Drought-Tolerance Characteristics and Water-Use Efficiency of Three Typical Sandy Shrubs

Abstract

Elucidating shrub ecohydrological adaptation is critical for optimizing vegetation-restoration strategies in arid regions and maintaining regional ecological stability. This study examined typical desert shrubs at the northern edge of the Mu Us Sand Land. During the growth peak season (July–September), we measured understory-soil δ¹⁸O, soil water content (SWC), leaf δ¹³C_p, stem δ¹⁸O, and gas-exchange rates, and evaluated shrub drought resistance and water-use efficiency using Mantel tests and principal component analysis (PCA). Based on the VPDB standard, the δ13Cp values of leaves ranked as follows: Caragana microphylla (−27.21‰) > Salix psammophila (−27.80‰) > Artemisia ordosica (−28.48‰). The results indicate that leaf δ¹³C_p and water δ¹⁸O are effective indicators of shrub water-use efficiency, reflecting Cᵢ/Cₐ dynamics and water-transport pathways, respectively. The three shrubs exhibit distinct water-use strategies: Caragana microphylla follows a conservative strategy that relies on deep-water sources and tight stomatal regulation; Salix psammophila shows an opportunistic strategy, responding to precipitation pulses and drawing from multiple soil layers; Artemisia ordosica displays a vulnerable, shallow-water-dependent strategy with high drought susceptibility. SWC was the primary driver of higher Long Water Use Efficiency (WUE), whereas Mean Air Temperature (MMAT) and Mean Relative Humidity (MMRH) exerted short-term regulation by modulating the vapor-pressure deficit (VPD). We conclude that desert-shrub water-use strategies form a complementary functional portfolio at the community scale. Vegetation restoration should prioritize high-WUE conservative species, complement them with opportunistic species, and use vulnerable species cautiously to optimize community water-use efficiency and ecosystem stability.

1. Introduction

Water availability is a primary limiting factor for plant growth and development, and its spatiotemporal variability directly influences plant performance and distribution patterns [1]. Under global climate change, hydrological regimes in arid and semi-arid regions have shifted markedly, affecting plant growth [2] and intensifying survival pressures. For decades, vegetation planting has been a central strategy for ecological restoration in arid and semi-arid regions [3], and its success depends strongly on the match between species and regional water regimes. Accordingly, plant adaptability to water conditions is fundamental for sustaining normal growth and for establishing stable, long-lived windbreak and sand-fixing vegetation systems [4]. With advances in stable-isotope techniques, oxygen (δ¹⁸O) and carbon (δ¹³C_p) isotopes have become, respectively, natural tracers of plant water sources [5] and indicators of water-use efficiency [6]. δ¹⁸O is increasingly preferred over traditional approaches such as root surveys/excavations [7] and sap-flow measurements [8], providing a more sensitive and accurate means to quantify plant water sources [9]. These isotopic methods are widely applied in croplands [10,11], plantation forests [12,13], and wetland vegetation to resolve water sources [14,15]. McCole et al. [16] showed that root depth structures water-source allocation: deep-rooted trees and shrubs predominantly use relatively stable deep-soil water, whereas shallow-rooted herbs rely on more variable near-surface water. Baoqing Liu et al. [17] quantified soil-water uptake depths for Pinus tabulaeformis, Caragana microphylla, and Artemisia frigida in the Horqin Sandy Land, and Su Wenxu et al. [18] recorded water-use depths for Larix principis-rupprechtii and Salix psammophila in the Hunshandake Sandy Land. Together, these studies consistently corroborate this pattern. Plant water-use efficiency is commonly evaluated as WUE and WUEi. WUE, a core comprehensive parameter characterizing plant physiological status [19], is derived by measuring leaf δ¹³C_p of plants, calculating Δ¹³C, and converting the resulting values. It also serves as a critical criterion for evaluating species suitability under climate change and water scarcity [20]. Plant carbon isotopes (δ¹³C_p) reflect time-integrated carbon assimilation, thereby indicating adaptation to water use and water stress over the integration period. δ¹³C_p is positively correlated with WUE [4]. Environmental drivers such as precipitation, relative humidity, and air temperature are primary regulators of WUEi [21]. Consistently, Zhao et al. [22] reported that soil moisture, plant transpiration, and air temperature are the dominant factors controlling WUEi.

Δ¹³C and δ¹⁸O can be jointly analyzed to evaluate and apply plant drought-resistance traits [23]. At physiological and molecular levels, drought stress perturbs plant cells by triggering reactive-oxygen-species (ROS) bursts, disrupting osmotic balance, and damaging membrane systems. Moderate ROS acts as signals that induce stomatal closure and stress-response gene expression, whereas excessive accumulation causes oxidative injury. Redox homeostasis is maintained by the glutathione–ascorbate cycle together with enzymatic defenses (SOD, POD, CAT) and non-enzymatic antioxidants [24]. Concurrently, proline—a key osmotic regulator [25]—acts with soluble sugars to alleviate osmotic stress [24]. Studies of arid-zone plants by Lü, X.-P., Yang, F., and Chang, Y. conclude that organic osmolytes are central to adaptation under prolonged drought [26,27,28].

During vegetation surveys in the Mu Us Sandy Land, we observed that, despite a shared macroclimate, Caragana microphylla, Salix psammophila, and Artemisia ordosica showed distinct growth performance and physiological responses during the growth peak season. Caragana microphylla maintained stable growth, whereas Salix psammophila and Artemisia ordosica frequently exhibited leaf wilting. This observation raises a core question: within the same climatic zone, which mechanisms determine interspecific differences in drought tolerance and water-use efficiency? Based on preliminary surveys and prior literature, we hypothesize that these differences arise from the joint effects of plant physiological regulation and environmental moisture regimes.

This study investigated gas-exchange rates, SWC, and stable isotopes for three typical desert shrubs along the northern edge of the Mu Us Sandy Land, including depth-resolved δ¹⁸O of soil and stems and leaf δ¹³C_p. Through repeated sampling over a period of months during the growth peak season, we assessed drought-resistance traits and temporal changes in water-use efficiency. To elucidate physiological response mechanisms and dominant drivers under variable environments, we analyzed species-specific water-use strategies and provided guidance for species selection and configuration in ecological restoration of arid regions.

