A Comparison of Water Uptake by Transpiration from Different Soil Depths among Three Land Cover Types in the Arid Northwest of China

Abstract

In recent decades, the frequency, intensity, and extent of extreme drought events have posed serious threats to ecosystems in vulnerable regions. With low annual precipitation, the arid area in northwest China is a typical ecologically fragile area, and extreme drought events will aggravate desertification in this area. In order to control desertification, various experimental plantations have been established in Northwest China. However, there is no consensus on which plantations are more suitable to become widespread. To explore this, we conducted a comparative study on different plantations from the perspective of long-term deep (100 cm depth) soil moisture balance. In our study, six typical ecosystems were selected for comparison of the variation of soil moisture and control factors. The results showed three main findings. First, the soil moisture of all six ecosystems showed a similar hierarchy of increasing moisture with the increasing depth of the soil layer. However, the deep layer soil moisture (mean = 0.33 ± 0.22 cm³·cm⁻³) of the artificial poplar (Populus alba) forest exhibited a downward trend over time after the fifth year, but did not at the shallow layer for this ecosystem. Second, the trends of the maximum canopy coverage between the different ecosystems from 2010 to 2019 showed significant differences from one another, with the maximum value of the leaf area index for the poplar forest being the highest (Maximum = 7.13). Third, a negative correlation (R² = 0.52) was found between deep soil moisture and transpiration for the poplar forest, and a positive correlation (R² ≥ 0.23) between these two metrics was found for the other five ecosystems. The results revealed that transpiration processes had a different consumption of deep soil moisture due to the differences in the root and canopy density of several plantations. Among these ecosystems, the transpiration of the artificial poplar forest is noticeably large, resulting in a unilateral decline in soil moisture.

1. Introduction

In recent decades, the frequency and intensity of extreme drought events have increased worldwide, and they have had a significant and adverse impact on ecologically fragile areas [1,2,3]. Ecologically fragile zones, also known as ecotones, are transitional zones between two different types of ecosystems [4,5]. The main characteristics of the system are: (1) a weak anti-interference ability, (2) sensitivity to global climate change, (3) strong temporal and spatial volatility, (4) a significant edge effect and (5) high environmental heterogeneity [6]. One potential consequence is the expansion of desertification for ecologically fragile zones in arid areas [4,5]. Generally, the primary limiting factor for vegetation in ecosystems in ecologically fragile zones in arid regions is water [1,7,8]. On one hand, natural precipitation is relatively low, on the other hand, due to dry air, there is high potential evaporation resulting in a vicious cycle of water scarcity [7,9]. The arid Ningxia Hui Autonomous Region of Northwest China, located in central Asia, has been experiencing increasingly substantial desertification and ecological degradation resulting from low precipitation and increasing annual temperatures [10]. The Ningxia Hui Autonomous Region is a typical dry watershed with annual precipitation of 289 mm·year⁻¹, which supports low vegetation cover [11]. Poor land-use strategies and excessive land reclamation have led to significant land degradation within this area [12]. To solve this problem, artificial forests and grass establishment projects, such as the Grain for Green Project, have become important methods to ecologically restore arid areas, where the threat of desertification has become extremely serious [13,14]. However, as the plantations continue to grow and demand more water, more problems have been exposed [15]. The lack of natural precipitation in arid areas makes it difficult to meet the growth demand of plantation forests, thus resulting in the eventual death of large areas of the plantations [16]. Furthermore, high transpiration consumption of water leads to rapid soil water loss, which has a detrimental impact on the original soil ecosystem and surface ecosystem [7]. This results in a worse post-death ecosystem than the original ecosystem, before the conversion of farmland to forest. Therefore, it is important to master the effects of current conversions of farmland to forest in order to better evaluate the potential growth state and influence the direction of the consumption of soil moisture by different species of plantations.

Soil moisture is a key factor for the survival of soil ecosystems and surface ecosystems, and its variation determines the type and health of vegetation [17,18]. For a given climate, the perennial state of soil moisture determines the type and limit of plants that can be supported. The threshold of soil water necessary for the survival of different plants varies greatly [19]. In turn, vegetation impacts water balance by regulating surface evapotranspiration and soil moisture [20,21]. Plants themselves have the function of conserving water, and at the same time, they are also the subject of water consumption [22,23]. The outcome of the game between water retention and water consumption depends on the climatic characteristics of a region, that is, the characteristics of continuous water supply and consumption [24]. The continued stability of surface vegetation ecology comes from the long-term balance of soil moisture [23,25]. This is why, when dealing with desertification, artificial plantations are thought to ensure the long-term stability of soil moisture [26]. As one of the important components of evapotranspiration, transpiration bonds water, the energy cycle, and carbon cycles, and it is an important reason for soil moisture loss to the air of ecosystems [27,28]. Vegetation canopy density is an important index to evaluate the vegetation of an ecosystem. It accelerates soil moisture loss through transpiration but it also reduces soil moisture loss through evaporation due to its shading and cooling function [29]. Because of the differences in climate and soil conditions, various canopy densities have completely different effects on soil water balance [30,31]. Although there have been numerous relevant studies, none have provided substantial guidance for preventing the expansion of desertification within arid areas.

