Checkerboard Barriers Attenuate Soil Particle Loss and Promote Nutrient Contents of Soil

Abstract

In the arid and semi-arid regions of the world, checkerboard barriers play a vital role in ecological restoration. However, the improvement of soil quality in the later stages of lattice barrier-laying is an issue that is not yet known. This study selected dunes lined with Salix psammophila (SL) and high-density polyethylene (HDPE) checkerboard barriers in the desert zone of an arid zone, and no-barrier dunes were used as the control area. We assessed the soil properties of different dunes’ windward slopes using a combination of particle size characteristics and soil nutrients and using soil nutrients to build a soil quality index (SQI). The laying of checkerboard barriers has a positive effect on the accumulation of fine soil particles and the improvement of soil quality. It ultimately leads to an increase in fractal dimension in the 0–2 cm and 2–4 cm soil layers. SQI and soil nutrients show consistent trends. The SQI of the different dune types follows the order: SL (0.22) > HDPE (0.14) > CK (0.12). We also found an interaction between soil nutrients and particle characteristics through statistical analysis. With a comprehensive assessment of checkerboard barriers installed on dunes, SL can provide better soil restoration benefits.

1. Introduction

Desertification is a major environmental problem and is inextricably linked to regional economic and social development [1]. The financial losses due to desertification in China are reported to be estimated at USD 6.8 billion per year. Increased infrastructure protection costs, devaluation of land values, loss of crops, and damage to human livelihoods are some of the obvious impacts of desertification [2,3]. As a result, the United Nations Conference on Desertification was held in the 1980s, after which China officially began a multidisciplinary and integrated approach to “desertification”. Over the past three decades, researchers have made some positive progress in the control of desertification, with much of the work taking place in areas of fragile ecological conditions and high human activity [4]. Several scholars working on aeolian research have successively proposed ecological and engineering interventions, such as artificial vegetation and checkerboard barriers, to control desertification hazards in northern China [5,6]. However, extremely desertified land will move under the action of wind, and the survival rate of vegetation was relatively low, only about 15% [7]. Compared with afforestation, planting drought-tolerant grasses and shrubs improves the water-use efficiency and survival rate of vegetation [1,8], but areas where this method is employed ultimately lack vegetation cover and are often buried by sandy sediments [9]. Therefore, engineering measures (such as checkerboard barriers) are usually used first for desertification control [10]. Checkerboard barriers can disrupt wind–sand flow and reduce the sand carrying capacity, reducing wind erosion and transport of sand particles and increasing deposition in different areas [11,12]. This process leads to the accumulation of fine particle sediments in the checkerboard barrier and provides support for soil development and vegetation growth. In addition to this, it also facilitates soil nutrient retention, carbon sequestration, and oxygen release in desert ecosystems. As an indicator of soil structure and texture, the soil particle size distribution is a crucial factor in measuring soil properties due to the interaction of various factors. The grain size distribution records the evolution of the depositional environment and the information of many studies on desertification (e.g., Mz (mean size), Sd, Sk and Kg) [13]. In addition to this, particle size is highly correlated with soil erosion and degradation (e.g., fractal dimension) [14] and also influences soil fertility and nutrient transformation [15]. Salix psammophila and high-density polyethylene (HDPE) checkerboard barriers are two common artificial sand fixation measures in northwest China because of their long-lasting resistance to wind. In many studies, after laying Salix psammophila checkerboard barriers and HDPE checkerboard barriers in the desert, researchers have demonstrated that the aeolian activity was weakened [9,16,17,18]. These are the significant impacts that can be achieved in the early stages of sand barrier-laying. Some explanations for the collapse of Salix psammophila checkerboard barriers have also been provided by some researchers, who suggest that continuous UV radiation in the atmospheric environment and microbial and fungal activities in the sandy soil lead to the destruction of lignin and cellulose components. However, fewer studies have been conducted on the effects of soil particle properties at later stages of Salix psammophila checkerboard barriers and the placement of HDPE checkerboard barriers in the desert. In addition, the evolution of soil properties after checkerboard barrier restoration remains unclear in large-scale desert areas.

The Hobq Desert, the seventh-largest desert in China, borders the Yellow River in its western, northern, and eastern parts. Among them, 10 tributaries of the Yellow River cross the desert in the eastern part of the Hobq Desert, known as the Ten Tributaries. The total watershed area of these 10 tributaries is 10,767 km² [19]. The risk of wind erosion was assessed by IWEMS (Integrated Wind-Erosion Modeling System) and RWEQ (Revised Wind Erosion Equation) models and compared with observations from local government and hydrological stations; the saltation emission in the Hobq Desert was 31 t/m between 2001 and 2010, and the saltation sediment in the Yellow River was about 6.35 × 10⁶ t from 2001 to 2010 [19,20]. Desertification has been effectively controlled after the Chinese government and local enterprises implemented strategies to combat wind erosion hazards [21,22].

This study investigated the soil particle size and soil nutrient distribution on the windward slopes of dunes under two specific engineering sand control measures in the Hobq Desert, China. This study aimed to (1) characterize soil particle size and soil nutrient distribution within two checkerboard barriers in the study area, and (2) assess the soil structure and restoration status of the two checkerboard barriers using soil quality and particle size characteristics as indicators, as well as identify new, environmentally conscious, effective materials that are more suitable for sand fixation in the sandy areas of the Hobq Desert. This suggests a reasonable strategy for restoring soil structure and improving soil quality.