2. Materials and Methods

2.1. Overview of the Study Area

As shown in Figure 1, the study site is in the Wushen Banner, southwestern Ordos City, Inner Mongolia Autonomous Region, China. The area lies at the northern edge of the Mu Us Sandy Land and typifies a composite ecosystem of wind-blown sand, vegetation, and water resources. The study area is at an elevation of 1310–1330 m. Soils are typical sandy loam and are prone to severe wind erosion, which produces coarse soil structures. Groundwater generally occurs at depths of 3–5 m, and surface water consists mainly of seasonal rivers and lakes. The region has an arid–semi-arid temperate continental climate with pronounced diurnal temperature fluctuations. The annual average temperature is 6.0–8.5 °C, and annual precipitation is 350–400 mm, predominantly from May to August (over 75% of the annual total). Annual evaporation averages 2000–2800 mm, equivalent to 5–6 times the annual precipitation. Average annual sunshine is 2911.5 h, and the frost-free period is 123–156 d. Dominant plant species in this region include Salix matsudana, Salix psammophila, Juniperus sabina, Hippophae rhamnoides, Caragana microphylla, Artemisia ordosica, and Stipa grandis.

Figure 1. Location of the study area and vegetation cover characteristics. Note: The base map is produced by the National Basic Geographic Information Center of the Ministry of Natural Resources, with the map review number “GS (2024) 0650”, and the boundaries of the base map have not been modified.

2.2. Experimental Design and Sample Selection

Ecological restoration at the northern edge of the Mu Us Sandy Land relies heavily on local xerophytic shrubs. As dominant, functionally pivotal species, these shrubs are critical for maintaining ecosystem structure and function. The three most representative dominant shrubs in this area are Caragana microphylla, Salix psammophila, and Artemisia ordosica [4]. These shrubs play a leading role in local plant communities, exhibiting very high occurrence frequency and broad ecological niche breadth. This study was conducted during the peak vegetation growth season spanning July to September 2024. This study comprises three research sites, each dominated by a single plant species. The area of each site, delineated via UAV surveys and the corresponding vegetation coverage, is as follows: Caragana microphylla (28,164 m², 34%), Salix psammophila (31,174 m², 31%), and Artemisia ordosica (27,811 m², 37%). Within each plot, three representative plants of similar topography and vigorous growth were selected. Three branches per plant were used as replicates. At a level site within each plot, a rain gauge, an evaporation pan, and an air temperature–humidity sensor were installed to monitor MMR (Monthly Mean Rainfall, mm), MME (Monthly Mean Evaporation, mm), MMAT (Monthly Mean Air Temperature, °C), and MMRH (Monthly Mean Relative Humidity, %). Data were logged every 10 min throughout the experimental phase. Table 1 summarizes plant growth status in the plots; Figure 2 presents the meteorological characteristics of the study area.

Table 1. Growth status of plants in the plot.

Plant Name	Plant Height/m	Crown Width/m	Plant Type	Carbon Fixation Pathway Functional Types
Caragana microphylla	2.12 ± 0.17	3.58 ± 0.10 × 3.75 ± 0.50	Shrub	C3
Salix psammophila	2.68 ± 0.28	2.44 ± 0.13 × 1.60 ± 0.05	Shrub	C3
Artemisia ordosica	1.00 ± 0.12	1.85 ± 0.09 × 1.58 ± 0.24	Shrub	C3

Figure 2. Monthly variation characteristics of meteorological factors in the study area in 2024. MMR, NNE, MMAT, and MMRH denote the Monthly Mean Rainfall, Monthly Mean Evaporation, Mean Air Temperature, and Mean Relative Humidity, respectively.

Figure 2 presents the monthly variations in meteorological factors in the study area for 2024. Based on monthly records for 2024, the cumulative precipitation was 467.5 mm, the annual mean air temperature was 8.35 °C, and the annual mean relative humidity was 5.75%. MMR was uneven across months, and MMAT exhibited pronounced month-to-month variability. MMRH was strongly influenced by both MMR and MMAT, and the growth peak season occurred from July to September. The study used MME values for July, August, and September—179.9, 135.6, and 78.7 mm, respectively.

The total rainfall during the growth peak season in the study area was 270.4 mm, accounting for 57.84% of annual precipitation. Within this period, August recorded the highest MMR at 147 mm. July was exceptionally dry, with MMR of only 19 mm. Nevertheless, July recorded the highest MMAT of the year (24.3 °C), and MME reached 179.9 mm. At the same time, the evaporation in that month was 9.5 times the precipitation, indicating a severe drought climate. Relative humidity exhibited an inflection in July; the combination of low MMR and high MMAT likely subjected plants to drought stress. With increased rainfall in August, air temperature trended downward, whereas relative humidity continued to rise, peaking at 79.7% in September.

2.3. Research Methods

2.3.1. Determination of Leaf Carbon Isotopes

Field Sample Collection

The experiment was conducted during the growth peak season (July–September 2024). Sampling was performed three times per month on sunny days with no rainfall during the preceding three days. Sampling took place between 8:00 a.m. and 11:00 a.m. on each sampling day. Forty to fifty intact leaves were collected from the sun-exposed central canopy of each plant and used as replicates. After collection, leaves were placed in kraft-paper sample bags.

Indoor Testing

Plant leaf samples were sent to the Wind-Sand Physics Experiment Building of the Inner Mongolia Agricultural University and dried in an oven at 65 °C for 24 h. After drying, samples were ground in a pulverizer, sieved through a 100-mesh screen, and stored in sealed polyethylene bags to prevent moisture reabsorption. The powdered sample was placed in a tin boat, weighed to approximately 2 mg on a microbalance, and sealed to complete sample preparation. The prepared samples were combusted at high temperatures in an elemental analyzer. Carbon (C) was oxidized to CO₂, which was detected by a thermal conductivity detector. The carbon content was calculated by comparing sample peak areas with those of 2–3 working standards. The ¹³C/¹²C ratio of CO₂ was determined using an isotope mass spectrometer.

