Divergence in plant water relations mediates the clear niche differentiation of two congeneric desert shrub species

Abstract

Desert plants, despite overall well adapted to drought, exhibit considerable inter-specific differentiation of micro-habitats and variation in adaptive characteristics in the generally much water-limited yet spatially heterogeneous water regimes. Investigating the interplay between fine-scale spacial variation of environmental water regime and inter-specific differences in water-related physiology can provide an important basis for understanding the functional diversity and co-existence of desert plants. Here, we addressed this question by taking advantage of the remarkable niche differentiation of two dominant shrub species, Haloxylon ammodendron occupying the inter-dune lowlands and H. persicum colonizing the sand dunes in Gurbantunggut Desert of Northwest China. We precisely quantified the locations of each individual of the two congeneric species in transects established across the inter-dune lowland to dune crest landscape, measured soil water contents at different transect positions, and investigated water-related physiological traits of each individual. We found that the clear niche differentiation between the two sister species was paralleled by divergences in a suite of water-related functional traits. Particularly, such divergence indicates a clear trade-off between efficient water acquisition and drought tolerance. Higher hydraulic efficiency in H. persicum contributed to greater competitiveness in the sand dunes with relatively high soil water availability. In contrast, greater hydraulic safety and higher tissue water storage capacity in H. ammodendron contributed to its endurance in the more drought stressed inter-dune lowlands. Our findings indicate that variation of fine-scale soil moisture due to local factors like topography is a critical driver in shaping the species diversity in the heterogeneous environments of deserts.

Highlights

• •Two congeneric desert shrub (Haloxylon) species show clear niche differentiation.
• •H. ammodendron occupying inter-dune lowlands exhibited higher hydraulic safety.
• •H. persicum distributed on sand dune slopes possessed greater hydraulic efficiency.
• •Fine-scale variation in soil water regime contributes to plant diversity in deserts.

1. Introduction

Shrubs are often foundation species of deserts and play vital roles in desert ecosystems by stabilizing the sand dunes, protecting land from wind erosion, as well as providing essential shelter and food for desert wildlife (Zhao et al., 2007; Chaneton et al., 2010; Zhang et al., 2024). Although desert shrub species are overall well adapted to drought-stressed environments, they still exhibit considerable interspecific variations in adaptive characteristics in the generally much water-limited yet spatially heterogeneous and highly dynamic water regimes of deserts (McDowell et al., 2008; Li et al., 2016b). Such interspecific variations in shrub environmental adaptation can be vividly reflected by their non-random association with different micro-environments of the deserts, such as preferences to sand dunes of different developmental stages, different positions of the dunes or inter-dune lowlands, or habitats with different soil properties (Li et al., 2002; Zhang et al., 2015). Among the various drivers causing niche differentiation of shrubs in the heterogeneous environments of deserts, the temporal–spatial change of water regimes have likely played a core role (Li et al., 2007; Hao et al., 2013; Gong et al., 2020; Guo et al., 2020; Ostria-Gallardo et al., 2023). Investigating the interplay between shrub adaptation in terms of water relations and the temporal–spatial variation of environmental water availability can contribute to revealing the mechanisms underlying the patterns of niche differentiation of different desert shrub species, and hence provide an important basis for their conservation as well as utilization in vegetation restoration of desertified lands.

The xylem hydraulics of vascular plants reflects their adaptation to different water regimes and the diversification of hydraulic characteristics has been found to be an important mechanism underlying the niche differentiation of different tree and shrub species along environmental water gradients at different scales (Hao et al., 2013; Zhang et al., 2021; Rutherford and Archer, 2023). Species adapted to habitats with higher water availability tend to have more efficient xylem systems for water delivery, which favors higher rates of leaf gas exchange and thus higher growth potential for enhanced competitiveness (De Micco et al., 2008; Olano et al., 2017). In contrast, plants inhabiting water-limited habitats usually have xylem systems of low hydraulic efficiency but greater resistance to drought-induced hydraulic dysfunction, which contribute to plants' persistence in environments with strong drought stress (Brodribb et al., 2010; Gauthey et al., 2024). Differentiation in xylem hydraulics between species adapted to more mesic and more xeric environments have contributed to the relative advantages in their respective habitats, which is underlain by a trade-off between xylem hydraulic efficiency and resistance to drought-induced embolism (Pratt et al., 2021; Yao et al., 2021). Such an efficiency-safety trade-off is resulted from a compromise in xylem construction, i.e. xylem structural characteristics at both the tissue and pit levels that enhance water permeability are often at the cost of reduced resistance to drought-induced embolism, and vice versa (Wheeler et al., 2005; Sperry et al., 2006; Hacke et al., 2007). Considering the crucial role of xylem hydraulics in affecting plants’ adaptation to environments of different water availability, inter-specific differentiation in xylem hydraulics has likely contributed to the sorting of different shrub species into various microhabitats of the heterogeneous desert environments in terms of water regimes at fine scales. Knowledge in this aspect is scarce but is much in need for the conservation of desert shrubs in their fragile ecosystems and in informing their usage in revegetation projects for desertification control.