2. Materials and Methods

2.1. Description of the Study Area

The study area is located in the northern part of the Hobq Desert, China (108°40.50′~108°55.05′ E, 40°29.37′~40°26.02′ N). The average annual precipitation is 188.4 mm. The mean annual temperature is 7.8 °C, with a minimum value of 6.5 °C and a maximum value of 9.5 °C. Northwest and westerly winds prevail in the area, with an annual average wind speed of 3.2 m∙s⁻¹ and a maximum wind speed of 25 m∙s⁻¹. The frequency of west and northwest winds is the highest (Figure 1). The original landforms in the study area are moving dunes, and the dunes are mainly barchan dunes and barchan dune chains with northwest-to-southeast alignment that are mobile and fast-moving. Salix psammophila checkerboard and HDPE checkerboard barriers have been established over the years to mitigate wind and sand damage to towns and industrial parks and the Yellow River basin.

2.2. Study Area Field Investigation

In this study, dunes where Salix psammophila checkerboard barriers and HDPE checkerboard barriers had been established for seven years were selected as the experimental group, and no-barrier dunes were used as the control group (CK). Before the setting up of checkerboard barriers, all dunes were in the same environmental conditions.

Table 1. Plant species and general coverage of different sand dune types.

Type	Herbaceous Species	Coverage
SL barrier dunes	Artemisia desertorum; Agriophyllum squarrosum; Psammochloa villosa	20–45%
HDPE barrier dunes	Agriophyllum squarrosum; Corispermum patelliforme; Psammochloa villosa; Bassia dasyphylla	20–35%
No-barrier dunes	Agriophyllum squarrosum	<3%

shows the cover of herbaceous plant species growing on the three types of dunes.

The vegetation coverage of the control dunes is less than 3%; by comparison, the vegetation cover of dunes with SL barriers is 20–45%, and the cover of dunes with HDPE barriers is 20–35%. The dunes we selected have a windward slope of 19–20° pointing northwest and a leeward slope of 35–36°. Their windward slope length varies from 33 to 35 m, and the leeward slope length varies from 16 to 18 m. The width of the dunes varies from 30 to 35 m, while the separation distance between these three types of dunes is about 10 m. The checkerboard barriers of Salix psammophila or HDPE were established on the windward and leeward slopes (size: 1 m × 1 m; height: 30 cm). One side of the grid is perpendicular to the main wind direction. Figure 2 shows the field survey of no-barrier dunes, SL barrier dunes, and HDPE barrier dunes.

2.3. Soil Sampling

In this study, we selected different locations within the checkerboard barriers on the windward slopes of the dunes to collect soil. This is because soil loss occurs mainly on the windward slopes of dunes, and studies have shown that the aeolian conditions and their results on soil particles are different at different dune locations. To avoid the accumulation of soil nutrients by herbaceous plants, we collected soil samples within checkerboard barriers without plants whenever possible. Three similar dunes with similar height, slope, and migration characteristics were selected for each type of dune. Soil samples were collected as shown in Figure 3 on the surface soil profiles at the bottom, lower middle, middle, upper middle, and top (E to A) of the windward slope of each dune, with the sampling area being the size of the checkerboard barrier layout (1 m × 1 m). In each sampling area, soil samples were collected at five different points within the checkerboard barrier, in the order a, b, c, d, and e (Figure 3). We collected topsoil from 0–2 cm and 2–4 cm in the sampling area at different locations (E to A). Five samples from the same soil layer in each sampling area were combined. The sampling area was located on the centerline perpendicular to the dunes’ strike (E2 to A2), and two other parallel sampling areas (E1 to A1 and E3 to A3) were located 8 m from the centerline. Soil from the same position in the dune was collected and mixed into one test soil sample, which was brought back to the laboratory for processing; debris was removed through a 2 mm soil sieve before analysis [6].

2.4. Physical and Chemical Analysis

We measured the particle size distribution of soil samples using an Analysette 22 NanoTec laser particle size meter (FRITSCH, Idar-Oberstein, Germany). Before measurement, the soil samples were sieved through a 2 mm soil sieve and treated with organic matter removal and desalination. Soil samples were passed through a 0.15 mm soil sieve to measure their organic carbon (SOC), total nitrogen (TN), and total phosphorus (TP) contents. Soil passed through the 1 mm soil sieve was used to determine alkaline hydrolysis nitrogen (AN), available phosphorus (AP), and available potassium (AK) contents. SOC was measured using the volumetric heating method with potassium dichromate. Total N was measured by the semi-micro-Kjeldahl method using a titration of 0.01 mol·L⁻¹ 1/2 H₂SO₄. AN was measured by the Conway diffusion dish method, using 1.0 mol·L⁻¹ NaOH to hydrolyze the soil and 20 g·L⁻¹ H₃BO₃ to absorb the NH₃ released by diffusion and titrated with 0.005 mol·L⁻¹ H₂SO₄ [24]. For TP analysis, soil subsamples were fused with NaOH, and P concentrations were measured with the ammonium molybdate method in a U/V spectrophotometer [25]. AP was extracted with 0.5 mol·L⁻¹ NaHCO₃, measured by Mo-Sb colorimetry, and quantified by spectrophotometry. TK was determined in the fused soil samples with a flame photometer. AK was extracted with 1 mol·L⁻¹ NH4OAc (leaching), and its concentration was measured with a flame photometer.

