Vegetation water consumption can be influenced by the phase changes of the soil water and low temperature, particularly in the seasonally frozen region. There are also significant differences between the freeze-thaw and non-freeze-thaw periods. While it is still unclear whether the effects of the seasonal freeze-thaw on the water sources, utilization strategies, and hydrological responses of the vegetation. Thus, the coupling of the ecohydrological process of the "soil-vegetation-atmosphere" system can be expected for the seasonally frozen soil regions, especially for the vegetation restoration and fragile ecological protection of the sandy area. In this study, an in-situ site was established to monitor the soil water and temperature in the different depths, vegetation water consumption, and meteorological elements. The research objects were selected as the Salix psammophila and Artemisia desertorum in the Mu Us Sandy Land. Meanwhile, the samples were collected from the rainfall, soil water, groundwater, and vegetation during regular monitoring. The isotope compositions of the different samples were tested for measurement. Bayesian mixed model and Random Forest were used to explore the distribution, variation patterns, influencing factors, and the significance of the soil water, temperature, and isotope under different vegetation systems during both freeze-thaw and non-freeze-thaw periods. Results showed that the distribution and transformation of the soil moisture and temperature displayed a significantly seasonal trend under freeze-thaw processes. The shallow 0-40 cm layer was the most active area for the soil water and heat dynamics. The isotopic composition of the soil water was closely related to the precipitation and evaporation. There were significant differences in the isotopic fractionation during freezing and melting. The variation of δ¹⁸O ranged from-7.03 to 1.77‰ during the freezing period, while during the melting period, δ¹⁸O was accumulated to -3.27~5.71‰. Meanwhile, the soil profile was divided into three layers, including the shallow layer (0-40 cm), the middle layer (40-90 cm), and the deep layer (90-150 cm), according to the distribution and variation of the soil water, temperature, and isotope. Soil water content and soil temperature were the two most important environmental influencing factors on the soil water isotope, thereby contributing 32.6% and 44% to the soil water isotope in the shallow layer during freeze-thaw periods, respectively, while the contribution decreased with increasing depth. Besides, the isotope of the rainfall was also one environmental influencing factor on the isotopic changes of the soil water in the middle and deep layers during the freeze-thaw period. The distribution of the vegetation roots and the groundwater level was dominated in the different sites. Among them, the Salix psammophila and Artemisia desertorum also exhibited different strategies of water consumption. In site 1 with the large groundwater depth, the Salix psammophila relied mainly on the soil water in the shallow and middle layer during the non-freeze-thaw periods, while the water absorbing layer shared the downward trend during the freezing period, with the ratio of deep soil water up to 43.5%. Meanwhile, the Artemisia desertorum was utilized in the soil water of the shallow and middle layers, with the highest water utilization rate in the shallow layer reaching 68.1%. There was a decrease in the water consumption of both Salix psammophila and Artemisia desertorum in the site 2 with the shallow groundwater depth. The seasonal freeze-thaw process can play a critical role in the soil water, temperature, isotope distribution, and vegetation water uptake, thus serving as a key influencing factor on the ecohydrological coupling in the arid areas. Isotope modules can be expected to integrate into the numerical models in order to explore the hydrological cycle of the typical vegetation in cold and arid regions. The soil water transport can be further predicted in seasonally frozen areas, from the perspectives of the hydrodynamics, isotope, and water cycle. The finding can also provide scientific support to the desertification prevention and vegetation restoration of the vulnerable areas.

This study aims to in situ monitor the soil water use patterns of the vadose zone in the Mu Us sandy land during the growing season (March-September). The water-heat-δ¹⁸O simulation was coupled with particle tracking. Three fractionation models were selected as the No Fractionation (NF), Gonfiantini (GF), and Craig-Gordon (CG). The results showed that the kinetic characteristics of δ¹⁸O in the study area were best in the GF model, compared with the NF and CG models. The evaporation of the soil water was characterized by a temperature-controlled isotope fractionation migration. The particle and δ18O migration paths were established to clarify the spatial and temporal occurrence of the soil water recharge-evaporation, according to the precipitation with the vegetation water uptake. When the precipitation was scarce in spring and summer, the evaporation intensity and drought duration were used to determine the evaporation rate and evaporation depth of soil water. Among them, the evaporation depth of the soil reached 0-120 cm and 0-135 cm in April-May and June-July, respectively, due mainly to the strong evaporation. In addition, the priority of the infiltration-evaporation process depended on the initial water content of the soil caused by the precipitation distribution. That is, the infiltration rate of the dry soil was slow after light rainfall recharge. The evaporation capacity was enhanced significantly, where the depth of evaporation (0-13 cm) was equal to the depth of infiltration. At the same time, all particles entering the soil with the precipitation were dominated by the evaporation to leave from the ground surface; The depth of infiltration (20 cm) was greater than that of evaporation (16 cm) after the effective recharge followed by light rainfall. As such, the soil was effectively recharged after the process. Therefore, the soil water in the root zone was sourced from the frequently small-medium precipitation and a single large precipitation event, and then passed through the evaporation susceptibility zone (0-40 cm), finally remaining at a depth of 40-90 cm for a long period of time. Water use strategies of Salix soil depended on the root distribution and shallow soil water availability. The ablated soil stagnant water from the shallow (0 to 40 cm) zone was the main source of water consumption during the rejuvenation season. If the precipitation was abundant during the growing season, the shallow soil water (0-40 cm) was still the main source of the water uptake. The monthly precipitation also fully met the growth demand; Once the precipitation was scarce, the soil water was reduced at the surface soil to transfer the uptake layer. The main source of water uptake was attributed to the middle soil water (60-100 cm) that was retained by the precipitation in the previous period. As such, the process further clarified the hydrological coupling of the precipitation-soil-vegetation. The finding can also provide a scientific basis for the effective management of water resources and vegetation restoration in the dry zone.