Besides enhancing hydraulic safety, leaf structural and physiological adjustments in favor of water conservation, water storage and tolerance to low water potentials are also of crucial importance for surviving drought-stressed environments (Su et al., 2009). Leaf structure determines functions including photosynthetic carbon assimilation, transpiration and water storage, which represents a compromise in balancing competing requirements of leaf functioning (Rhizopoulou and Psaras, 2003; Li et al., 2021; Yang et al., 2021). Plants in arid environments usually have small leaves to reduce water loss and increased leaf tissue thickness or density to resist dehydration (Scoffoni et al., 2011; Liu and Zheng, 2024). Some desert plants have been found to be able to maintain tissue turgor and biological activity under water potentials as low as −8 MPa (Meinzer et al., 1986; Ogle and Reynolds, 2002). Many desert plants have evolved succulent leaves or stems, which represent the strategy of enhancing internal water storage to survive periods with limited water access (Urs and Reto, 2009; Guo et al., 2024). Leaves of some desert shrubs have been extremely degraded and the photosynthetic function is instead performed by a special structure called assimilative branch, which reflects the functional integration and optimization of these shrubs under extreme conditions (Meng et al., 2022, 2024; Zhou et al., 2022; Ávila-Lovera et al., 2020; Yang et al., 2022b). Understanding the adaptive significance of structural and functional variation of leaves or assimilative branches of desert shrubs along environmental gradients can also be critically important in revealing the mechanisms of desert shrubs in adapting to different microhabitats.

Haloxylon ammodendron (C.A. Mey.) Bunge and H. persicum Bunge ex Boiss. et Buhse are two sister taxa shrub species commonly found in the inland deserts of Asia, which both have important ecological functions of sand dune fixation and prevention of wind erosion (Dai et al., 2015; Li et al., 2023). Particularly, H. ammodendron is widely used in projects of desertified land revegetation in arid areas of Northwest China. In Gurbantunggut Desert, the second largest desert of China, natural distribution of the two shrub species of the same genus shows spectacularly clear segregation of microhabitats, i.e. the H. ammodendron plants are concentrated in the inter-dune lowlands while the H. persicum plants scatter over the upper parts of the sand dunes (Liu et al., 2022). It has been proposed that the differences in seed salinity tolerance and seeding tolerance to sand burial have contributed to the niche differentiation of the two shrub species. Compared with H. persicum, the seeds of H. ammodendron are more tolerant to salinity, which facilitates its occupation of the saline inter-dune lowlands (Tobe et al., 2000; Song et al., 2005; Dai et al., 2015). Despite of lower seed tolerance to salinity, H. persicum seedlings exhibit higher root and shoot elongation rates after seed germination and show higher survival rate in face of sand burial compared to H. ammodendron (Liu et al., 2010; Li et al., 2011). Although the divergences in adaptive characteristics during the early stages of plant development have likely contributed to their niche differentiation, it may not be the whole story and other aspects of adaptation may also be involved and deserve further investigation.

The sand dunes and inter-dune lowlands in deserts have been observed to differ substantially in abiotic and biotic conditions (e.g. vegetation coverage, soil texture, salinity of soil water, groundwater depth), which can lead to large difference in soil water regimes (Sperry and Hacke, 2002; An et al., 2015; Zhao et al., 2023). For example, the inter-dune lowlands usually have more compact surface soil that is less permeable to rainfall infiltration, which may lead to more extreme low soil water potentials during the long rainfall intervals (Rosenthal et al., 2005; Huang, 2017). The inter-dune lowlands may also differ greatly in biological soil crust coverage from the sand dunes, which may further affect rainfall infiltration as well as soil water evaporation and hence contribute to the divergence in soil water regime between the two habitats (Fick et al., 2019; Guida et al., 2023). Considering the potential differences in soil water regimes between inter-dune lowlands and sand dunes, we proposed that the divergence in characteristics related to xylem water transport and plant water management are coupled to the clear niche differentiation of the two Haloxylon species. Here, we quantified the niche differentiation of the two shrub species in transects that established across the inter-dune lowland and dune crest landscape perpendicular to the contour lines. We measured soil water contents, soil crust coverage, and soil crust thickness at different positions of the transects. Simultaneously, we measured stem xylem hydraulic traits, structural and physiological traits of assimilative branches of individuals of the two shrub species at different positions of the transects. Specifically, we tested the following hypotheses: (1) Compared to H. ammodendron that grow on the inter-dune lowlands, H. persicum plants growing on the sand dunes would show better water status due to more favorable soil water regime; (2) H. ammodendron has lower stem hydraulic conductivity but higher resistance to drought-induced embolism compared to H. persicum, reflecting a trade-off between efficiency and safety of xylem water transport; (3) The assimilative branches of H. ammodendron would have higher water storage capacity and physiological characteristics of enhanced drought tolerance than H. persicum.

2. Materials and methods

2.1. Study site and plant materials

The study site was located in the southern edge of Gurbantunggut Desert, Xinjiang Uygur Autonomous Region, Northwest China (87°92′E, 44°40′N). This area is characterized by a typical temperate continental arid climate with a hot and dry summer from June to August and a cold winter with snow coverage from December to March (Fig. S1). The average annual temperature of this region is 6.6 °C and annual precipitation ranged from 70 to 180 mm with relatively large inter-annual variation (Xu et al., 2007). The annual potential evaporation is over 2000 mm. The prolonged snow cover with maximum depth about 20 cm is one of the most distinctive characteristics of this desert and the snowpack can last 100–150 days in winter. The melting snow in April constitute a significant portion of the water source for the desert vegetation (Zhou et al., 2009).