2.5. Data Analysis

We used the USDA soil size classification criteria, which divide samples into clay (<0.002 mm), silt (0.002–0.05 mm), and sand (0.05–2 mm) (Among the latter are very fine sand (0.05–0.1 mm), fine sand (0.1–0.25 mm), medium sand (0.25–0.5 mm), coarse sand (0.5–1 mm), and very coarse sand (1–2 mm). The mean size (Mz), standard deviation (Sd), skewness (Sk), and kurtosis (Kg) were also calculated according to the equations devised by Folk [26]. The fractal dimension (D) values of the soil particle size distribution were calculated according to Tyler and Wheatcraft [27]. We used the soil quality index (SQI) to describe the soil quality. Soil quality is described by taking into account the nutrient status of the soil. According to previous studies [28,29,30], there are three main steps to calculate SQI. The three main steps in calculating the SQI are (1) making a selection of suitable parameters, (2) converting and weighing these parameters, and (3) combining the best parameters into one index. Prior to the principal component analysis (PCA), all data were standardized (Equation (1)), and KMO and Bartlett’s sphericity tests were performed. The mean SQI was calculated using (Equation (2)).

$$X\prime =\frac{\mathrm{X}-\mathrm{Xmin}}{\mathrm{Xmax}-\mathrm{Xmin}}$$

(1)

$$\mathrm{SQI}={{\displaystyle \sum }}_{i=1}^{n}WiPCi$$

(2)

where X′, X, Xmin, and Xmax represent the normalized value, actual value, minimum value, and maximum value, respectively; PCi is the PC selected, and Wi is the weight of the PC in the PCA [28].

The SQI is a comprehensive indicator of reflecting soil quality. According to the relevant studies mentioned [31]. We performed PCA on 10 soil nutrient indicators and extracted the principal components that could explain the variation in the observed values. For each PC, the appropriate parameter is >90% of the maximum loading value [32]. When significant correlations between these parameters were detected by Pearson correlation analysis, the component with the highest loading value was selected to construct each PC. We calculated the coefficients in the SQI formula using the following equation (coefficient = $\frac{\mathrm{the} \mathrm{highest} \mathrm{loading} \mathrm{value}}{\sqrt{eigenvalue}}\times \frac{eigenvalue}{\sum eigenvalue}$) [28].

2.6. Statistical Analysis

One-way ANOVA and the Games–Howell test were used to determine whether there was statistical significance. The Shapiro–Wilk test was used to check the normality of the variables before performing the data analysis. For those variables that do not conform to the normal distribution, data transformation was performed by obtaining the natural logarithm, square root, and derivative to make them conform to the normal distribution. The LSD procedure, as well as Duncan’s multiple comparison probability level of 0.05 for comparison and contrast, were used to identify significant differences. Two-way ANOVA was used to test the effects of dune type (T), windward slope sites (D), and their two-way interactions on soil nutrients. Redundancy analysis (RDA) was applied to analyze the relationship between soil nutrients and physical properties. Soil nutrients were considered the dependent variable, while physical factors were the explanatory variables. In addition, we identify environmental factors that have a significant effect on soil nutrients.

3. Results and Analysis

3.1. Composition and Distribution of Sand Dunes Particle Size

We summarize the particle size fractions and parameters of samples from the three types of dunes in

Table 2. Characteristics of soil particles and parameter values in different sand dunes.

Soil Layer	Types	Soil Particle Size Content (%)			Mz	Sd	Sk	Kg	D value	Determination Coefficient (R2)
		Clay	Silt	Sand
0–2	SL	0.75 ± 0.11 a	4.42 ± 0.77 a	94.80 ± 0.86 a	2.22 ± 0.10 b	0.88 ± 0.15 a	0.15 ± 0.08 a	1.60 ± 0.31 a	2.25 ± 0.02 a	0.92
	HDPE	0.67 ± 0.08 a	3.75 ± 0.96 a	95.35 ± 1.48 a	2.39 ± 0.06 a	0.77 ± 0.17 a	0.13 ± 0.10 a	1.46 ± 0.35 a	2.22 ± 0.02 b	0.91
	CK	0.29 ± 0.05 b	1.34 ± 0.40 b	98.34 ± 0.43 a	2.11 ± 0.09 c	0.51 ± 0.05 b	0.11 ± 0.05 a	1.06 ± 0.09 b	2.06 ± 0.03 c	0.87
2–4	SL	0.93 ± 0.12 a	2.60 ± 0.94 b	96.50 ± 1.02 a	2.34 ± 0.12 b	0.66 ± 0.12 a	0.09 ± 0.09 b	1.19 ± 0.18 a	2.26 ± 0.03 a	0.87
	HDPE	0.83 ± 0.12 a	3.89 ± 1.20 a	95.27 ± 1.32 a	2.47 ± 0.13 a	0.74 ± 0.07 a	0.21 ± 0.07 a	1.31 ± 0.11 a	2.25 ± 0.03 a	0.92
	CK	0.28 ± 0.02 b	1.08 ± 0.12 c	98.62 ± 0.14 a	2.09 ± 0.07 c	0.48 ± 0.03 b	0.08 ± 0.01 b	1.01 ± 0.02 b	2.05 ± 0.01 b	0.85

Note: Values are means ±SD (n = 15, from the bottom of the dune to the top). SL represents SL barrier dunes, HDPE represents HDPE barrier dunes, and CK represents no-barrier dunes. Different lowercase letters represent significant difference between dunes (p < 0.05, LSD test).

and Figure 4 (details in Appendix A 

Table A1. Proportion (%) of soil PSD in the 0–2 cm soil layer at different dune locations.