This study aims to clarify the variation tendency of soil moisture for several typical vegetations and characterize the controls and their relationship with soil moisture for arid areas of Northwest China.

2. Materials and Methods

2.1. Study Area

As shown in Figure 1, the study sites were conducted in Yinchuan, Zhongwei, and Tongxin of the Ningxia Hui Autonomous Region, China, located at 36°0′ to 39°23′ N and 104°17′ to 107°39′ E, covering an area of approximately 51,400 km². The Yellow river enters into Ningxia Hui Autonomous Region in Zhongwei, which flows into Inner Mongolia after passing through Shizuishan [32]. The geomorphic type is alluvial plain, with flat terrain and free canals with an elevation ranging from 1064 to 3523 m above sea level. It is a variable arid zone with an annual average sunshine duration between 2750–2950 h and an annual average wind speed ranging from 1.7 to 2.5 m·s⁻¹. The area receives sufficient sunshine and heat for vegetation growth. The temperature difference is broad, but the frost-free period is 105–163 days. With an average annual precipitation of 180–220 mm, the river basin is a continental climate. There is a period of drought and little rain, and the annual precipitation varies from north to south. There is a relatively significant and uneven distribution of monthly precipitation in the Yellow River irrigation area. In the central and southern mountains, where seasonal rainfall variation occurs, the drought zone also moderately varies. The precipitation in July-September accounts for 60%–70% of annual precipitation. This leads to marked dry and wet seasons as well as strong potential evapotranspiration that can reach up to an annual average of 1100–1600 mm [11]. Soil types are mainly sierozem (33.0%), loessal soil (20.2%), aeolian sandy soil (13.1%), alluvial soil (8.0%), and rhogosol (5.3%) [33,34]. In the basin, land use gives priority to grasslands and croplands which account for 33.74% and 48.76%, respectively. The main crops in this area are wheat, rice, corn, soybean, and wolfberry [35]. Areas occupied by forests and water account for 3.60% and 1.43%, respectively, and the main tree species include Cypress, spruce, Pinus tabulaeformis, Robinia pseudoacacia, national Robinia pseudoacacia, poplar, and willow [36]. Finally, the unused areas (mainly saline and alkaline land, desolate beach, sand land, and sand and gravel mixed land) account for 11.01% [37].

2.2. Field Measurements, Data Collection, and Processing Analysis

We used the cutting-ring method in our study to collect soil samples and measure soil moisture for six ecosystems. As shown in

Table 1. Basic information for study sites.

Tegetation Types	Location	Coordinates	Elevation (m)	Annual Precipitation (mm)	Annual Air Temperature (°C)	Soil Types	Tree Spacing (m × m)	Establishment Date
Natural Grassland	Zhongwei	36°38′40″ N 106°6′25″ E	1400			Sierozem	NAN	2010
Artificial Grassland				145.6–276.4	9.2–11.2		0.15 × 0.15	2010
Artificial Elaeagnus angustifolia	Tongxin	38°9′20″ N 106°18′28″ E	1120			Sierozem	5.0 × 5.0	2010
Artificial Poplar forest				153.4–289.4	9.1–10.6		5.0 × 8.0	2010
Artificial Sand willow forest	Yinchuan	38°25′06″ N 106°10′35 E	1115			Anthropogenic alluvial soil	6.5 × 7.5	2010
Artificial Buddleia				146.3–286.7	9.3–10.8		0.5 × 3.0	2010

, these ecosystems include two arbor forests (artificial poplar and artificial Elaeagnus angustifolia), two shrub wood forests (artificial sand willow and artificial buddleia), and two grasslands (one natural and one artificial). Two samples were taken with three duplicate samples at each of the following layers within the soil of every ecosystem: at shallow layer (10 and 20 cm depth), middle layer (40 and 60 cm depth), and a deep layer (100 cm depth). The samples were collected once a month from 2010 to 2019 while simultaneously taking leaf area index (LAI) measurements using a LAI-2200C plant canopy analyzer (LI-COR Ltd., Lincoln, NE, USA).