To investigate the vertical propagation of meteorological drought into vadose zone soil drought at the southern margin of the Mu Us Sandy Land in the Yellow River Basin, this study utilized the Hydrus-1D model to simulate soil moisture dynamics at 0-150 cm depth from 1980 to 2020, based on in-situ monitoring data from grassland and bare land as well as long-term meteorological data. Using the standardized precipitation evapotranspiration index (SPEI), the past four decades were categorized into alternating wet-dry periods (S1), drought periods (S2), and wet periods (S3). Combined with the soil moisture deficit index (SMDI), methods including maximum correlation coefficient analysis, wavelet coherence analysis, and drought propagation indices were employed to explore drought propagation processes in grassland and bare land under different climatic regimes. The results indicated that meteorological drought propagation time to soil drought increased with soil depth. Propagation intensity generally weakened with depth, but the root zone of grassland exhibited weaker drought propagation intensity compared to other depths due to root distribution. Under different land cover conditions, grassland soil drought during the same period was more severe than bare land, characterized by longer drought duration, higher intensity, and deeper affected soil layers. Grassland responded more sensitively to meteorological drought-wet shifts, with drought propagation times 1-2 months faster than bare land across depths, and stronger propagation intensity. Drought propagation processes were influenced by climatic regimes. During S3, propagation time was shortest due to high precipitation and temperature, but propagation intensity was strongest in S1, followed by S3 and weakest in S2, owing to variations in evapotranspiration. Vertical drought propagation was jointly regulated by hydrological connectivity, root distribution, and evapotranspiration intensity. Given the stronger propagation intensity in grassland, vegetation density was suggested to be controlled in desertification control and ecological restoration projects to mitigate soil water overconsumption. The research results reveal the vertical propagation process from meteorological drought to soil drought, and illustrate the synergistic influence of vegetation cover and climate model on drought propagation process, which provides theoretical reference and practical basis for the work of desertification control in Mu Us Sandy Land.

Vapor-water exchange among atmosphere, soil and groundwater plays an important role in the surface ecological restoration in the vadose zones of arid and semiarid regions. However, it is limited to in-situ monitoring of water vapor flux, due to the complex and uncertain process. The simulation of water vapor cannot be fully verified by the measurement. It is still lacking in the vapor-water migration at different spatial and temporal scales. This study aims to clarify the spatiotemporal variation and migration of vapor-water exchange in the vadose zone of Mu Us Sandy Land by in-situ vapor-water monitoring, isotope tracer and numerical analysis. The results showed that the δD_L and δ¹⁸O_L of precipitation were enriched in spring and summer, but depleted in autumn and winter, due to the atmospheric water vapor source and local circulation. Soil evaporation was enriched in the oxygen isotope (δ¹⁸O_a) of surface soil vapor-water in summer more than that of soil freezing in winter. Moreover, there was a significant positive correlation (P<0.01) in the δ¹⁸O profile of the soil vapor-water and liquid water at different depths. The reason was that the evaporation intensity source and migration mode of soil vapor-water varied greatly in seasons. The δ¹⁸O values of vapor water in spring and summer showed an extremely significant linear positive correlation with the liquid water (P<0.01), while there was no significant correlation in winter (P=0.12). Moreover, there was a positive linear relationship between water vapor flux and δ¹⁸O_a. Furthermore, the δ¹⁸O_a was enriched in the surface layer at the large water vapor flux, including both downward and upward migration. The surface δ¹⁸O_a was enriched in the period of soil freezing in winter, particularly with the decrease of water vapor density (November-March). But the surface layer δ¹⁸O_a increased in summer (May-October) as well with the increase of surface water vapor density. Driven by the gradient of soil temperature in the vadose zone, the replenishment relationship of water vapor throughout the profile varied greatly at different periods. The water vapor of shallow soil was the supply source of deep vapor-water in summer, while the recharge was received from the deep layer in winter. But there were the temperature convergence and divergence zero flux planes in spring and autumn, respectively. The recharge characteristics caused the complicated relationship of vapor recharge in the profile. The soil δ¹⁸O_a was controlled by the water vapor migration, atmospheric evaporation and soil freeze-thaw in the vadose zone. In winter, the reduced evaporation and upward transport of water vapor can concurrently cause to enrich the surface δ¹⁸O_a, thus resulting in a decrease in the correlation between soil water vapor δ¹⁸O_a and liquid water δ¹⁸O_L. While in summer, the dominant effect can be from the diurnal evaporation and condensation cycle of soil water. The findings can provide a scientific basis to clarify the migration of soil water vapor in the water circulation