Soil Layer	Types	Position	Soil Particle Size Content (%)
			Clay	Silt	Very Fine Sand	Fine Sand	Medium Sand	Coarse Sand	Very Coarse Sand
0–2 cm	SL	Top	0.65 ± 0.05 a	3.98 ± 0.13 a	3.39 ± 0.02 a	49.71 ± 0.51 b	37.77 ± 0.33 b	2.79 ± 0.31 a	1.70 ± 0.28 a
		Upper middle	0.81 ± 0.06 a	4.78 ± 0.11 a	2.91 ± 0.04 b	52.76 ± 0.59 b	33.59 ± 0.39 a	3.47 ± 0.16 a	1.67 ± 0.24 ab
		Middle	0.78 ± 0.05 a	4.18 ± 0.10 b	2.65 ± 0.07 b	71.23 ± 1.56 a	16.26 ± 0.31 c	2.06 ± 0.77 a	2.83 ± 1.03 a
		Lower middle	0.74 ± 0.05 a	3.48 ± 0.05 a	2.20 ± 0.04 b	56.86 ± 0.26 c	31.52 ± 0.58 b	1.74 ± 0.41 a	1.45 ± 0.10 a
		Bottom	0.91 ± 0.06 a	5.69± 0.05 a	3.49 ± 0.05 a	49.25 ± 0.12 c	33.30 ± 0.66 a	4.96 ± 0.62 a	2.40 ± 0.15 a
	HDPE	Top	0.62 ± 0.01 a	3.92 ± 0.01 a	3.32 ± 0.11 a	63.19 ± 0.33 a	26.05 ± 0.54 c	1.67 ± 0.14 b	1.23 ± 0.27 ab
		Upper middle	0.71 ± 0.06 a	3.75 ± 0.14 b	4.09 ± 0.05 a	69.31 ± 0.88 a	16.98 ± 0.77 c	2.32 ± 0.63 a	2.83 ± 0.81 a
		Middle	0.82 ± 0.08 a	5.21 ± 0.35 a	5.03 ± 0.21 b	60.88 ± 2.55 b	19.83 ± 0.92 b	4.10 ± 1.81 a	4.12 ± 2.45 a
		Lower middle	0.80 ± 0.05 a	3.25 ± 0.08 a	4.72 ± 0.08 a	72.37 ± 0.09 a	16.88 ± 0.19 c	0.92 ± 0.21 b	1.06 ± 0.11 b
		Bottom	0.64 ± 0.06 b	2.34 ± 0.08 b	3.22 ± 0.04 b	70.69 ± 0.16 a	21.29 ± 0.40 c	0.80 ± 0.41 b	1.01 ± 0.17 c
	CK	Top	0.21 ± 0.02 b	0.79 ± 0.01 b	0.69 ± 0.03 b	48.36 ± 0.20 c	48.61 ± 0.33 a	0.28 ± 0.09 c	1.02 ± 0.02 b
		Upper middle	0.32 ± 0.02 b	1.11 ± 0.01 c	0.51 ± 0.01 c	67.52 ± 0.24 c	28.79 ± 0.16 b	0.57 ± 0.16 a	1.17 ± 0.12 b
		Middle	0.30 ± 0.02 b	2.01 ± 0.02 c	2.64 ± 0.05 a	48.10 ± 0.28 c	44.93 ± 0.52 a	0.59 ± 0.08 a	1.40 ± 0.24 a
		Lower middle	0.31 ± 0.01 b	1.43 ± 0.01 b	1.07 ± 0.03 c	60.29 ± 0.15 b	35.33 ± 0.22 a	0.64 ± 0.15 b	0.91 ± 0.12 b
		Bottom	0.34 ± 0.01 c	1.33 ± 0.02 c	0.78 ± 0.04 c	63.66 ± 0.41 b	31.12 ± 0.53 b	1.03 ± 0.05 b	1.69 ± 0.20 b

Note: Values are means ±SD (n = 3). SL represents SL barrier dunes, HDPE represents HDPE barrier dunes, and CK represents no-barrier dunes. Different lowercase letters represent significant difference between dunes (p < 0.05, LSD test).

and

Table A2. Proportion (%) of soil PSD in the 2–4 cm soil layer at different dune locations.

Soil Layer	Types	Position	Soil Particle Size Content (%)
			Clay	Silt	Very Fine Sand	Fine Sand	Medium Sand	Coarse Sand	Very Coarse Sand
2–4 cm	SL	Top	1.07 ± 0.07 a	4.31 ± 0.04 a	3.52 ± 0.07 b	63.26 ± 0.69 a	25.39 ± 0.19 b	1.23 ± 0.09 b	1.21 ± 0.37 a
		Upper middle	0.78 ± 0.05 a	1.61 ± 0.05 b	2.16 ± 0.04 b	63.98 ± 0.23 c	27.28 ± 0.16 a	2.76 ± 0.10 a	1.44 ± 0.25 a
		Middle	0.96 ± 0.04 a	2.09 ± 0.06 b	2.93 ± 0.02 b	81.44 ± 0.25 a	10.79 ± 0.16 c	0.73 ± 0.10 b	1.07 ± 0.24 b
		Lower middle	0.94 ± 0.05 a	2.16 ± 0.04 b	2.33 ± 0.33 b	73.71 ± 0.48 a	19.12 ± 0.26 b	0.86 ± 0.50 a	0.88 ± 0.25 a
		Bottom	0.85 ± 0.06 a	2.75 ± 0.15 a	2.66 ± 0.06 b	56.89 ± 1.71 c	30.10 ± 0.92 c	4.42 ± 1.65 a	2.32 ± 0.73 a
	HDPE	Top	0.90 ± 0.07 b	4.02 ± 0.09 b	6.99 ± 0.10 a	60.25 ± 0.43 b	24.68 ± 0.69 b	1.61 ± 0.14 a	1.56 ± 0.48 a
		Upper middle	0.90 ± 0.06 a	3.92 ± 0.10 a	5.97 ± 0.09 a	70.45 ± 0.39 a	16.49 ± 0.38 b	1.04 ± 0.13 b	1.23 ± 0.19 ab
		Middle	0.85 ± 0.05 a	2.94 ± 0.06 a	4.06 ± 0.12 a	62.04 ± 0.46 b	26.18 ± 0.82 b	1.55 ± 0.29 a	2.38 ± 0.30 a
		Lower middle	1.11 ± 0.05 a	5.83 ± 0.11 a	9.13 ± 0.06 a	74.23 ± 1.03 a	8.61 ± 0.47 c	0.24 ± 0.17 a	0.84 ± 0.45 a
		Bottom	0.74 ± 0.10 a	2.37 ± 0.09 b	5.14 ± 0.05 a	71.90 ± 0.90 a	16.87 ± 0.26 b	1.21 ± 0.51 a	1.75 ± 0.49 a
	CK	Top	0.26 ± 0.01 c	0.87 ± 0.01 c	0.70 ± 0.02 c	51.51 ± 0.23 c	44.95 ± 0.22 a	0.30 ± 0.15 c	1.39 ± 0.20 a
		Upper middle	0.32 ± 0.02 b	1.11 ± 0.01 c	0.49 ± 0.01 c	69.07 ± 0.71 b	27.93 ± 0.76 a	0.24 ± 0.06 c	0.82 ± 0.30 b
		Middle	0.29 ± 0.01 b	1.24 ± 0.01 c	0.76 ± 0.04 c	55.51 ± 0.21 c	40.39 ± 0.31 a	0.51 ± 0.24 b	1.29 ± 0.14 b
		Lower middle	0.28 ± 0.01 b	1.09 ± 0.02 c	0.99 ± 0.03 c	53.91 ± 0.11 b	41.46 ± 0.27 a	0.86 ± 0.30 a	1.41 ± 0.10 a
		Bottom	0.29 ± 0.01 b	1.06 ± 0.01 c	0.41 ± 0.01 c	61.71 ± 0.16 b	35.19 ± 0.40 a	0.19 ± 0.10 a	1.11 ± 0.18 a