Calculation Method

The δ¹³C_p value of the sample was calculated by comparing the measured CO₂ ¹³C/¹²C ratio with the international standard (VPDB). δ¹³C_p (plant leaf carbon-isotope ratio) reflects isotopic fractionation during photosynthesis. The calculation formula is:

$${\mathsf{\delta }}^{13}{\mathrm{C}}_{\mathrm{p}}={\displaystyle \frac{{({}^{13}\mathrm{C}/{}^{12}\mathrm{C})}_{\mathrm{p}}-({}^{13}\mathrm{C}/{}^{12}\mathrm{C}{)}_{\mathrm{VPDB}}}{{({}^{13}\mathrm{C}/{}^{12}\mathrm{C})}_{\mathrm{VPDB}}}}$$

(1)

δ¹³C_p denotes the per-mille (‰) deviation of the sample (¹³C/¹²C) ratio relative to the standard sample, ¹³C/¹²Cp represents the sample (¹³C/¹²C) ratio, and (¹³C/¹²C)_VPDB denotes the (¹³C/¹²C) ratio of the standard material VPDB.

The atmospheric CO₂ carbon-isotope ratio (δ¹³C) is a standardized parameter that describes the carbon-isotope composition of atmospheric CO₂ [29]. It is used to remove the influence of atmospheric background values when calculating leaf stable carbon isotope ratios (Δ¹³C). The calculation formulas are:

$${\mathsf{\delta }}^{13}{\mathrm{C}}_{\mathrm{a}}=-6.429-0.006\exp \left(0.0217\left(\mathrm{t}-1740\right)\right)$$

(2)

δ¹³C_a is expressed in per mille (‰), t denotes the sampling year; in this study, t = 2024. Substituting into Equation (2) yields δ¹³C_a = −9.28‰.

Δ¹³C (stable carbon-isotope fractionation in plant leaves) reflects the extent of isotopic fractionation during photosynthesis. The calculation method for this indicator is as follows:

$${\mathsf{\Delta }}^{13}\mathrm{C}={\displaystyle \frac{{\mathsf{\delta }}^{13}{\mathrm{C}}_{\mathrm{a}}-{\mathsf{\delta }}^{13}{\mathrm{C}}_{\mathrm{p}}}{1+{\mathsf{\delta }}^{13}{\mathrm{C}}_{\mathrm{p}}/1000}}$$

(3)

Δ¹³C is expressed in per mille (‰). δ¹³C_p and δ¹³C_a represent the carbon-isotope ratios of plant leaves and atmospheric CO₂, respectively.

C_a (Atmospheric CO₂ concentration) denotes the molar mixing ratio of CO₂ in air, indicating the availability of carbon for plant photosynthesis. It is a pivotal variable linking the global carbon cycle to plant physiological processes; its dynamics directly regulate ecosystem productivity and climate adaptability. The calculation formula is as follows:

$${\mathrm{C}}_{\mathrm{a}}=277.78 + 1.350\exp \left(0.01572\left(t-1740\right)\right)$$

(4)

t denotes the sampling year; in this study, t = 2024. Substituting t = 2024 into Equation (4) yields Ca = 395.14 μmol/mol.

WUE (Long Water Use Efficiency) is the amount of carbon fixed per unit of water consumed over a plant’s growth cycle. It can be indirectly estimated from carbon-isotope resolution (Δ¹³C), thereby indicating a plant’s integrated regulation of the carbon-water balance and its adaptive strategy under variable environments. The calculation formula is:

$$\mathrm{WUE}={\displaystyle \frac{{\mathrm{C}}_{\mathrm{a}}\left(\mathrm{b}-{\mathsf{\Delta }}^{13}\mathrm{C}\right)}{1.6\left(\mathrm{b}-\mathrm{a}\right)}}$$

(5)

WUE is expressed in μmol mol⁻¹; a = −4.4‰ denotes the diffusive fractionation coefficient for CO₂ entering stomata; b = 27‰ represents the fractionation coefficient during CO₂ carboxylation by Rubisco; 1.6 is the diffusivity ratio of water vapor to CO₂ in air; and Ca denotes the atmospheric CO₂ concentration.

2.3.2. Plant and Soil Oxygen Isotope Analysis

Field Sample Collection

The experiment was conducted during the growth peak season (July–September 2024). To minimize sampling error, three sampling rounds were conducted per month. Sampling was carried out on sunny days with no rainfall during the preceding three days, between 8:00 a.m. and 11:00 a.m. Each sampling point was located midway along the transect from the plant center (origin) to the canopy edge above the root zone. To minimize isotopic enrichment caused by stomatal transpiration, stems older than two years were selected. Branch segments approximately 0.3–0.5 cm in diameter and 3–5 cm in length were sampled from fully lignified, mature secondary woody branches. Three replicates were collected per plant after removing the outer bark. Within a 1 m radius of each plant, soil was extracted from 0 to 100 cm depth using a soil auger. The soil profile was partitioned into five layers, each 20 cm thick. All samples were immediately placed in 10 mL amber glass vials, sealed with parafilm, and stored in a portable refrigerator at 4 °C.

Indoor Testing

Within ten days following the completion of sampling, at the Wind-Sand Physics Laboratory of Inner Mongolia Agricultural University, the LI–2100 Automatic Vacuum Cold Condensation Extraction System, developed by Beijing Ruijia United Technology Co., Ltd., was employed to extract moisture from plant stem and soil samples. Impurities were removed using a 0.22 µm membrane filter, and oxygen isotopes in the water samples were analyzed with an isotope mass spectrometer.

Calculation Method

The δ¹⁸O value (oxygen-isotope ratio of water from plant stems and soils) is calculated by comparing the sample ¹⁸O/¹⁶O ratio with the international standard (VSMOW). The calculation formula is:

$${\mathsf{\delta }}^{18}\mathrm{O}={\displaystyle \frac{{({}^{18}\mathrm{O}/{}^{16}\mathrm{O})}_{\mathrm{sample}}}{{({}^{18}\mathrm{O}/{}^{16}\mathrm{O})}_{\mathrm{SMOW}}}}$$

(6)

δ¹⁸O is expressed in per mille (‰), (¹⁸O/¹⁶O)_sample denotes the isotopic ratio of oxygen-18 to oxygen-16 in the sample, and (¹⁸O/¹⁶O)_SMOW denotes the corresponding ratio in the standard material.