Most of the sand dunes in Gurbantunggut Desert are fixed and semi-fixed, and the dominant shrub species distributed along the sand dune slopes were Haloxylon ammodendron (C.A.Mey.) Bunge, Haloxylon persicum Bunge ex Boiss. et Buhse, Calligonum leucocladum (Schrenk) Bunge and Ephedra distachya L. These shrub species usually have clear habitat preferences in the desert due to their specific environmental adaptations. H. ammodendron and H. persicum are two sister species from the same genus and dominant species of the typical plant communities in the Gurbantunggut Desert. A miraculous phenominon in the Haloxylon shrub communities is that the two sister species show clear niche differentiation (Fig. 1), with H. ammodendron mostly growing in the inter-dune lowlands and the bottom of the windward sand dune slopes and H. persicum mostly occupying the middle and upper part of the windward sand dune slopes. Both species are also the dominant xerophytic shrubs in Central Asia deserts. H. ammodendron is widely distributed from Central Asia and Mongolia to northern China, while H. persicum has a broader distribution range from North Africa and the Middle East to Central Asia, with its main area of distribution in China situated at Junggar Basin of Xinjiang. These two species show significant overlap in distribution in Central Asia (Buras et al., 2012; Li et al., 2019).

Fig. 1.

(a) The diagram showing a sampling transect along the topography of sand dune slope with Haloxylonammodendron and H. persicum occurring at lower and higher positions, respectively. The black dots represent the points for measuring soil mass water content. (b) The frequency distribution of the two species along the topography showing the total number of all individuals within a certain range of distance to inter-dune lowland (DDL). Data for the three transects were pooled together.

2.2. Quantification of species distribution, soil water content and soil crust characteristics along the sand dune slopes

Three transects of 20 m in width and ca. 90 m in length from the inter-dune lowland to the top of the sand dunes were established at a typical shrub community mainly composed of Haloxylon ammodendron and H. persicum (Fig. 1a). The positions of all individuals of these two shrub species occurring in the three transects were determined using a handheld GPS positioner (Model P20, Zhuolin Electronic Science and Technology Company, Anhui Province, China, with a vertical precision of ±0.5 m). The relative elevation of distribution in the transects for each shrub measured at the base of the plant was quantified as its vertical distance to the inter-dune lowland (DDL):

DDL = A_i – A_l

where A_i indicates the elevation of the base of a specific shrub individual, and A_l is the elevation of the inter-dune lowland bottom at the study site. All individuals with basal diameter ≥4 cm were surveyed and later on sampled for physiological measurements. The diameter at ground height (DGH, cm), the plant height (H, m) and crown area (CA, m²) of all individuals were also measured (Fig. S2). The crown area of shrubs were viewed as ellipses in a two-dimensional projected area, the geometric mean of major and minor axis widths were used as diameters (Montès et al., 2000).

Soil samples were taken at three representative points (inter-dune lowland, dune slope and dune crest) in each transect to measure water contents at different depths (Fig. 1a). Measurements were done three times from late Autumn 2023 through to Spring 2024, i.e. late November 2023 before snowfall, mid April 2024 when snow has just melted, and late April. At each sampling point, soil samples were collected using a boring auger at every 10 cm depth intervals from 0 to 80 cm depth and sealed immediately in individual aluminium boxes. The fresh mass of soil samples was determined using an electronic balance with a precision of 0.0001 g, then oven-dried under 105 °C for 48 h to constant mass. Soil mass water content (θ_m) was calculated as follows:

$${\mathrm{\theta }}_m=\left(m_1-m_2\right)/\left(m_2-m_0\right)\times 100\%\text{,}$$

where m₀ is the mass of each aluminum box, m₁ and m₂ are the mass of the aluminum box containing the fresh and dried soil samples, respectively. Moreover, soil crusts were identified as biological soil crusts (BSCs) primarily composed of cyanobacteria, lichens and mosses based on surface coloration and morphology (Lan et al., 2013). Soil crust coverage was surveyed in 20 plots (1 × 1 m in size) at inter-dune lowland, dune slope and dune crest in each of the three transects (Fig. S3) and the high-resolution photos captured from the sampling plots were analyzed to calculated the proportion of coverage by using ImageJ software (Wayne Rasband, National Institute of Health). Soil crust thickness was measured following the criteria by Li et al. (2002) in five randomly selected points avoiding the coverage by shrubs and weeds at each of the three locations of each transect, i.e. inter-dune lowland, the middle part of the dune, and the top of the dune. Soil samples (0–5 cm depth) including crust and topsoil were collected and the crust thickness was also calculated based on photos of soil profile by using ImageJ software.

2.3. Stem hydraulic traits and related characteristics

One branch of approximately 1.3 m in length from each of the surveyed individuals of Haloxylon ammodendron and H. persicum in the three transects were collected in August 20 and 21, 2023. Samples were collected before dawn and after excision the branches were re-cut immediately under water in a bucket and transported to the laboratory with the cutting ends submerged in water and the whole branches wrapped in black plastic bags. Stems of approximately 20 cm in length and 1 cm in diameter were obtained from different individuals of each species with both ends trimmed using a razor blade under water for measuring the stem hydraulic conductivity (K_h, kg m s⁻¹ MPa⁻¹) with a tubing apparatus filled with 20 mM KCl solution (Niu et al., 2017). Afterward, stem segments were flushed with degassed KCl solution under a positive pressure of 0.1 MPa to refill embolized vessels and then maximum stem hydraulic conductivity (K_max, kg m s⁻¹ MPa⁻¹) was measured. Then the percent loss of conductivity (PLC, %) was calculated as:

$$\text{PLC}=\left(K_{\mathit{max}}-K_h\right)/K_{\mathit{max}}\times 100\%\text{,}$$

The sapwood area (SA) of each stem segment was determined using the dye-staining method on two short segments cut from both ends of the segment. All the assimilative branches terminal to the stem segment were scanned and the total projected area (A_ab) was calculated using ImageJ software (Wayne Rasband, National Institutes of Health). The sapwood specific conductivity (K_s, kg m⁻¹ s⁻¹ MPa⁻¹) and assimilative branch area specific conductivity (K_ab, kg m⁻¹ s⁻¹ MPa⁻¹) were calculated as K_h divided by SA and A_ab, respectively. Huber value (Hv, m² m⁻²) was calculated as SA of a stem segment divided by A_ab of all the terminal assimilative branches. Wood density (WD, g cm⁻³) was determined by the water displacement method.