Note: Values are means ±SD (n = 3). SL represents SL barrier dunes, HDPE represents HDPE barrier dunes, and CK represents no-barrier dunes. Different lowercase letters represent significant difference between dunes (p < 0.05, LSD test).

). Overall, significantly different results in terms of soil particle content and parameters were observed for no-barrier dunes and dunes with checkerboard barriers (Table 2) (p < 0.05). For all dunes, the Mz of the aeolian sands ranged from 2.10 Φ to 2.43 Φ. The most dominant soil particles were fine and medium sands, ranging from 48.10% to 81.44% and 10.79% to 44.95% of the samples, respectively. Clay, silt, coarse sand, and very coarse sand made up only a small portion. The contents of clay, silt, and very fine sands in SL barrier dunes and HDPE barrier dunes were higher than in no-barrier dunes (CK). In addition, they were more abundant in SL barrier dunes and HDPE barrier dunes in the different soil layers from the bottom to the top part of the dunes than in the no-barrier dunes in the corresponding positions (p < 0.05).

The Sd of the dune sands ranged from 0.48 Φ to 0.51 Φ, with a mean of 0.49 ± 0.04 in no-barrier dunes, and the sands were mostly well sorted. Sk shows that the no-barrier dune sands ranged from 0.08 to 0.11, with an average of 0.09 ± 0.04. No-barrier dunes were nearly symmetrically distributed. Kg ranged from 1.01 to 1.06 and averaged 1.04 ± 0.07, indicating a mesokurtic distribution. Sand was moderately sorted over SL barrier dunes (mean, 0.77 ± 0.17 Φ) and HDPE barrier dunes (mean, 0.76 ± 0.13 Φ). SL barrier dunes (mean, 0.12 ± 0.09) and HDPE barrier dunes (mean, 0.17 ± 0.09) were mostly fine-skewed, and those represented a leptokurtic distribution. In addition, we calculated the fractal dimension using soil particles. We found that the D values of SL barrier dunes were significantly higher than those of HDPE barrier dunes and no-barrier dunes (p < 0.05).

Figure 4 shows the particle size trends for the windward slopes of the three types of dunes. From bottom to top, there are five positions. At each position, three representative dunes of each type of dune were selected to depict the particle size distribution of the sand.

The particle distribution of SL barrier dunes is mainly characterized by becoming finer to the middle of the dune and then becomeing coarser towards the top of the dune. The Mz in the subsurface (2–4 cm soil layer) is finer grained than the surface (0–2 cm) soil. The Mz of the surface (0–2 cm) soil becomes finer as it moves from the bottom of the dunes (2.17 Φ) to the top of the dunes (2.13 Φ). The dunes of surface soil were moderately sorted but well sorted in subsurface soil. The Sk of subsurface soil showed variation from the bottom, which had a symmetrical distribution (−0.007), to the top, which was fine-skewed (0.24), but the Sk of surface soil showed slight variation from the bottom (0.13) to the top (0.20), which were mostly fine-skewed. The Kg of surface soil displayed a constant leptokurtic trend along the windward slope, and kurtosis was almost constantly in a leptokurtic distribution over the subsurface of the dune, with values from 1.32 to 1.46.

The particle distribution of HDPE barrier dunes tended to gradually become finer but suddenly became coarser at the top of the dune, producing a symmetrical M-shaped distribution. Similar to SL barrier dunes, the Mz of HDPE barrier dunes exhibited coarser surface soils than in subsurface soils, and they were poorly sorted. The dunes’ surface soil showed a fine-skewed grain size distribution, with little variation among positions on the dune, and weas mostly leptokurtic (1.14–1.32) from bottom to top. The surface soil of the dune mostly had a fine-skewed grain-size distribution (0.12 to 0.21), and the skewness increased (from 0.12 to 0.21) towards the dune’s top (0.21). The Kg showed a trend from the bottom (1.07) to middle (1.87), changed from a leptokurtic distribution to a very leptokurtic distribution, and then decreased from 1.87 at the middle to 1.39 at the top, with the top showing a leptokurtic distribution.