2.3.3. Determination of Soil Moisture Content

Field Sample Collection

Experiments were conducted during the growth peak season (July–September 2024). Sampling was conducted three times per month on sunny days with no rainfall during the preceding three days. Samples were collected between 8:00 a.m. and 11:00 a.m. on each sampling day. Within a 1 m radius of each study plant, soil was extracted from 0 to 100 cm depth using a soil auger. The soil profile was partitioned into 10 layers, each 10 cm thick. Collected soil samples were placed in pre-weighed aluminum boxes (55 × 35 mm), and the total mass was recorded. Each soil layer was sampled in triplicate.

Indoor Testing

After sampling was complete, the samples were transported to the Wind-Sand Physics Experiment Building of Inner Mongolia Agricultural University. The samples were weighed on a precision balance with an accuracy of 0.0001 g, and the mass was recorded as the combined mass of fresh soil and the aluminum container. The container was opened and placed in an oven at 105 °C. The samples were then dried for 24 h, then removed and weighed. The mass was recorded as the combined mass of dried soil and the aluminum container.

Calculation Method

SWC (Soil Water Content) is the mass of water per unit mass of dry soil and is a fundamental metric of soil moisture. The test employed the oven-drying method using aluminum boxes; the gravimetric formula is:

$$\mathrm{SWC}={\displaystyle \frac{\left({\mathrm{M}}_{\mathrm{wet}}-{\mathrm{M}}_{\mathrm{dry}}\right)}{\left({\mathrm{M}}_{\mathrm{dry}}-{\mathrm{M}}_{\mathrm{tin}}\right)}}\times 100$$

(7)

WSC is expressed as a percentage (%), M_wet represents the combined mass of wet soil and the aluminum container, M_dry denotes the combined mass of oven-dried soil and the aluminum container, and M_tin indicates the mass of the aluminum container.

2.3.4. Measurement of Gas Exchange in Plants

Gas Exchange Measurements

The experiment was conducted during the growth peak season (July–September 2024). Measurements were taken three times per month to reduce measurement error. Field measurements were conducted on sunny days with no rainfall during the preceding three days, between 8:30 a.m. and 11:30 a.m. Ten unshaded, well-illuminated, fully expanded leaves were selected as replicates from the sunlit canopy. Measurements were conducted using the Li-6800 Portable Photosynthesis System (manufactured by LI-COR Biosciences, Lincoln, NE, USA), which is equipped with a 2 cm²; fluorescent leaf chamber; external instrument parameters were adjusted to align with in situ field conditions. Pn and Tr were measured monthly for plants in the study area.

Calculation Method

WUEi (Instantaneous Water Use Efficiency) is the amount of CO₂ fixed per unit water lost by a plant at a given moment. The calculation formula is:

$$\mathrm{WUEi} = {\displaystyle \frac{\mathrm{Pn}}{\mathrm{Tr}}}$$

(8)

WUEi is expressed in mmol·mol⁻¹; Pn denotes net photosynthetic rate, Tr denotes transpiration rate.

2.4. Data Processing and Analysis

One-way ANOVA tested monthly differences in photosynthetic parameters, soil water content, soil and stem δ¹⁸O, and plant water-use efficiency. Before ANOVA, data normality and homoscedasticity were assessed with the Shapiro–Wilk and Levene’s tests, respectively. Mantel tests quantified correlations between plant indices and meteorological factors; results were visualized as correlation heatmaps. For principal component analysis (PCA), variables were selected a priori based on ecological relevance and prior studies; all variables were standardized to z-scores to remove scale effects, and key drivers of plant water-use dynamics were identified using a 3D PCA plot. Data processing and statistical analyses were performed in Excel 2010 and SPSS Statistics 27, and figures were generated in Origin 2024 and R 4.5.0.

3. Results

3.1. Soil Moisture Content and Oxygen Isotope Characteristics in Shrubland Understory Soils

During the growth peak season, SWC (soil water content) in the study plots of Caragana microphylla, Salix psammophila, and Artemisia ordosica ranged from 0.90 to 3.22%, 0.37–3.37%, and 0.53–3.14%, respectively. The mean SWC values were 2.37% ± 0.68%, 2.44% ± 0.01, and 2.28% ± 0.03, respectively. Average SWC were ranked as follows: Salix psammophila > Caragana microphylla > Artemisia ordosica, indicating higher soil moisture under Salix psammophila. As shown in Figure 3a–c, the monthly mean SWC in Caragana microphylla plots generally increased initially and then declined. During the growth peak season, SWC rose continuously at 0–30 cm, increased and then decreased at 30–60 cm (peaking at 2.77% in August), and reached a maximum of 3.12% at 60–100 cm in July. In the Salix psammophila plot, monthly mean SWC generally exhibited an upward trend. During the growth peak season, SWC at 0–30 cm increased markedly, peaking at 2.49% in September. At 30–60 cm, SWC first increased and then decreased, peaking at 2.72% in August. At 60–100 cm, SWC remained stable before rising rapidly to a peak of 3.15%, the highest among all plots. In the Artemisia ordosica plot, the monthly mean SWC showed an overall gradual increase with a decelerating rate. During the growth peak season, SWC at 0–30 cm continued to rise, whereas at 30–60 cm it first decreased and then increased, reaching a trough of 2.46% in August. At 60–100 cm, SWC first increased and then decreased, peaking at 2.83% in August.

Figure 3. Soil moisture characteristics in the shrubland understory. (a) Soil moisture content of Caragana microphylla, July–September; (b) soil moisture content of Salix psammophila, July–September; (c) soil moisture content of Artemisia ordosica, July–September.