2.4. Stem xylem light microscopy

In August 2023, a short stem segment (10–15 mm in diameter) were collected from each of the surveyed individuals of both shrub species within every transect for xylem anatomical measurements. Stems were cut into segments approximately 0.5 cm in length and fixed in 70% FAA solution before they were embedded in paraffin. The embedded samples were then cut into slices of 8–15 μm in thickness using a microtome (Leica Rm2016, Leica Biosystems, Germany), stained with 1% safranin solution and made into permanent slides by sealing with neutral balsam. Xylem cross-sections were examined under an optical microscope (Olympus BX40, Olympus Corporation, Japan), and images were taken under magnifications of ×40, ×100 and ×400 to produce integral cross-sectional images. The vessel wall thickness (T_VW, μm) and fiber wall thickness (T_FW, μm) were measured using ImageJ software. Vessel diameter (D_V, μm) of each individual was estimated based on the lumen area of all vessels in the analyzed images, assuming a circular shape of vessels. The vessel thickness to span ratio (T_VW/D_V)², an indicator of resistance to drought-induced embolism, was calculated following the method described by Hacke et al. (2001).

2.5. Measurement of water potentials and construction of pressure-volume curves for assimilative branches

The predawn (Ψ_pd) and midday (Ψ_md) water potentials for assimilative branches of H. ammodendron and H. persicum were measured on two consecutive sunny days in August 2023 using a pressure chamber (Model 1505D, PMS Instrument, Albany, USA). Small twigs for Ψ_pd and Ψ_md measurements were excised during 6:00–8:00 and 13:00–14:00 (solar time), respectively. Samples were immediately sealed in small individual plastic bags containing moist paper towels in order to prevent water loss. All water potential measurements were completed within 2 h of excision.

The pressure-volume (PV) relations were determined on three consecutive days in late August 2023. One branch from each individual of the two species in the three transects (a total of 36 branches for each species) were sampled and allowed to rehydrate for ca. 1 h in the dark prior to measurements. After rehydration, small twigs with assimilative branches attached were taken for determination of PV relations. Paired measurements of masses and water potentials were done periodically for each of the twigs following the natural dehydration process. Measurements were stopped until branches show obvious symptoms of damage due to dehydration. Finally, the dry mass of the twigs was measured after oven-dried at 70 °C for 72 h. All the mass measurements were made using a 4-digit electronic balance (BSA124S, Satorius Inc) and water potential measurements were done using a PMS pressure chamber. Osmotic potentials at full turgor (π¹⁰⁰, MPa), turgor loss point (π⁰, MPa), water capacity (C, g MPa⁻¹) and modulus of elasticity (ε, MPa) of assimilative branches were calculated using the methods of Nguyen et al. (2017). The diurnal water potential change for assimilative branch (ΔΨ, MPa) was calculated as the difference between Ψ_pd and Ψ_md. The safety margin for assimilative branch (ABSM, MPa) was defined as the difference between Ψ_md and π⁰, according to Ziegler et al. (2019).

2.6. Measurements of structural and physiological traits of assimilative branches

The assimilative branch thickness (ABT, cm) and length were measured using the same set of branches sampled for the hydraulic conductivity measurements. The masses of saturated (F_ab, g) and oven-dried (D_ab, g) assimilative branches were measured with a precision of 0.0001 g using the analytical balance. Relative water content (RWC, %) was calculated as:

$$\text{RWC}=\left(F_{ab}-D_{ab}\right)/F_{ab}\times 100\%$$

Mass per projected area (MPA, g cm⁻²) of the assimilative branches was calculated as dry mass divided by their projected area. Tissue density of the assimilative branches (ABTD, g cm⁻³) was calculated as dry mass divided by volume, with an assimilative branch assumed as a long cylinder and its volume calculated as the product of the base area and length. Proline content in the assimilative branch (Pro, μg g⁻¹) was determined according to the method described by Bates et al. (1973). First, an extract was prepared using sulfosalicylic acid and the grind solution of 0.2 g fresh assimilative branches. Then, extract was reacted with acidified ninhydrin solution while being heated at 100 °C for 10 min, and the proline concentration was calculated based on absorbance obtained spectrophotometrically at 520 nm using a UV/VIS spectrophotometer (Model Lambda 35, PerkinElmer, USA).

2.7. Statistical analysis

All data were tested for normality and homogeneity of variance. The Hv, ABTD, RWC were log-transformed and ABT was ln-transformed before analysis to ensure the homogeneity of data variance, respectively. After analysis of variance (ANOVA), least significant difference (LSD) test was conducted to compare the differences in soil water content, soil crust coverage and thickness between different positions along the sand dune slope. The correlations between all pairwise functional traits were evaluated by using Pearson's correlation analysis (Table 1). Principal component analysis (PCA) was conducted to examine the multivariate relationships among 22 key functional traits using all the individuals of the two species. All analyses were conducted using R (v.4.5.0).

Table 1.

List of functional traits measured for the two studied shrub species.