3.2. Soil Nutrients and Soil Quality Index Characteristics of Sand Dunes

3.2.1. Characteristics of Soil Nutrients

The soil nutrient contents of windward slopes of sand dunes differed at varying degrees among treatments of barriers (

Table 3. Soil nutrients and ecological stoichiometry characteristics of different types of dunes.

Soil Nutrients	Types			Mean	Cv
	SL	HDPE	CK
SOC (g∙kg−1)	3.06 ± 1.41 a	2.40 ± 0.60 a	0.65 ± 0.08 b	2.04 ± 1.35	66.27
AP (mg∙kg−1)	1.24 ± 0.22 a	1.18 ± 0.31 a	0.80 ± 0.13 b	1.07 ± 0.30	27.93
AK (mg∙kg−1)	82.67 ± 9.12 a	71.73 ± 15.42 b	56.03 ± 2.71 c	70.14 ± 15.13	21.57
AN (mg∙kg−1)	11.90 ± 4.56 b	19.14 ± 6.71 a	6.66 ± 2.41 c	12.56 ± 7.08	56.31
TP (g∙kg−1)	0.40 ± 0.04 a	0.33 ± 0.08 b	0.32 ± 0.04 b	0.35 ± 0.07	19.59
TK (g∙kg−1)	24.14 ± 2.28 a	20.37 ± 2.55 b	20.05 ± 1.89 b	21.52 ± 2.92	13.57
TN (g∙kg−1)	0.20 ± 0.08 a	0.21 ± 0.05 a	0.04 ± 0.002 b	0.15 ± 0.10	64.15
C/N	14.79 ± 2.01 b	11.25 ± 1.63 c	16.81 ± 2.13 a	14.28 ± 3.00	21.05
C/P	7.52 ± 3.12 a	8.52 ± 5.84 a	2.05 ± 0.26 b	6.03 ± 4.77	79.04
N/P	0.50 ± 0.18 b	0.75 ± 0.47 a	0.12 ± 0.01 c	0.46 ± 0.39	84.97

Note: Values are means ±SD (n = 15, from the bottom of the dune to the top). SL represents SL barrier dunes, HDPE represents HDPE barrier dunes, and CK represents no-barrier dunes. Different lowercase letters represent significant difference between dunes (p < 0.05, LSD test). Abbreviations: soil organic carbon (SOC), total phosphorus (TP), total potassium (TK), total nitrogen (TN), alkaline hydrolysis nitrogen (AN), available phosphorus (AP), available potassium (AK), carbon/nitrogen ratio (C/N), carbon/phosphorus ratio (C/P), nitrogen/phosphorus ratio (N/P).

). Soil nutrients differed significantly among dune types (p < 0.01) (

Table 4. ANOVA results for soil nutrients and ecological stoichiometry influenced by dune type (T), location (D), and their interactions (T × D).

	SOC	AN	AP	AK	TN	TP	TK	C/N	C/P	N/P
F
T	103.76	61.03	36.91	95.49	98.12	38.05	144.22	135.10	113.74	120.50
D	1.12	2.15	2.83	8.07	1.98	6.13	38.22	5.30	2.00	2.33
T × D	2.37	1.98	2.45	4.59	1.51	5.52	24.53	9.35	3.83	2.54
P
T	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
D	0.352	0.083	0.030	<0.001	0.107	<0.001	<0.001	0.001	0.103	0.064
T × D	0.025	0.060	0.021	<0.001	0.169	<0.001	<0.001	<0.001	0.001	0.017

Abbreviations: soil organic carbon (SOC), total phosphorus (TP), total potassium (TK), total nitrogen (TN), alkaline hydrolysis nitrogen (AN), available phosphorus (AP), available potassium (AK), carbon/nitrogen ratio (C/N), carbon/phosphorus ratio (C/P), nitrogen/phosphorus ratio (N/P).

). Except for soil SOC, AP, TN, AN, and C/P, the contents of other soil nutrients in SL barrier dunes were higher than those in HDPE and no-barrier dunes (p < 0.05) (Table 3). The soil AK and TP of the SL barrier dunes reached 82.67 mg∙kg⁻¹ and 0.40 g∙kg⁻¹, respectively. However, it is worth mentioning that the soil SOC content of SL barrier dunes and HDPE barrier dunes were 3.06 and 2.40 g∙kg⁻¹, respectively. The SOC content of barrier dunes was 369% to 470% higher than that of no-barrier dunes. The soil AN of SL barrier dunes and HDPE barrier dunes reached 11.90 and 19.14 mg∙kg⁻¹, respectively, and were 178% to 287% higher than that of no-barrier dunes. Among different dune types, soil nutrients, which had medium variation, displayed coefficients of variation of 13.57% and 84.97%.

The distribution of nutrients in different positions of the soil under different types of dunes had evident gradation (accompanying diagram Figure A1 and Figure A2), and soil nutrient content generally followed a trend of 0–2 cm > 2–4 cm. Except for soil AN, AP, and N/P, other soil nutrient contents of the same location in the 0–2 cm layer followed the order of SL barrier dunes > HDPE barrier dunes > no-barrier dunes. The trends of TP, TK, AP, and AK contents of soils in the same position of different types of dunes in the 2–4 cm soil layer were also similar to those in the 0–2 cm layer. Soil AP was significantly (p < 0.05), and AK, TP, TK, and C/N were extremely significantly (p < 0.001) affected at different locations on the windward slope of the dune. In the two types of checkerboard barrier dunes from top to bottom, the 0–2 cm soil layer showed roughly high enrichment of soil nutrients at the top or bottom except for soil AN. However, in the 2–4 cm soil layer, soil AK, TP, TK, and TN contents showed enrichment in the dunes’ middle parts. The interaction (T × D) between checkerboard barrier types and dune position significantly affected soil SOC, AP, and N/P (p < 0.05) and extremely significantly affected AK, TP, TK, C/N, and C/P (p < 0.001).