As shown in Figure 4a, δ¹⁸O values of soil water beneath Caragana microphylla shrubland generally decreased with increasing soil depth in July. At 0–20 cm, enhanced evaporation in July enriched δ¹⁸O in the surface soil, reaching 1.92‰. At 80–100 cm, δ¹⁸O was depleted, reaching −6.51‰. In August, a sharp increase in rainfall drove δ¹⁸O to a minimum of −15.18‰ across the Caragana microphylla soil profile, while values at 80–100 cm were −7.00‰, indicating strong precipitation effects in shallow layers and relative stability at depth. In September, δ¹⁸O values beneath Caragana microphylla converged across depths (−6.83 to −6.85‰), indicating thorough mixing of soil water. As shown in Figure 4b, in July, δ¹⁸O in soil water beneath Salix psammophila shrubland declined steeply and then stabilized. At 0–20 cm, enhanced evaporation enriched δ¹⁸O in surface soil, peaking at 5.04‰. At 40–60 cm, values decreased sharply to −2.92‰; at 80–100 cm, they stabilized with a slight increase to −2.24‰. In August, δ¹⁸O in soil water beneath Salix psammophila was uniformly depleted, indicating rainfall influenced the entire 0–100 cm profile. By September, δ¹⁸O at all depths converged to −5.71 to −6.42‰, suggesting a uniform water source. As shown in Figure 4c, in July, δ¹⁸O in soil water beneath Artemisia ordosica shrubland declined steeply and then stabilized. At 0–20 cm, enhanced evaporation enriched δ¹⁸O in the surface layer, reaching 4.37‰. At 60–80 cm, values rose slightly to −0.35‰. This depth range may have experienced short-term δ¹⁸O enrichment due to increased plant root exudates that enhanced local clay content, forming a biogenic microstructure layer that impeded water flow and intensified evaporation. In August, δ¹⁸O in soil water beneath Artemisia ordosica was uniformly depleted from 0 to 60 cm, indicating rainfall influence across this interval. At 80–100 cm, values reached −7.84‰. In September, δ¹⁸O beneath Artemisia ordosica converged across all depths (−6.45 to −6.05‰), indicating thorough soil-water mixing and a uniform water source.

Figure 4. Soil-water oxygen-isotope characteristics. (a) Soil-water δ¹⁸O beneath Caragana microphylla, July–September; (b) Soil-water δ¹⁸O beneath Salix psammophila, July–September; (c) Soil-water δ¹⁸O beneath Artemisia ordosica, July–September.

3.2. Gas Exchange Variables

As shown in Figure 5, photosynthetic parameters of the three shrubs showed highly significant correlations (p ≤ 0.01), and Pn exhibited an overall monthly upward trend from July to September, the net photosynthetic rates of Caragana microphylla, Salix psammophila, and Artemisia ordosica increased by 16.88%, 22.46%, and 23.74%, respectively, peaking in September at 12.27, 10.10, and 8.63 µmol·m⁻²·s⁻¹. Tr showed an overall decreasing trend; from July to September, the reported percentage changes for Caragana microphylla, Salix psammophila, and Artemisia ordosica were +38.00%, +54.19%, and +58.64%, respectively, with trough values in September of 1.57, 1.27, and 1.53 mmol·m⁻²·s⁻¹. Accordingly, during the growth peak season, the stability of photosynthetic parameters ranked as: Caragana microphylla > Salix psammophila > Artemisia ordosica.

Figure 5. Characteristics of gas exchange variables. Different lowercase letters above bars and lines denote significant differences in photosynthetic parameters among plant species within the same month (p < 0.05).

3.3. Characteristics of Stable Carbon Isotope Parameters in Typical Shrub Leaves

As shown in Figure 6, leaf δ¹³C_p ranged from −27.60 to −26.92‰ for Caragana microphylla, −28.66 to −27.05‰ for Salix psammophila, and −28.92 to −27.51‰ for Artemisia ordosica. The results indicate that during the growth peak season, the δ¹³C_p values ranked in descending order as Caragana microphylla (−27.21‰) > Salix psammophila (−27.80‰) > Artemisia ordosica (−28.48‰). Caragana microphylla reached its highest δ¹³C_p in July (−26.92‰); Salix psammophila also peaked in July (−27.25‰); Artemisia ordosica peaked in August (−28.15‰) during the growth peak season.

Figure 6. Carbon−isotope ratio characteristics of shrub leaves. Different lowercase letters denote significant differences in leaf carbon−isotope ratios (δ¹³C_p) among plant species within the same month (p < 0.05).

Δ¹³C of plant leaves (stable carbon-isotope fractionation) reflects the extent of isotopic fractionation in plant tissues relative to the atmosphere. As shown in Figure 7, in July, the ¹³C fractionations during CO₂ fixation were ranked as follows: Artemisia ordosica (19.32‰) > Salix psammophila (17.95‰) > Caragana microphylla (17.61‰). In August, Δ¹³C were ranked as follows: Artemisia ordosica (18.84‰) > Salix psammophila (18.70‰) > Caragana microphylla (17.82‰). In September, ¹³C fractionations were ranked as follows: Artemisia ordosica (19.36‰) > Salix psammophila (18.84‰) > Caragana microphylla (18.29‰). Accordingly, Caragana microphylla exhibited the lowest ¹³C fractionation throughout the growth peak season.

Figure 7. Isotopic resolution. Different lowercase letters denote significant differences in Δ¹³C among plant species within the same month (p < 0.05).

3.4. Typical Desert Shrubs’ Drought Resistance Characteristics

In July, MMP was 19 mm, MME and MMAT peaked at 179.9 mm and 24.3 °C, respectively, characterizing a dry month. As shown in Figure 8a, leaf δ¹³C_p in July was −26.92‰ (Caragana microphylla), −27.75‰ (Salix psammophila), and −28.63‰ (Artemisia ordosica). These values indicate δ¹³C_p enrichment consistent with high WUE across all three species. Stem δ¹⁸O values were −8.56‰ (Caragana microphylla), −8.89‰ (Salix psammophila), and −6.23‰ (Artemisia ordosica), implying that Caragana microphylla and Salix psammophila maintained higher stomatal conductance—likely via groundwater uptake through deep root systems [30,31,32]—whereas Artemisia ordosica showed lower conductance due to its shallow root distribution [33]. Overall, July drought tolerance ranked as: Caragana microphylla > Salix psammophila > Artemisia ordosica.

Figure 8. Monthly variation characteristics of stem oxygen isotopes and leaf carbon isotopes. (a) δ¹³C_p of leaves and δ¹⁸O of stems of three shrubs in July; (b) δ¹³C_p of leaves and δ¹⁸O of stems of three shrubs in August; (c) δ¹³C_p of leaves and δ¹⁸O of stems of three shrubs in July and August. Different lowercase letters below bars denote significant differences (p < 0.05) in leaf carbon-isotope ratios (δ¹³C_p) and stem oxygen-isotope ratios (δ¹⁸O) among plant species within the same month.