Acronyms	Definition	Units
Hydraulic related traits of stems
Ks	Sapwood specific conductivity	kg m−1 s−1 MPa−1
Kab	Assimilative branch specific conductivity	× 10−4 kg m−1 s−1 MPa−1
Hv	Huber value	× 10−4 m2 m−2
PLC	Percent loss of conductivity	%
WD	Wood density	g cm−3
Stem xylem anatomical traits
TVW	Vessel wall thickness	μm
TFW	Fiber wall thickness	μm
DV	Vessel diameter	μm
(TVW/DV)2	Vessel wall thickness to span ratio	–
Physiological traits of assimilative branches in plant-water relations
Ψpd	Predawn water potential	MPa
Ψmd	Midday water potential	MPa
ΔΨ	Diurnal water potential change	MPa
ABSM	Safety margin for assimilative branch	MPa
π0	Osmotic potential at turgor loss	MPa
π100	Osmotic potential at full turgor	MPa
C	Water capacity	g MPa−1
ε	Modulus of elasticity	MPa
Pro	Proline content	μg g−1
Traits of assimilative branches reflecting degree of succulence
MPA	Mass per projected area	g m−2
ABT	Assimilative branch thickness	cm
ABTD	Assimilative branch tissue density	g cm−3
RWC	Relative water content at full saturation	%

3. Results

Our results showed that Haloxylon ammodendron and H. persicum distributed in the lower and higher positions of the desert landscape differed in plant size and diverged significantly in a suite of water-related functional traits, with the former shrub species showing characteristics adaptive to environmental conditions of stronger drought stress (Table 2). H. persicum occupying the sand dunes on average had significantly larger basal diameter and plant height (P < 0.05, one-way ANOVA) and marginally significant larger crown size (P = 0.1, one-way ANOVA) than H. ammodendron distributed in the inter-dune lowlands. With the increase of vertical distance to the inter-dune lowland (DDL), plant size measures all exhibited a trend of significant increase (Fig. S2a–c, P < 0.05, linear regression).

Table 2.

Key functional traits of the two studied shrub species (Haloxylonammodendron and H. persicum). Data show species mean values ± 1 SE. CV, coefficient of variation (%). P values indicate level of statistical significance (t-test). Trait name abbreviations are as defined in Table 1.

	H. ammodendron	CV	H. persicum	CV	P value
Ks	0.14 ± 0.02	64.20%	0.21 ± 0.02	59.36%	0.01
Kab	1.19 ± 0.16	82.69%	2.23 ± 0.19	51.40%	<0.001
Hv	3.11 ± 0.02	4.64%	11.78 ± 0.78	39.62%	<0.001
PLC	57.73 ± 3.33	34.59%	52.82 ± 2.84	32.30%	0.27
WD	0.89 ± 0.01	5.55%	0.88 ± 0.01	5.54%	0.27
TVW	4.77 ± 0.07	8.83%	3.50 ± 0.05	9.16%	<0.001
TFW	1.66 ± 0.01	4.86%	1.51 ± 0.03	11.14%	<0.001
DV	16.27 ± 0.38	14.11%	18.75 ± 0.48	15.50%	<0.001
(TVW/DV)2	0.10 ± 0.01	34.31%	0.04 ± 0.01	31.92%	<0.001
Ψpd	−3.40 ± 0.06	12.08%	−3.32 ± 0.06	11.55%	0.41
Ψmd	−5.28 ± 0.07	7.50%	−5.05 ± 0.07	8.13%	0.02
π0	−5.80 ± 0.06	6.32%	−5.70 ± 0.07	7.66%	0.29
ΔΨ	1.89 ± 0.08	25.09%	1.73 ± 0.08	26.02%	0.16
ABSM	0.52 ± 0.09	103.23%	0.65 ± 0.08	78.43%	0.29
π100	−4.09 ± 0.07	10.60%	−3.96 ± 0.11	17.15%	0.32
C	4.41 ± 0.27	37.36%	2.08 ± 0.14	41.19%	<0.001
ε	0.44 ± 0.04	57.70%	0.51 ± 0.04	46.42%	0.25
Pro	37.16 ± 2.12	34.23%	60.34 ± 2.75	27.34%	<0.001
MPA	380.25 ± 15.13	23.87%	477.10 ± 16.96	21.33%	<0.001
ABT	1.58 ± 0.03	10.89%	1.04 ± 0.02	9.50%	<0.001
ABTD	0.29 ± 0.01	13.14%	0.59 ± 0.01	14.04%	<0.001
RWC	69.13 ± 0.50	4.36%	46.22 ± 1.07	13.86%	<0.001

Measurement of soil water contents of different depths on the dune crest, middle part of the dune slope, and the inter-dune lowland showed that soil water regime differed substantially at different positions of the transects. In late November of 2023, the end of the growing season, the surface soil layer (0–10 cm in depth) showed relatively high water content and did not differ among the three positions (Fig. 2a). Soil water content consistently decreased with the increase of depth and in soil layers of 20–40 cm depth the dune crest had substantially higher water content than those of the dune slope and inter-dune lowland, although some of the differences were not statistically significant due to the small number of replicates (Fig. 2a). In mid April of 2024, when snow melting had just completed, sand layers at 0–20 cm depth of the inter-dune lowland had the highest water content among the three positions, while the dune crest had the highest sand water content at deeper (40–60 cm) layers (Fig. 2b). In late April, as the water from snow melting infiltrated to deeper layers and evaporated from soil surface, water contents in the top soil layers (0–60 cm) became considerably low at all the three positions of the transects. At the deepest layers sampled (70–80 cm), the dune crest showed the highest sand water content but no statistically significant difference was found (Fig. 2c). Besides the differences in soil water regimes, the coverage and thickness of biological soil crust differed substantially among dune crest, dune slope and inter-dune lowland (Fig. S3 and S4). Soil surface had consistently higher crust coverage and thickness towards the upper part of the dunes (Fig. S4).

Fig. 2.