3.2.2. SQI Characteristics of Different Types of Dunes

The distribution of SQI values for different types of dune soils is shown in Figure 5. The total SQI values of different types of dunes followed the order of SL (0.22) > HDPE (0.14) > CK (0.12). Previously, the comprehensive evaluation of SQI was classified into very high level (>0.8), high level (0.6–0.8), medium level (0.4–0.6), low level (0.2–0.4), and very low level (<0.2) according to the equal interval method [28,33]. The SQI of SL barrier dunes in the study area was low, while the SQI of other dune types was very low. Overall, the SQI levels in the study area were mostly low.

3.3. Relationships between Soil Nutrients and Parameters

According to RDA analysis, 97.78% of the variation in soil nutrients in this study could be explained by particle size characteristics combined. Tests on the first axis (pseudo-F = 50.6, p < 0.01) and all axes (pseudo-F = 14.4, p < 0.01) were significant, thus indicating that the results of the RDA analysis were credible. The first two canonical axes explained 65.48% and 15.69% of the variation in soil nutrients, with correlation coefficients of 0.9166 and 0.7111 for the studied particle size characteristics and soil nutrients, respectively. The simple-term effects showed that Silt (explained 57.4%, p < 0.01), Kg (explained 16.4%, p < 0.01), Mz (explained 9.7%, p < 0.01), and Sk (explained 6.9%, p < 0.01) could significantly explain the changes in soil nutrients.

The angle between other particle size characteristics except sand and soil nutrients indices is less than 90° in Figure 6, proving that there is a positive correlation between these two parameters—the smaller the angle, the stronger the correlation. The arrow angle between environmental factors and soil nutrients is greater than 90°, which proves that there is a negative correlation between these two parameters. Most of the arrow angles between environmental factors and soil nutrients were less than 90°, indicating a positive correlation between these environmental factors and soil nutrients. The angles between Sand and the other soil nutrient indicators were greater than 90° except for soil C/N. Sk was negatively correlated with AK, TP, and C/N. This was also confirmed by the Pearson correlation results (

Table 5. Pearson correlation coefficients between soil nutrients, ecological stoichiometry, and particle size characteristics.

	Silt	Clay	Sand	Mz	Sd	Sk	Kg	D
SOC	0.692 **	0.694 **	−0.700 **	0.324 **	0.817 **	0.364 **	0.749 **	0.725 **
AN	0.478 **	0.376 **	−0.391 **	0.537 **	0.521 **	0.206	0.527 **	0.476 **
AP	0.601 **	0.437 **	−0.462 **	0.526 **	0.290 **	0.138	0.164	0.611 **
AK	0.711 **	0.627 **	−0.644 **	0.406 **	0.404 **	−0.136	0.351 **	0.617 **
TN	0.319 **	0.414 **	−0.406 **	−0.085	0.340 **	0.217 *	0.234 *	0.329 **
TP	0.453 **	0.407 **	−0.418 **	0.142	0.202	−0.111	0.050	0.414 **
TK	0.772 **	0.751 **	−0.760 **	0.483 **	0.866 **	0.372 **	0.785 **	0.789 **
C/N	−0.550 **	−0.437 **	0.454 **	−0.641 **	−0.366 **	−0.073	−0.255 *	−0.512 **
C/P	0.538 **	0.491 **	−0.501 **	0.419 **	0.653 **	0.176	0.682 **	0.571 **
N/P	0.548 **	0.487 **	−0.498 **	0.504 **	0.638 **	0.153	0.665 **	0.569 **

* indicates significant at the 0.05 level, ** at the 0.01 level (double-tailed test). Abbreviations: soil organic carbon (SOC), total phosphorus (TP), total potassium (TK), total nitrogen (TN), alkaline hydrolysis nitrogen (AN), available phosphorus (AP), available potassium (AK), carbon/nitrogen ratio (C/N), carbon/phosphorus ratio (C/P), nitrogen/phosphorus ratio (N/P), fractal dimension values (D values).

).

4. Discussion

4.1. Soil Particle Distribution of Dunes

Soil particle distribution plays a vital role in influencing soil physicochemical properties, which can affect soil nutrients such as soil C, N, and P [18,28,34]; therefore, soil particle distribution has also been used to evaluate soil quality. The particle size distribution on a dune’s surface is a product of the interaction between the dune and airflow, resulting in the dune indicating changes in sand transport rate and erosion pattern. In contrast, the erosion state of the windward slope of the dune mainly depends on the wind shear speed [35,36,37]. For the wandering dunes in the study area, the streamline on the windward slope of the dune is compressed, and the airflow was accelerated and gradually increased along the bottom of the slope, reaching its maximum at the top [38,39,40]. Thus, wind erosion was the strongest on the windward slope, and the fine particle fraction of saltation was lost with the strengthening of erosion, making the particle size coarser. Therefore, the Mz of the no-barrier dunes gradually decreases from the bottom to the top. The content of fine particles at the top was lesser, while being well sorted at each position of the windward slope of the dune, and the sorting at the top was better. The particle size distribution shows a positive skewed distribution and moderate kurtosis (e.g., [41]).