In August, MMR rose to 147 mm, MME and MMAT reached 135.6 mm and 21 °C, respectively, indicating relief from drought stress that month. MMRH increased to 70.8%. As shown in Figure 8b, leaf δ¹³C_p in August was −27.12‰ (Caragana microphylla), −28.01‰ (Salix psammophila), and −28.15‰ (Artemisia ordosica). For Salix psammophila, a relatively enriched δ¹³C_p is consistent with maintaining higher WUE, whereas a more negative δ¹³C_p signal has been associated with photosynthetic recovery under pulse precipitation [34]. The slightly more negative δ¹³C_p of Artemisia ordosica and its reduced WUE are consistent with the findings of Yan Jiang et al. [35]. Stem δ¹⁸O values for the three shrubs ranged from −6.03‰ to −6.42‰. For Salix psammophila, stem δ¹⁸O increased to −6.03‰, indicating elevated stomatal conductance and high water consumption after rainfall pulses [36]. Artemisia ordosica remained relatively stable [37], showing a slight decrease to −6.53‰ and becoming the species with the lowest stomatal conductance during the same period. Taken together, drought tolerance in August ranked as: Caragana microphylla > Salix psammophila > Artemisia ordosica.

In September, MMR declined to 104.4 mm, and MME and MMAT decreased to 78.7 mm and 15.8 °C, respectively, indicating moderate drought conditions. As shown in Figure 8c, Caragana microphylla maintained high WUE (−27.06‰) throughout September, whereas the WUE of Salix psammophila and Artemisia ordosica continued to decline while sustaining autumn photosynthetic carbon assimilation [38]. Stem δ¹⁸O values further increased, ranging from −4.59‰ to −5.81‰. Artemisia ordosica exhibited the highest value (−4.59‰) and the greatest stomatal conductance, indicating strong reliance on prior water reserves [39]. Salix psammophila showed a slightly lower value (−4.79‰), reflecting moderate regulation of stomatal conductance. Overall, September drought resistance ranked as: Caragana microphylla > Salix psammophila > Artemisia ordosica.

3.5. Water Use Efficiency Characteristics of Typical Sandy Shrubs

WUEi is a core indicator for assessing plant adaptation to arid environments in desert ecosystems [40]. Among-species comparisons of WUEi were highly significant (p ≤ 0.01). As shown in Figure 9, July WUEi were ranked as follows: Caragana microphylla (4.14 mmol·mol⁻¹) > Salix psammophila (2.97 mmol·mol⁻¹) > Artemisia ordosica (1.89 mmol·mol⁻¹), indicating a pronounced among-species disparity. This disparity may arise from Caragana microphylla’s extensive root system [41], which enables deeper soil-moisture uptake [42] and results in higher WUEi. Artemisia ordosica exhibited the lowest WUEi during this period, as its shallow root distribution hampers maintaining a balance between carbon gain and water loss under arid conditions [43]. In August, WUEi were ranked as follows: Caragana microphylla (5.50 mmol·mol⁻¹) > Salix psammophila (5.19 mmol·mol⁻¹) > Artemisia ordosica (4.87 mmol·mol⁻¹). This pattern suggests that increased regional precipitation partially alleviated drought stress, allowing WUEi across the three shrubs to increase more uniformly [44]. In September, the WUEi of the three shrubs reached their peak values, following the order of Salix psammophila (7.94 mmol·mol⁻¹) > Caragana microphylla (7.81 mmol·mol⁻¹) > Artemisia ordosica (5.65 mmol·mol⁻¹). Salix psammophila maintained efficient growth through photosynthesis with stable photosynthetic enzyme activities in September [45]. Overall, during the peak growth period, the WUEi of these three typical psammophytic shrubs showed the order of Caragana microphylla > Salix psammophila > Artemisia ordosica.

Figure 9. WUEi characteristics of shrubs. Different lowercase letters above bars denote significant differences in WUEi among plant species within the same month (p < 0.05).

WUE is a core ecophysiological indicator integrating photosynthetic carbon fixation and water loss, directly reflecting plant carbon–water coupling. As shown in Figure 10, overall WUE ranked as Caragana microphylla > Salix psammophila > Artemisia ordosica. During the growth peak season, Caragana microphylla and Salix psammophila declined month by month, reaching troughs in September with reductions of 7.24% and 9.89%, respectively. This pattern suggests decreased Rubisco activity, with a greater decline in net photosynthetic rate than in transpiration rate [46], reflecting high responsiveness to environmental change. Artemisia ordosica showed little WUE variation across the growth peak season, following a symmetric increase-decrease pattern and peaking at 90.68 μmol·mol⁻¹ in August.

Figure 10. WUE characteristics of shrubs.

4. Discussion

4.1. Physiological Response Differences Among Three Shrub Species During the Growth Peak Season

Plant community adaptation to environmental conditions drives adjustments in physiological function. This study shows that during the growth peak season (July–September), three desert shrubs expressed differentiated physiological responses under distinct water-use efficiency strategies [4]. Across the growth peak season, Pn increased month by month in Caragana microphylla, Salix psammophila, and Artemisia ordosica by 16.88%, 22.46%, and 23.74%, respectively, while Tr decreased (Figure 5). These dynamics indicate that as the growth peak season progresses, plants regulate stomatal behavior and photosynthetic metabolism to track changing environmental conditions. Temporal shifts in shrub photosynthetic parameters are tightly coupled to seasonal water availability [47]. Caragana microphylla maintained high WUE and WUEi throughout the growth peak season and showed the most stable Pn and Tr trajectories. This performance is consistent with smaller xylem vessel diameters and thicker walls that confer higher resistance to pit-related hydraulic failure [48], sustaining water transport through September. Mechanistically, guard cells precisely regulate stomatal conductance (gs) via abscisic acid (ABA) signaling, while Pn is maintained by stable photosynthetic enzyme activity in mesophyll cells [49]. These traits reflect strong buffering against environmental stress and a conservative water-use strategy [48]. By contrast, Pn in Salix psammophila surged under intense August precipitation, evidencing sensitivity to precipitation pulses and an opportunistic water-use strategy [48], thereby maximizing carbon acquisition under ample water [50]. Artemisia ordosica exhibited lower WUE and WUEi with the greatest Tr reduction, revealing a drought-vulnerable water-use strategy linked to shallow rooting. Its early-drought stomatal-closure efficiency was lower than in the other two shrubs. Although larger xylem vessels facilitate rapid post-rainfall transport, they entail a higher risk of pit-mediated failure. Short-term drought induces vessel plugging, and even after water conditions improve, hydraulic function recovers slowly [51]. The conservative strategy of Caragana microphylla, the opportunistic strategy of Salix psammophila, and the vulnerable strategy of Artemisia ordosica align with Schwinning’s predictive framework [52]. Thus, water-use strategies in sandy ecosystems are non-random; they represent adaptive optima shaped by precipitation regimes to maximize carbon gain.