Soil mass water content of different depths measured at dune crest, dune slope and inter-dune lowland for (a) late November 2023, (b) mid April 2024, and (c) late April 2024, respectively. Error bars show ±1 SE (n = 3). Different lowercase letters indicate significant differences between micro-sites at different soil depths (P < 0.05, LSD post-hoc test following one-way ANOVA).

Compared to Haloxylon persicum, H. ammodendron showed significantly lower stem hydraulic conductivity with K_s (0.14 ± 0.02 vs. 0.21 ± 0.02 kg m⁻¹ s⁻¹ MPa⁻¹; P = 0.01, t-test) and K_ab (1.19 ± 0.16 vs. 2.23 ± 0.19 × 10⁻⁴ kg m⁻¹ s⁻¹ MPa⁻¹; P < 0.001, t-test) both significantly lower. On the other hand, H. ammodendron showed stem xylem characteristics of enhanced hydraulic safety than H. persicum, e.g. vessel wall thickness and fiber wall thickness were both significantly higher in H. ammodendron (P < 0.001, t-test). Although water potential, osmotic potential at full turgor, and osmotic potential at turgor loss were largely the same between the two shrub species, H. ammodendron showed characteristics of substantially greater degree of succulence than H. persicum. For example, H. ammodendron had significantly higher assimilative branch thickness, higher assimilative branch relative water content, as well as larger water storage capacity than H. persicum (P < 0.001, t-tests).

Changes of the key functional traits along the sand dune slope in individuals of the two shrub species follow the unified patterns. Midday water potential and osmotic water potential at turgor loss both showed significant positive correlations with DDL of the measured individuals (Fig. 3a and b; P < 0.01, linear regression). Similarly, traits related to stem hydraulic efficiency and safety showed consistent patterns of change. Both sapwood area specific hydraulic conductivity (K_s) and assimilative branch area specific hydraulic conductivity (K_ab) of the stems showed positive correlations with DDL, although it was only significant for K_ab (Fig. 4a and b; P < 0.01, linear regression). Consistently, stems of plants growing at higher positions of the dune slope tended to have larger vessels, as indicated by the positive correlation between D_v and DDL (Fig. 4c; P < 0.01, linear regression). Huber value also showed a positive correlation with DDL (Fig. 4d; P < 0.01, linear regression), i.e. plants growing towards the top of the sand dune tended to bear more assimilative branches for certain area of stem cross-sectional area. Contrastingly, traits reflecting stem hydraulic safety tended to decrease with the increase of height of plant distribution, i.e. both T_FW and (T_VW/D_V)² showed significant negative correlations with DDL (Fig. 4e and f; P < 0.01, linear regressions). The structural and physiological traits of assimilative branches related to water storage also showed consistent pattern of change from the inter-dune lowland to the dune crest. The thickness, relative water content, and water storage capacity all showed significant negative correlations with DDL (Fig. 5a–c; P < 0.001, linear regressions), while modulus of elasticity showed significant positive correlation with DDL (Fig. 5d; P < 0.05, linear regression).

Fig. 3.

(a) The correlation between midday water potential of assimilative branches (Ψ_md) and the vertical distance from the inter-dune lowland (DDL), and (b) the correlation between osmotic potential at turgor loss (π⁰) and DDL across all individuals of Haloxylonammodendron (HA) and H. persicum (HP) occurring in the three transects. Inset boxplots show the comparison between the two species and the asterisk in each panel indicates a significant difference at P < 0.05 level.

Fig. 4.

Correlations between (a) sapwood specific conductivity (K_s), (b) assimilative branch specific conductivity (K_ab), (c) vessel diameter (D_V), (d) Huber value, (e) fiber wall thickness (T_FW), (f) vessel wall thickness to span ratio (T_VW/D_V)² and the vertical distance from the inter-dune lowland (DDL) across all individuals of Haloxylonammodendron (HA) and H. persicum (HP) occurring in the three transects. Inset boxplots show the comparison between the two species and the asterisk in each panel indicates a significant difference at P < 0.05 level.

Fig. 5.

Correlations between (a) assimilative branch thickness (ABT), (b) relative water content (RWC), (c) water capacity (C), (d) modulus of elasticity (ε) and the vertical distance from the inter-dune lowland (DDL) across all individuals of Haloxylonammodendron (HA) and H. persicum (HP) occurring in the three transects. Inset boxplots show the comparison between the two species and the asterisk in each panel indicates a significant difference at P < 0.05 level.

Results of PCA analysis on pooled data of Haloxylon ammodendron and H. persicum individuals in the transects based on the stem and assimilative branch functional traits showed that PC1 and PC2 axis explained 30.4% and 11.2% of the total variation, respectively (Fig. 6a). Functional traits associated with higher resistance to drought-induced embolism and assimilative branch traits reflecting higher degrees of succulence clustered at the negative side of PC axis 1, whereas the traits contributing to greater hydraulic conductivity and lower degrees of succulence clustered at the positive side of PC axis 1. The individuals of the two shrub species were clearly divided into two groups on PC axis 1, while no separation was found along PC axis 2 (Fig. 6b).

Fig. 6.

Results of the principal component analysis of the two studied shrub species using 22 measured functional traits. Trait name abbreviations are as defined in Table 1. (a) the correlations of the principal components with each of the original variables, (b) the component scores for all individuals of Haloxylonammodendron and H. persicum. Ellipses represent 95 % confidence intervals for each species group. The light blue, red and dark blue arrows indicate functional traits related to hydraulic efficiency, safety, water relations and storage capacity, respectively.