Meteorological data were obtained from http://nx.cma.gov.cn/ (accessed on 1 May 2023) and hydrological data were obtained from http://www.hwswj.com.cn/ (accessed on 1 May 2023). Transpiration was estimated using the modified Shuttleworth–Wallace (SW) model and the equation can be expressed as:

$$P{M}_c=\frac{\mathsf{\lambda }E-C_{s}P{M}_s+E_i}{C_c}$$

(1)

where PM_s is evaporation of surface soil, PM_c is the transpiration of the plantation, E_i is evaporation from canopy surface water intercepted by leaves C_c and C_s are the coefficients that depend on the resistances.

$$P{M}_s=\frac{△\left(R_n-G\right)+\left[\rho c_{p}D-△r_a^s\left(R_{ns}-G\right)\right]/\left(r_a^a+r_a^s\right)}{△+\gamma \left[1+r_s^s/\left(r_a^a+r_a^s\right)\right]}$$

(2)

$$P{M}_c=\frac{△\left(R_n-G\right)+\left[\rho c_{p}D-△r_a^c\left(R_{ns}-G\right)\right]/\left(r_a^a+r_a^c\right)}{△+\gamma \left[1+r_s^c/\left(r_a^a+r_a^c\right)\right]}$$

(3)

where ∆ (kPa·°C⁻¹) is the change rate for SVP (saturated vapor pressure) with Ta (temperature), and R_ns (MJ·m⁻²·d⁻¹) is the net radiation hitting the surface soil, and d R_n (MJ·m⁻²·d⁻¹) is the net radiation above the canopy of the vegetation. G (W·m⁻²) is the heat flux for underground soil, ρ (kg·m⁻³) is air density at the experimental site, D (Pa) is VPD (the vapor pressure deficit) at a particular height for reference, c_p (J·kg⁻¹·k⁻¹) is the specific heat at a particular pressure, and γ is the psychrometric constant. $r_s^s$ (m·s⁻¹) is the resistance of the surface soil, $r_s^c$ (m·s⁻¹) is the resistance for the bulk canopy, and $r_a^c$ (m·s⁻¹) is the bulk boundary layer resistance between a hypothetical level in the canopy and the canopy elements. $r_a^a$ (m·s⁻¹) is the aerodynamic resistances between the surface soil and some hypothetical level in the canopy, and $r_a^s$ (m·s⁻¹) is the aerodynamic resistances between the surface soil and the reference height [38].

Some parameters cannot be measured directly and need to be estimated indirectly from other data resources. For example, r_a (m·s⁻¹) is the aerodynamic resistance between the atmosphere and the canopy, and it can be obtained from the equation below [39,40,41,42]:

$$r_a=\frac{\ln \left(\frac{z-d}{z_0}\right)\ln \left(\frac{z-d}{z_e}\right)}{k^2\mu }$$

(4)

where k and μ are the Von Karman constant (k = 0.41) and wind speed at height z (m·s⁻¹), respectively. d is the zero plane displacement height (taken as 66% of the plant canopy height h in meters, d = 0.67 h, m), and z_e and z₀ are the roughness length for vapor transport (z_e = 0.1 z₀, m) and the movement roughness length (z₀ = 0.13 h, m), respectively [42,43].

The parameter r_c (m·s⁻¹) is influenced by climatological, biological, and agronomical factors and is the resistance of the canopy to the diffusion of water vapor from the inner leaves into the surrounding atmosphere [42]. It is calculated as:

$$r_c=\frac{r_s}{LA{I}_e}$$

(5)

where LAI_e is leaf area index for the effective leaves in the transpiration process. It’s easy to see, r_c (m·s⁻¹) is dependent on the function of the daily mean stomatal resistance of single leaves r_s and the LAI_e (m²·m⁻²). It can also be obtained from the equations below [42,44,45,46]:

$$r_c=\frac{r_{min}}{LA{I}_e}\times F_1\times F_2{}^{-1}\times F_3{}^{-1}\times F_4{}^{-4}$$

(6)

$$F_1=\frac{1+0.55\frac{R_G}{R_{GL}}\times \frac{2}{LA{I}_e}}{\frac{r_{min}}{r_{max}}+0.55\frac{R_G}{R_{GL}}\times \frac{2}{LA{I}_e}}$$

(7)

$$LA{I}_e=\left\{\begin{array}{c}LAI LAI\le 2\\ 2 2<LAI<4\\ \frac{LAI}{2} LAI\ge 4\end{array}\right.$$

(8)

where F₁ is the function for stomatal resistance, photosynthetically active radiation, and LAI. r_max (m·s⁻¹) and r_min (m·s⁻¹) are the maximum and the minimum stomatal resistance of vegetation, respectively. LAI_e is the effective LAI, R_G (MJ·m⁻²·d⁻¹) is active radiation that reaches the vegetation canopy, and R_GL (MJ·m⁻²·d⁻¹) is a limit value of 100 W·m⁻² for vegetation [42].

