O and δ2H in plant water at the grassland site (2.46 and -24.57, respectively) were higher than the slope and intercept at the shrub site (1.54 and -33.75, respectively). The differences likely resulted from substantial variabilities for leaf water at shrub and grassland sites, evaporative distinctions of 2H/1H and18O/17O on the primary of leaf water(Farquhar et al., 2007), and the transpiration from an area of grass is greater than the transpiration from a similar area of shrubs.
Because the two sampling sites neighbored each other, they had similar topographical and geological conditions, and precipitation δ18O and δ2H were also similar. The δ18O in soil water and plant water showed significant differences between shrub and grassland sites (p < 0.05) in contrast to the non-significant differences in δ2H between the two sites (p = 0.068 and 0.06 for soil water and plant water, respectively), resulting in different slopes and intercepts for the water line equations from soil water and plant water between the two sites. Isotope-fractionated differences between δ18O and δ2H were probably associated with local microenvironment and heterogeneity differences of surroundings between the grassland and shrub sites despite the same general conditions (Ellsworth&Williams, 2007).