4. Discussion

Our results show that Haloxylon ammodendron and H. persicum, two sister shrub species of the same genus, have clear separation of micro-habitats in the Gurbantunggut desert of NW China. Individuals of H. ammodendron are mostly limited to the inter-dune lowlands and the foot of the sand dunes, while H. persicum disperse over the upper parts of the slopes all the way to the dune crest. The niche differentiation of the two sister shrub species is paralleled by clear divergences in a suite of water-related functional traits, with H. ammodendron showing syndromes related to stronger drought stress compared to H. persicum. Comparisons among all investigated individuals of the two species, with elevations of each precisely determined, in transects established across the landscape from the inter-dune lowland to dune crest revealed that all the individuals of the two species follow the same unified patterns of changes. A trade-off between hydraulic efficiency and safety, together with the divergence in tissue water storage capacity, have likely contributed to the niche differentiation of the two sister shrub species, which is driven by the change in soil water availability from the inter-dune lowland towards the top of the sand dunes.

4.1. Clear niche differentiation parallels consistent change in fine scale water regime

The non-random association of the two sister shrub species with different micro-habitats in the deserts seems to have a tight association with their divergent adaptation to different degrees of drought stress at their respective micro-habitats. It is known that ecological processes in deserts are mainly limited by water availability and at fine scales within the same climate regime, plant growth and community composition is often coupled to topography due to its strong influence on ecohydrological processes (Noy-Meir, 1973; Rajaniemi et al., 2012; Wu et al., 2016). With the increase of the vertical distance from the inter-dune lowland (DDL), shrubs show increasingly higher midday water potential and turgor loss point osmotic potential, which are indicative of higher environmental water availability to plants (Filella and Penuelas, 2003; Hao et al., 2008; Zhang et al., 2023). Consistently, larger maximum plant height, basal diameter and crown size of plants towards the top of the sand dunes may also indicative of higher environmental water availability. For example, changes in maximum tree height observed at local to global scales have been consistently attributed to the variation in climatological measures of water availability (Moles et al., 2009; Larjavaara, 2014; Liu et al., 2019). Although the pattern of changes in the current investigation (i.e. better water status and growth performance at higher positions of the sand dunes) seem to be counter intuitive, measurements of both physiological parameters and plant growth-related characteristics all suggest that plants have better water availability towards the top of the sand dunes.

Measurements of soil water content at different positions of the transects, in combination with the quantification of soil surface characteristics influencing local hydrology of desert habitats, provide additional evidence to higher water availability at higher positions of the sand dunes. Within the same climatic regime of a desert, variation in topography and surface soil textures often lead to high spatial and temporal heterogeneity of soil water availability due to their strong influences on local hydrological processes, such as run-off, infiltration, and evapotranspiration (Noy-Meir, 1973; Sperry and Hacke, 2002; Muñoz-Reinoso and García Novo, 2005). For example, sandy soils with larger particle sizes in dunes allow easy rainfall infiltration and store more water at greater depth, while clay soils at inter-dunes may increase the water withholding capacity of the rooting zones making water more available to plants over extended dry periods (Ma et al., 2017; Benaafi et al., 2020). In the present study, we found that instead of the difference in soil texture the contrasts in biological soil crust at the very top of the soil surface have likely contributed to changes in soil water regime across the landscape.

Biological soil crust, although contributing to sand stabilization and the retention of soil water, can result in less effective infiltration of rainfall due to their low permeability to water (Li et al., 2002; Coppola et al., 2011; Chamizo et al., 2013). It may play a particularly important role in shaping the fine-scale hydrological process of deserts, where water availability usually exhibit a “pulse and reserve” pattern (Reynolds et al., 2004; Schwinning et al., 2004). In Gurbantunggut desert where the present study was carried out, water from melting snow accounts for a considerable portion of annual water supply to the ecosystem (Li et al., 2010; Wu et al., 2019). If water infiltration is impeded by the thick biological soil crust, this important water source can be lost through run-off or transpiration and become unavailable to the local community (Li et al., 2016a). The coupled spacial variation in soil crust characteristics and patterns of soil moisture in early spring across the landscape shows vividly that biological soil crust has a strong influence on hydrological process of different micro-habitats of the desert. Collectively, topography and the coupled changes in soil crust have at least partially contributed to the sorting of the two sister shrub species with different water requirements by modulating soil water availability across the landscape (Liu et al., 2014). Furthermore, it should be noted that soil salinity might have acted as another critical environmental filter contributing to the niche differentiation of these two shrub species. Though our study highlights the divergence in water availability, the higher salt tolerance of H. ammodendron likely facilitates its success in more saline inter-dune lowlands compared to sand dunes occupied by H. persicum (Tobe et al., 2000; Huang et al., 2003). Future research incorporating both soil water and salinity gradient would provide a more comprehensive understanding of adaptive strategies of these desert shrub species.

4.2. Hydraulic efficiency and safety trade-off in mediating niche differentiation

The consistent changes in stem hydraulic traits with DDL for Haloxylon ammodendron and H. persicum reveal that the trade-off between hydraulic efficiency and safety have likely played an important role in mediating the pattern of niche differentiation along the water gradient caused by variation of topography. Xylem hydraulics has been found to be crucial in determining growth performance and drought tolerance of plants and hydraulic traits play an important role in shaping the fast-slow economics spectrum across species (Poorter et al., 2008; Song et al., 2022). Towards one end of the spectrum, greater hydraulic conductivity supports higher photosynthesis rate and faster growth that would be advantageous under environments with relatively adequate water supply (Agele et al., 2006; Taylor and Eamus, 2008). Towards the other end of the spectrum, lower hydraulic efficiency but enhanced resistance to drought-induced hydraulic failure contributes to greater drought tolerance although often at the cost of reduced resource use capacity and growth potential under resource rich conditions (Larter et al., 2017; Onoda et al., 2017). Consequently, differentiation in xylem hydraulics along such a fast-slow strategy spectrum have been found to be coupled with the sorting of species to different water regimes at different scales (Hao et al., 2011; Gong et al., 2020; Oliveira et al., 2021; Zhang et al., 2023).