After a checkerboard barrier is arranged, it can influence the soil particle distribution. By increasing the surface roughness [42], the checkerboard barrier disturbed the wind flow in the area around the barrier. It changed the pattern of airflow for secondary distribution of airflow. The acceleration zone formed in the barrier reduced or prevented the accumulation of sand grains in the barrier [43], which usually resulted in the accumulation of sand before and after the barrier and wind erosion in the middle of the sand barrier [18]. This accumulation intercepts the saltation component of sand grains and the creep component, so the accumulation of sand grains presents a diversity of grain levels. In this study, the fine particle content of soil particles was higher in both the 0–2 cm soil layer and the 2–4 cm soil layer in the SL barrier dunes and HDPE barrier dunes than in the no-barrier dunes, which also made the sand particle sorting on the windward slope of the barrier dunes worse compared to the no-barrier dune. Among them, the bottom position of the SL barrier dune had the highest fine particle content and was also consistent with the distribution of Mz [6]. The wind speed at the bottom of SL barrier dunes was much lower than that of the sand-laden wind, so the fine particle components such as very fine sand, clay, and silt particles were difficult to erode. The airflow accelerated to the middle of the slope in the climbing process, and the wind protection effectiveness of the checkerboard barrier at this time was less than that of the sand-laden wind. The shear stress gradually increased and became unstable and unsaturated. Some fine particle components began to be eroded, resulting in an increase in the content of fine sand and medium sand from the middle to the top of the slope [44]. Unlike the conventional checkerboard barrier with a certain degree of permeability, the HDPE barrier was a mesh checkerboard barrier with a certain degree of permeability, which has a lifting effect on the airflow [45], accelerates the airflow convergence, and causes stronger wind erosion in the middle and bottom position of the windward slope of the dune. The erosion force of unsaturated wind-drift sand was stronger. When the wind-drift sand reached the top of the dune, the coarse particles settled and accumulated due to the interception effect of the checkerboard barrier and the reduction of the sand carrying capacity of the near-saturated wind-drift sand [44]. Therefore, the Mz at the bottom, middle, and top of the dune was lower.

4.2. Soil Nutrient Characteristics of Dunes

Soil SOC and available and total nutrient contents can be used to assess the effect of checkerboard barriers on soil texture [25,32,46]. The setting of checkerboard barriers enables the nutrients carried by rainfall and dust to accumulate and decompose in the process of fixing. The deposition and accumulation process of fine particles will enhance the mutual cementation between soil particles, promote the development of physical soil crust, reduce the loss of soil nutrients in the sandy environment, and facilitate the accumulation of soil’s available nitrogen and total nitrogen in the surface layer of HDPE barrier dunes [17,47,48,49]. Some scholars have also found that SL checkerboard barriers have been irreversibly damaged over time due to microbial or fungal influences on the biological structure and chemical composition of the sand-buried section. The structure and composition can alter the nutrient characteristics of the soil [49,50,51,52]. In this study, soil nutrients were higher in the bottom position of the checkerboard barrier dunes than in other parts of the dunes because the bottom position of the SL barrier dunes and HDPE barrier dunes belonged to the weak wind erosion zone compared with the middle and top parts of the dunes, which was more conducive to soil nutrient enrichment. Dai [6] also confirmed this by studying soil SOC in different position of dunes. Although soil SOC, available nutrients, and total nutrients can reflect soil fertility [53,54], it is impossible to accurately determine soil quality based on specific soil properties because of the interdependence between these properties [55]. Therefore, the SQI was established to assess the soil comprehensively, and the SQI can effectively evaluate the impact of checkerboard barriers on soil quality [31,56]. The SQI of SL barrier dunes was higher than that of HDPE barrier dunes, which was also consistent with their nutrient contents.

Soil and vegetation are two integral parts of the ecosystem [25]. Checkerboard barriers improve the soil on the one hand and intercept plant seeds carried by wind–sand flows on the other, facilitating the establishment of dune seed banks [57,58,59]. Improving the aeolian environment through checkerboard barriers provides protection for plant growth, and the improving soil quality positively contributes to plant growth [60]. The relationship between checkerboard barriers, soil, and plants develops into a benign ecological restoration strategy [9]. Based on the above results, given the relatively low material and construction costs of SL checkboard barriers and whether it was increasing the content of fine particles or increasing the soil nutrient content, the effect of the SL checkboard barrier was better than that of the HDPE checkerboard barrier. Scholars also have made some conclusions that the D value decreases with the increase of desertification, which can effectively evaluate the change in soil quality caused by desertification [61,62,63]. In this paper, the D values of SL barrier dunes were significantly higher (p < 0.05) than those of HDPE barrier dunes and no-barrier dunes. Therefore, the environmental improvement effect of the SL checkerboard barrier was better than the HDPE checkerboard barrier. The laying of checkerboard barriers facilitates the reversal of desertification.

5. Conclusions

In arid or semi-arid ecosystems, checkerboard barriers play an essential role in restoring sandy soils and facilitate the establishment of seed banks of dune plants [9]. Checkerboard barrier placement has a positive effect on the accumulation of fine soil particles and improvement of soil quality, ultimately leading to an increase in D value. Specifically, soil particles in the study area are mainly fine and medium sands. The clay, silt, and very fine sand components in the 0–2 cm soil layer and 2–4 cm soil layer were significantly enhanced in the Salix psammophila (SL) and high-density polyethylene (HDPE) barrier dunes compared with the no-barrier dunes. The fine particle components tend to be concentrated in the bottom or middle and lower parts of the dunes. HDPE checkerboard barriers are beneficial to the enrichment of soil AN and TN content, while other soil nutrient contents are higher in SL checkerboard barrier dunes than HDPE checkerboard barrier dunes, and soil nutrients are easily enriched in the bottom or lower middle part of the dune. When SQI was used as an indicator of soil quality for artificial sand fixation measures in this area, the soil SQI of SL barrier dunes was higher than that of HDPE barrier dunes. There was a significant correlation between soil nutrients and particle size characteristics in our study. Soil grain characteristics and soil nutrient content have an interactive relationship. The increase in D values also proves that the laying of checkerboard barriers facilitated the reversal of the desertification process. Therefore, installing SL checkerboard barriers on dunes has significant ecological restoration benefits.