4.2. Plant Water Source Allocation and Drought Resistance Traits

Plant water-use strategies in arid environments are tightly linked to drought resistance, with source-water selection at the core of these strategies. By comparing stem-water and depth-resolved soil-water δ¹⁸O and integrating these with WUE, this study reveals fundamental interspecific differences in source-water allocation and the resulting drought-resistance traits [20]. Caragana microphylla followed a pattern typical of deep-rooted, drought-tolerant shrubs [47]. Prioritizing stable deep-water sources underpinned sustained high Pn and WUE during drought, providing the physiological basis for its drought resistance. This also reflects a conservative, stable drought-tolerance pattern [17]. By contrast, Salix psammophila exhibits an opportunistic water-use pattern. In July, stem-water δ¹⁸O (−8.89‰) approximated deep-soil water, indicating drought tolerance. After increased August precipitation, stem-water δ¹⁸O shifted toward enrichment to −6.03‰—the least negative among the three species. This shift paralleled soil-water δ¹⁸O depletion across depths, indicating rapid root-system reallocation and efficient use of pulse precipitation throughout the profile [53]. This strategy enables rapid growth during wet periods but renders water status and WUE highly sensitive to precipitation variability. Consequently, its drought tolerance is strongly environment-dependent and entails a higher risk [18]. However, Huang L. reported that Salix psammophila primarily relies on deep-soil water and groundwater as long-term sources [54]. This distinction likely reflects differences in groundwater recharge conditions and topographic variation within the study area. The water-use pattern of Artemisia ordosica reveals constraints inherent to shallow-water reliance. Throughout the growth peak season, its stem-water δ¹⁸O was the most enriched (least negative) among the three species—July (−6.24‰), August (−6.53‰), and September (−4.59‰). This pattern aligns with the δ¹⁸O signature of surface soil water, which is strongly enriched by evaporation and preferentially taken up. This constrained source-water allocation imposed severe drought stress during dry periods, resulting in persistently low WUE. The highest stem-water δ¹⁸O in September confirms heightened reliance on shallow soil water at the end of the growth peak season, further diminishing drought tolerance. This conclusion aligns with observations from the Tengger Desert reported by [54].

4.3. Plant Stomatal Control and Carbon Balance

In arid environments, plant survival hinges on balancing carbon fixation with water loss, a process primarily regulated by stomatal behavior [55]. Leaf δ¹³C_p, which integrates long-term variation in stomatal conductance and photosynthetic capacity, is a reliable indicator of carbon balance [49]. Across the growth peak season, leaf δ¹³C_p ranked as follows: Caragana microphylla (−27.21‰) > Salix psammophila (−27.80‰) > Artemisia ordosica (−28.48‰). This ordering corresponds to the interspecific ranking of WUE [Jan Binter]. According to the Farquhar model, a more negative δ¹³C_p indicates a higher C_i/C_a ratio at the leaf scale [56]. Such depletion typically reflects more open stomata or reduced carboxylation capacity, leading to stronger Δ¹³C and lower WUE. This point has been verified in the research of this article (Figure 7 and Figure 10). Nevertheless, under extreme stress, δ¹³C_p can yield biased inferences about stomatal control. Under severe drought or heat that markedly impairs Rubisco activity [57], the dominant constraint on photosynthesis shifts from stomatal to non-stomatal limitation. This shift causes an irreversible decline in Pn, whereas WUE fails to increase as expected. In this study, Pn in all three shrubs remained elevated and peaked in September. These patterns indicate that elevated δ¹³C_p together with high WUE primarily reflects efficient stomatal regulation that limits water loss while sustaining carbon fixation, rather than impaired photosynthetic function.

4.4. Effects of Meteorological Factors on Shrub Photosynthetic Parameters and Stable Isotopes

Principal Component Analysis (PCA) clarifies how environmental drivers intersect with species-intrinsic strategies to shape plant physiological patterns. As shown in Figure 11, PC1–PC3 together explained 95.08% of the total variance. Individually, PC1 = 59.43%, PC2 = 32.70%, and PC3 = 2.95%. PC1 captures the common physiological response of all species to synergistic water–heat dynamics during the growth peak season. Its gradient tracks the environmental continuum from July drought stress to improved moisture in September across the study area. This pattern aligns with Wang et al. [58], which emphasizes regional climate control over physiological rhythms and highlights the strong constraint of large-scale moisture regimes on physiological activity in sandy ecosystems. PC2 represents intrinsic differentiation in species’ water-use strategies. The three shrubs showed consistent and temporally stable positions along PC2. Caragana microphylla consistently occupied the positive end of PC2, maintaining stable Pn, elevated δ¹³C_p, and high WUE across the growth peak season. This configuration is consistent with a conservative strategy that combines deep-rooted access to stable water with efficient stomatal regulation [59]. However, the PC2 scores of Salix psammophila fluctuated markedly—especially after precipitation pulses—indicating a shift that prioritized growth over sustained WUE enhancement, a self-regulatory pattern consistent with an opportunistic strategy [59]. Artemisia ordosica consistently clustered at the negative end of PC2 and, in the dry month of July, simultaneously occupied low PC1 and low PC2 positions. This pattern corroborates vulnerability arising from shallow-water dependence and passive stomatal regulation, consistent with its narrow ecological niche [60].

Figure 11. PCA of meteorological factors, plant photosynthetic parameters, and carbon−oxygen isotopes.