Both the interspecific and intraspecific variation in hydraulic traits contributed to the sorting of the two sister shrub species along the water gradient across the desert landscape. All the measured functional traits related to hydraulic efficiency (K_s, K_ab and D_v) and safety (T_VW and (T_VW/D_V)²) showed significant differences between Haloxylon ammodendron and H. persicum, indicating clear divergence between the two sister species in xylem hydraulics. This is consistent with findings that closely related species in phylogeny or even different ecotypes of the same tree or shrub species can differ substantially in hydraulic characteristics that contribute to adaptation to different water regimes (Preston and Ackerly 2003; Hao et al., 2009, 2011, 2013; Zhang et al., 2017). The significantly lower stem hydraulic conductivity in H. ammodendron and xylem structural traits related to greater drought tolerance allowed stronger endurance in the inter-dune habitats with more extreme drought stress (Pockman and Sperry, 2000; Yang et al., 2022). Differently, higher xylem water transport efficiency would contribute to higher photosynthetic carbon assimilation and growth rate, allowing H. persicum take advantage of the relatively higher water availability on the sand dunes. The relatively high growth rate of H. persicum might be of great importance for this species occupying sand dune habitats, where keeping pace with the dune drifting and the avoidance of sand burial through fast growth are considered to have high adaptive significance (Niu et al., 2004; Guo et al., 2020).

Notably, Haloxylon persicum exhibited relatively high degrees of plasticity in hydraulic traits (Fig. 3), suggesting that a certain degree of plasticity is important for occupying the sand dune slope habitats with a relatively high spacial variation from the bottom to the top of the dunes. It has been known that hydraulic traits are considerably conservative and the interspecific variations are usually much larger than the within species differences caused by variations of proximal water conditions (Rosas et al., 2019; Andrés-Hernández and Rodríguez-Ramírez, 2025). Nevertheless, the relatively large variation in xylem traits related to hydraulic efficiency and safety in H. persicum along the sand dune slope indicates that a certain degree of plasticity in hydraulics at the intraspecific level endowed this species the ability to occupy the dune slope with a relatively large spacial heterogeneity in water availability.

4.3. Divergence in water storage in shaping adaptability to different water regimes

Comparisons of physiological characteristics at the assimilative branch (functionally equivalent to leaf) level between the two sister shrub species indicated that the divergence in tissue water storage capacity also played an important role in mediating their niche differentiation across the landscape. The success of Haloxylon ammodendron at inter-dune lowland with more extreme drought stress, including physiological drought related to high soil salinity, has likely benefited from its high tissue water storage capacity. In desserts, a high degree of tissue succulence contributes greatly to surviving extended dry periods (Fletcher et al., 2018; Guo et al., 2024). Characteristics related to higher water storage capacity can buffer fluctuations of water status and delay damage from drought to a certain extent (Gleason et al., 2014; Huo et al., 2022). In addition, succulent organs of desert plants having well-developed water storage structures (such as vacuoles) can also store soluble sugars related to osmotic regulation, thereby reducing the risk of cells being subjected to strong osmotic stress (Mediavilla et al., 2008).

Our results show that the thicker assimilative branches of Haloxylon ammodendron tend to exhibit higher water storage capacity and smaller modulus of elasticity, which have likely endowed it stronger adaptability to more challenging drought stress in the inter-dune lowlands compared to its sister species H. persicum. In deed, previous studies have shown that the water storage tissues in H. ammodendron are well developed, resulting in high water retention ability and reduced osmotic stress during drought (Xu et al., 2016; Liu and Zheng, 2024). Similar to the pattern found for stem hydraulic traits, water-related physiological traits (e.g. Ψ_pd, π⁰, C, and ε) in H. persicum also exhibited relatively high degrees of plasticity, which suggests that a functional trade-off of the assimilative branches at the intra-specific level have also contributed to the occupation of the overall less drought stressed but more heterogeneous water conditions on the sand dune slope (Janssen et al., 2020; Jiang et al., 2021).

5. Conclusions

Our results showed that, within the same climate regime of a desert, the fine-scale soil moisture can be strongly affected by local factors like topography and soil biological crust coverage that eventually result in the fine-tune niche differentiation of plant species across the landscape. Particularly, a fast-slow economics spectrum supported by a trade-off between hydraulic efficiency and safety as well as variation in tissue water storage capacity seem to be a critical mechanism explaining the distribution pattern of the two sister shrub species in the overall drought stressed but still heterogenous desert environment. Understanding the subtle but important differences in plant functionality in terms of water relations and environmental heterogeneity in water regime would contribute to more effective conservation of desert plant species and communities where water is the most important controlling factor.

CRediT authorship contribution statement

Chi Zhang: Investigation; Methodology; Formal analysis; Writing—original draft. Guang-You Hao: Conceptualization; funding acquisition; supervision; formal analysis; writing—review & editing. Chun-Yang Duan: Formal analysis; Review & editing. Jing Zhang: Review & editing. Ben-Feng Yin: Formal analysis. Yuan-Ming Zhang: Conceptualization; Funding acquisition; Project administration; Supervision; Writing—review & editing.

Data availability statement

All the data used in this article are available in the supporting information.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is/are the supplementary data to this article.