The factor F₂ took into account the effect of water stress on surface resistance. It will vary between 0 and 1 along with θ varying between θ_w and θ_f [42,47].

$$F_2=\left\{\begin{array}{c}1 \mathrm{if} \theta >{\theta }_f\\ \frac{\theta -{\theta }_w}{{\theta }_f-{\theta }_w} \mathrm{if} {\theta }_w\le \theta \le {\theta }_f\\ 0 if \theta <{\theta }_w\end{array}\right.$$

(9)

where θ (cm³·cm⁻³) is the soil water content, θ_f (cm³·cm⁻³) is the field water capacity, and θ_w (cm³·cm⁻³) is the wilting water content.

The factor F₃ represents the effects of the VPD of the atmosphere [44,48].

$$F_3=1-\beta \left(e_s-e_a\right)$$

(10)

where β is a species-dependent empirical parameter (0.025–0.061) and e_s (Pa) and e_a (Pa) are saturation vapor pressure and actual vapor pressure, respectively [42].

The factor F₄ introduces an air temperature dependence on the surface resistance and is expressed as [45,46]:

$$F_4=1-0.0016{\left(298.0-T_a\right)}^2$$

(11)

2.3. Drawing Platform and Analytical Method

In this paper, OriginPro 2016 was used for data mapping and Arcgis 10.4 was used for map processing. To analyze the relationship between transpiration and deep (at 100 cm) soil water content for the different ecosystems, a curvilinear equation and the determination efficiency (R²) were used [49,50], and the determination efficiency was calculated using the following equation:

$$R^2={\left[\frac{{{\displaystyle \sum }}_{t=1}^n\left(Q_m^t-{\overline{Q}}_m\right)\left(Q_O^t-{\overline{Q}}_O\right)}{\sqrt{{{\displaystyle \sum }}_{t=1}^n{\left(Q_m^t-{\overline{Q}}_m\right)}^2}\sqrt{{{\displaystyle \sum }}_{t=1}^n{\left(Q_O^t-{\overline{Q}}_O\right)}^2}}\right]}^2$$

(12)

where n is the number of data sets, $Q_m^t$ is the modeled value, $Q_o^t$ is the observed data, $\overline{Q_m}$ is the average modelled value, and $\overline{Q_o}$ is the average observed value.

3. Results

3.1. Annual Soil Moisture Variation at Different Depths for Different Ecosystems

The soil moisture of all six ecosystems showed a similar hierarchy of increasing moisture with the increasing depth of the soil layer. The absolute values of the soil moisture, however, vary from ecosystem to ecosystem for the same depths. For the two arbor forest ecosystems, the soil moisture of the Elaeagnus angustifolia forest is higher in the deep and shallow layers, but lower in the middle layer than in the poplar forest. For the two grassland ecosystems, the soil moisture of the artificial grassland is higher in the deep and shallow layers, but lower in the middle layer than in the natural grassland. For the two shrub wood forest ecosystems, soil moisture of the sand willow forest in the deep and middle layers is higher, but lower in the shallow layer than in the Buddleia Forest. With the growth of plants over time, the variation trends of soil moisture in different soil layers for each ecosystem gradually produced differences. Figure 2a shows, in detail, that the deep and middle layer soil moisture of the artificial poplar forest exhibited a downward trend over time after the fifth year, but not at the shallow layer for this ecosystem. In regard to the two grasslands and shrub wood ecosystems, deep soil moisture exhibited a stable trend over time and had little variability, while shallow soil moisture showed a high coefficient of variation.

3.2. Annual Canopy Coverage Variation for Different Ecosystems

Figure 3 shows the max LAI of each ecosystem for every year from 2010 to 2019 with the visible trends showing great differences between one another. As shown in Figure 3a, the LAI of the poplar forest increased faster than the Elaeagnus angustifolia forest, it being lower from 2010 to 2013 and higher starting in 2014. It also shows that the growth rate of the poplar forest is significantly faster than the Elaeagnus angustifolia forest. The LAI of the natural grassland shows a stable trend with a small fluctuation, while the LAI of the artificial grassland shows extremely rapid growth during the first five years, which then began to become a relatively stable trend. For the shrub wood forest ecosystem, the LAI of the sand willow forest exhibited a unilateral trend of steady growth, while the LAI of Buddleia forest almost reached its peak in the third year, followed by a steadier trend with minimal fluctuation throughout the rest of the study. Among all of the ecosystems, the poplar forest had the highest max value of the LAI (7.13), while the natural grassland had the lowest max value of the LAI (1.93).

