2.2. Sample collections and isotope analysis
2.2.1 Sample collection
We collected the samples, including the plant, soil, precipitation and groundwater. All the samples were frozen at -10°C prior to vacuum distillation, for consistent handling while avoiding the effects of fractionation. In addition, our preliminary experimental results showed that the isotope ratio spectrometer had more stable operating state after extraction of precipitation and groundwater.
The plant and soil samples were collected from the shrub and grassland sites during the re-green season, growing season, and withering season in 2021 (20 May, 15 June, 16 July, 16 August, 7 September, and 12 October). Plant species were randomly chosen on the sampling date in the two sites. The whole plants of herbs were used for experiments after removing leaves. Whereas, the stem sections were removing the outer bark from P. fruticosa were used for isotope analysis (Martín-Gómez et al., 2017). The soil samples were collected around the selected plants within 2 m by soil cores, the soil cores were divided into 10 layers (0–5 cm, 5–10 cm, 10–15 cm, 15–20 cm, 20–30 cm, 30–40 cm, 40–50 cm, 50–60 cm, 60–80 cm, and 80–100 cm), each soil samples were analyzed as separate samples. Three replicates for each of the soil layers and plants samples were collected in the two experimental sites. A total number of 72 and 360 samples of plants and soils, respectively, were sealed for isotopic analysis.
Precipitation and groundwater were sampled concurrently from May to October in 2021. The event-based precipitations were sampled using bottles at herder’s home, 2 km away from our sampling sites during the sampling period, the number of replicates (one, two or three) from every event-based precipitation were determined according to the amount of precipitation. Groundwater were sampled once a week from a well 2 km away from our sites, three replicates were collected. The well had a depth of 2 m, and was commonly used for groundwater monitoring. A total of 162 and 72 samples from precipitation and groundwater, respectively, were all transferred into clean polyethylene bottles.
The soil water content (SWC) was determined by an automatic soil moisture monitoring system (CR800; Campbell, USA) with sensors installed at depths of 5 cm, 10 cm, 15 cm, 20 cm, 30 cm, 40 cm, 50 cm, 60 cm, 80 cm, and 100 cm below the soil surface. Precipitation measurements were collected using a precipitation gauge (52,203, RM Young, USA) at a height of 0.5 m. Temperatures were obtained from a meteorological station (Molis 520; Vaisala, Finland). All data were recorded every 30 min.
2.2.2. Isotopic analysis
Precipitation, groundwater, plant water, and soil water samples were all extracted by the cryogenic vacuum distillation technique (LI-2100pro, Lica United Technology Limited, Beijing, China) in order to avoid the influence of high salinity moisture on the accuracy of the instrument. After all samples were equilibrated to room temperature, extraction was started, setting 3 h for plant water and soil water, 2 h for precipitation and groundwater samples according to West et al. (2006). All of the extracted water was transferred into 2 ml vials, then analyzed stable isotopes (δ2H and δ18O) using an isotope ratio spectrometer (LGR-TLWIA-912, Los Gatos Research, San Jose, CA, USA). The instrument was equipped with an autosampler (PAL-LSI) for sample injection, and post-processing software (LWIA Post Analysis Full Installer v4.4.1) for test diagnosis, checking, and quantifying problems in the analysis (e.g., interference from organic pollutants, injection volume error) through detailed analysis of high-resolution absorption spectra. The measurement precision was 0.3‰ for δ2H and 0.1‰ for δ18O. The organic contamination on plant water need correction procedures to eliminate the influence (Schultz et al., 2011). The isotopic compositions of δ18O and δ2H were expressed as an isotope ratio:
\begin{equation} \delta sam(\%0)=\left(\frac{R\text{sam}}{R\text{std}}-1\right)\times 1000\%0\nonumber \\ \end{equation}
where δsam was the isotopic difference, Rsam was the abundance ratios (18O/16O and2H/1H) of samples, Rstd was the abundance ratios of standard.
2.2.3. Statistical analysis
All statistical analyses were conducted using R (software version 4.0.3, https://www.r-project.org/), and all figures were plotted using Origin 9.1(https://www.originlab.com/). One-way analysis of variance (ANOVA) followed by the post hoc Turkey’s test at p = 0.05 was used to assess hydrometeorological parameters of sampling date and sites. Two-way ANOVA was performed to examine the significant effects of hydrometeorological parameters and their interactions. Pearson’s correlations were tested at the p = 0.05 level.
The Bayesian mixing model Mix SIAR (http://conserver.iugo-cafe.org/user/ brice.semmens/Mix SIAR) was used for identifying the proportions of contributions from each water source according to the δ2H and δ18O, which were considered as the mixture data of the potential water sources (Dawson&Ehleringer, 1993; Beyer et al., 2018). The inputs of original data (for example, the 0-20 cm soil layer data was input as 0-10 cm and 10-20 cm isotope data), the discrimination data (the TDF data in the model), the running time of Markov Chain Monte Carlo (MCMC), and the diagnosis method of the model results were according to (Zhou et al., 2021). The average value was output from the model. Three potential soil water sources were identified to facilitate subsequent analysis (i.e., shallow soil water (0–20 cm), middle soil water (20–60 cm), and deep soil water (60-100 cm)), according to the variability in the soil water content and the impacts of precipitation pulse.
3. Results