4.3. Implications for grassland shrubbization of alpine meadow
The plants at alpine P. fruticosa shrub extracted water from the middle soil layer to deep soil water preferentially from May to June, and extracted water from shallow, middle and deep soil in a balanced manner from July to October. The results were different from the results carried out on north of Qilian Mountains, where has a drier climate. Their results had shown that P. fruticosa used the shallow soil water except for August, when used the middle soil water (Zhang et al., 2022). This difference may be due to the different ages of plant (Wang, et al., 2021a), the different geographical locations (Li et al., 2022), or the different symbiotic species (Wu, et al., 2019). Additionally, the different sources of the soil water from the eco-hydrological processes maybe another reason.
Both δ18O and δ2H at the grassland site were enricher than at the shrub site during the growing season, but almost the same during the withering season. Grassland was more susceptible to drought stress, plants at grassland site had a stronger partitioning effect in dynamic transport, comparing with plants at the shrub site displayed more flexible water utilisations. From the perspective of plant water use patterns and evolution, alpine shrub on the northern QPT were formed from the long-term encroachment of P. fruticosa . This result is consistent with the results published previously by Klein et al. (2007) , who found that shrub populations are a result of proliferation and expansion of woody plant, and that areal expansion of shrubs is one of the most threatening forms of grassland degradation (Eldridge et al., 2011) . The contrasting plant water use patterns identified in our study provided theoretical basis for the alpine grassland shrubbization, contributed to predict vegetation functional dynamics and the ecohydrological process under current and future climatic conditions should take species-specific plasticity and responses to short-term water supply fluctuations into account.
5. Conclusions and future directions
The δ18O and δ2H values of precipitation, soil water, and plant water varied significantly over months at the alpine grassland and P. fruticosa shrub sites on the northern QTP, those of groundwater remained relatively stable during the sampling period. Considering groundwater, precipitation, soil water, and plant water, a progressive enrichment in δ18O or δ2H existed from groundwater and precipitation to soil water to plant water for each month. Alpine grassland was more susceptible to drought stress, had a stronger partitioning effect in dynamic transport than shrub. The P. fruticosa shrub displayed more flexible water utilisations, and was more competitive for water than grasslands. Furthermore, the plants in alpine shrub and grassland reached water use balance in August to October. Shrubs degraded from alpine grassland changed water use pattern of grassland, thereby changing soil water storage. These results contribute to understand the alpine grassland shrubbization from water use patterns of grassland and shrub plants on the northern Qinghai-Tibet Plateau, promote better understanding of the interface between plant and surrounding soils from the perspective of eco-hydrological processes on the QTP.
Our study explored the interaction of soil water and plant water at the alpine grassland and P. fruticosa shrub sites using stable isotopes, drawn important conclusions in the process of alpine grassland shrubbization. However, greater variety of species, more potential water sources (dew, creek and deeper groundwater), multi-site continuous observation, and longer time scales need to be took into consideration to further examine in the future.