3.3. Relationship between Transpiration and Deep Soil Moisture for Different Ecosystems

As shown in Figure 4, the relationship between deep soil moisture and transpiration for different ecosystems exhibited marked differences. The two were negatively correlated for the artificial poplar forest, with transpiration increasing over time and deep soil moisture decreasing, and positively correlated for the other five ecosystems. Transpiration showed an increasing trend over time, while deep soil moisture showed a decreasing trend for the poplar forest. This is likely due to the fact that the deep soil moisture contributes greatly to a plant’s transpiration process in this ecosystem. Transpiration for the artificial Elaeagnus angustifolia forest, artificial grassland, and artificial buddleia forest showed an increasing trend while deep soil moisture did not exhibit a decreased trend. This is mainly because the transpiration process of vegetation in these ecosystems consumes a large proportion of shallow water. The transpiration of natural grassland and artificial sand willow forest did not show an upward trend and the deep soil moisture did not show a downward trend. This indicated that the proportion of deep water and shallow water consumed by the transpiration process of these two ecosystems did not change.

4. Discussion

In our study, the deep soil moisture of the artificial poplar forest exhibited a downward trend over time. This means that the water consumed by the poplar forest through transpiration was in a rapid growth pattern since the fifth year [17,51]. In terms of the water balance, this also means that the consumption of soil water was greater than the supply in the deep soil layer [52,53]. An increase in transpiration is typically due to an increase in the water consumption capacity of plants [54,55]. However, another explanation is that an increase in water supply, or an increase in solar radiation and heat, could lead to an increase in the pressure difference of saturated water, thus forcing the plants to increase their water consumption capacity [28,56]. From the perspective of the transpiration capacity of plants, the most common control is the LAI, which is often positively correlated with transpiration [57,58]. In our study, the change trend of the LAI shows that the growth rate of the poplar forest is significantly faster than the Elaeagnus angustifolia forest or any other ecosystem. Although the LAI and transpiration of each ecosystem showed an increasing trend, the upper limit of the LAI of each ecosystem varied. The upper LAI limit of the poplar forest was relatively high and exhibited a steady and rapid growth. Correspondingly, the transpiration rate of the poplar forest was also in a stable and rapid increase. The study area is located in an arid zone, characterized by low precipitation. This means that the poplar forest consumes deep water faster than any other ecosystem, leading to a faster drop in deep soil water.

The shallow soil water consumption did not change much throughout the study period. This suggests that either the natural conditions, such as VPD (vapor pressure deficit), did not increase significantly during the study period or that the water supply and consumption were in a relative balance during this period of time [59,60]. In contrast, deep soil water consumption was higher and can be directly related to the distribution of roots within the soil [61,62,63,64,65,66]. This explains why the deep-water consumption of poplars is higher than other ecosystems since the roots of poplars are significantly denser at depth than those of grasses and shrubs. It can be concluded from the variation trend of shallow soil water that there has been little change in energy in this region within the last 10 years, which disproves that deep water consumption is the result of heat change or polar radiation change. When the deep soil moisture is reduced to a certain extent, the photosynthesis of trees and other large plants can no longer be greater than the respiration. This means that these large plants will stop growing, or even gradually degrade until they die due to the inertia of excessive consumption capacity [52,67,68]. According to the actual survival rate and effect of afforestation, large-scale death and degradation did occur in some forest lands, which further confirms the results of this study [13,15]. Extreme weather could also be a contributing factor. As plants grow to match the current water supply within the soil, if they are suddenly hit by an extreme drought, this causes them to gradually die since they cannot adjust their water conservation mechanisms [69,70,71,72].

5. Conclusions

In this study, the variation tendency of soil moisture for six typical ecosystems in Northwest China, both natural and artificial, was compared, then the impact controls and their relationship between moisture for arid areas of Northwest China were characterized. The deep soil moisture of the artificial poplar forest exhibited a more significant downward trend than the other ecosystems after the fifth year. This means that the water consumption through transpiration of the poplar forest from soil was in a rapid growth pattern since the third year. The reason that the deep soil water consumption was higher is likely directly related to the distribution of roots since the roots of poplars are significantly denser at depth than those of grasses and shrubs. This study has important directive and practical significance for desertification control by revealing the effect of plantations on soil moisture in arid environments.