This study elucidated multiscale regulatory mechanisms of shrub water-use strategies using Mantel correlations among meteorological factors, gas-exchange parameters, and carbon–oxygen isotopes (Figure 12). Leaf δ¹³Cp correlated positively with WUE, whereas Δ¹³C correlated negatively with WUE, collectively indicating that carbon-isotope fractionation integrates long-term variation in intercellular CO₂ concentration (C_i). According to the Farquhar model, enrichment in δ¹³C_p directly reflects a reduction in the C_i/C_a ratio during the growth peak season. These results indicate that the three shrubs achieved high WUE primarily via tighter stomatal regulation that lowered C_i/C_a, rather than via increased biochemical capacity [56]. The negative correlation between WUEi and Tr indicates that Tr is the dominant driver of short-term fluctuations in WUEi. The observed negative Pn–Tr correlation further supports this inference. Physiologically, the divergence between WUE and WUEi is clarified: δ¹³C_p integrates stomatal effects on CO₂ supply, whereas WUEi reflects transient stomatal regulation of water loss under short-term environmental variation. δ¹⁸O correlated positively with WUEi and negatively with Tr, indirectly corroborating stomatal regulation of water-transport processes. These findings align with the conclusions in Section 4.3.

Figure 12. Mantel-test correlations among meteorological factors, plant photosynthetic parameters, and carbon−oxygen isotopes. The size and labels of the squares, together with the line colors, represent Spearman’s correlation coefficients. Asterisks denote significance levels: *** p < 0.001, ** p < 0.01, * p < 0.05. Dotted lines indicate negative correlations, whereas solid lines indicate positive correlations.

Analysis of environmental drivers further clarifies the mechanisms above. The positive SWC−WUEi correlation indicates that sufficient soil moisture supports maintaining high Pn. The negative MMAT−WUEi correlation suggests that, at comparable moisture, higher temperature elevates vapor-pressure deficit (VPD), inducing partial stomatal closure and reducing water loss [61]. MMRH correlates positively with WUEi and negatively with Tr, indicating that atmospheric humidity modulates VPD and thereby contributes to multiscale regulation of water-use efficiency. Together, month-to-month increases in SWC, decreases in MMAT, and rises in MMRH underpinned the July–September increases in WUE and WUEi across the study area. Among these drivers, SWC was dominant, whereas MMAT and MMRH acted as short-term modulators that amplified or dampened the effects of stomatal regulation on carbon balance and water-use.

4.5. Limitations and Future Perspectives

This study jointly measured leaf gas-exchange parameters, leaf δ¹³C_p, and δ¹⁸O of stem water and soil water, and related these indicators to key environmental drivers, thereby robustly discriminating the water-use strategies of three xerophytic shrubs. This integrative approach overcomes single-indicator bias and provides a rigorous basis for interpreting the physiological and ecological mechanisms underlying shrub adaptation to arid environments. Nevertheless, several limitations remain: sampling was confined to a single growth peak season. While this captures strategy contrasts during a critical window, it does not encompass interannual climate variability or responses across phenological stages. In addition, we did not directly quantify key physiological signals and metabolites, including abscisic acid (ABA), reactive oxygen species (ROS), and proline, thereby limiting inference on the proximal regulation of stomata and osmotic adjustment. Future work should expand the indicator set and implement long-term in situ, fixed-site observations in the study area to assess generalizability, deepen mechanistic understanding of strategy formation from physiological and biochemical perspectives, improve the predictive skill of vegetation models, and inform climate-adaptive species selection for ecological restoration.

5. Conclusions

Focusing on the northern edge of the Mu Us Sandy Land, this study investigates three typical xerophytic shrubs—Caragana microphylla, Salix psammophila, and Artemisia ordosica. During the growth peak season, we quantified gas-exchange fluxes (Pn, Tr), leaf δ¹³C_p, and δ¹⁸O of stem and soil water, and evaluated their coupling with hydrometeorological (water–heat) factors. Our objective is to resolve the mechanistic basis of shrub water-use efficiency under divergent water-use strategies, guiding the optimization of species composition and selecting vegetation-restoration models in arid and semi-arid regions. The conclusion is as follows:

(1)

During the growth peak season (July–September), Pn increased month by month across all three desert shrubs, whereas Tr decreased. Both WUE and WUEi showed overall upward trends. Leaf δ¹³C_p was significantly positively correlated with WUE, whereas Δ¹³C was significantly negatively correlated with WUE. Stem-water δ¹⁸O was significantly positively correlated with WUEi and significantly negatively correlated with Tr. Together, these patterns indicate that δ¹³C_p and δ¹⁸O effectively diagnose the regulation of shrub water-use efficiency via C_i/C_a dynamics and water-transport pathways.

(2)

Three shrub species show significant differences in water source utilization and drought resistance strategies: Caragana microphylla relies on stable deep-water sources and, via efficient stomatal regulation, maintains the highest and most stable WUE and WUEi, exemplifying a conservative water-use strategy. Salix psammophila is highly responsive to precipitation pulses, rapidly elevating Pn and exploiting newly recharged water across the soil profile during wet periods, consistent with an opportunistic water-use strategy. Artemisia ordosica depends long-term on evaporatively enriched shallow-water sources, exhibits low average WUE, and shows persistently enriched stem-water δ¹⁸O during late-drought stages, revealing drought vulnerability driven by shallow rooting and passive stomatal regulation.

(3)

The combined influence of meteorological factors governs the temporal dynamics of shrub water-use strategies. SWC shows a significant positive correlation with WUEi. MMAT is significantly negatively correlated, whereas MMRH is significantly positively correlated with WUEi, indicating that temperature and atmospheric humidity modulate water-use efficiency over short timescales by regulating VPD and stomatal conductance, thereby underpinning the month-to-month increase in WUE and WUEi from July to September.

This study demonstrates that shrub water-use strategies in sandy regions form a complementary portfolio of conservative, opportunistic, and vulnerable types rather than a random assemblage. This conclusion provides guidance for species selection and configuration in ecological restoration in arid areas, prioritizing conservative species with high water use efficiency and reliable access to deep water sources as the foundation. It judiciously combines opportunistic species that respond rapidly to precipitation pulses and deploys shallow-water-dependent, drought-vulnerable species with caution. Such a combination is expected to enhance the efficiency of large-scale water use in arid regions and strengthen the stability of ecosystems.