In the context of global carbon neutrality goals, the development of renewable energy has become a critical strategy for green economic transformation. Desert regions, with their vast land areas and abundant solar resources, offer ideal conditions for large-scale photovoltaic projects. However, the construction and operation of solar panels in sandy environments often lead to secondary sand hazards, such as erosion and deposition around the panels, which can impair their efficiency and longevity. To mitigate these issues, low vertical sand barriers are employed between photovoltaic arrays to stabilize the surface and reduce wind-blown sand movement. This study investigates the sediment grain size characteristics under three types of low vertical sand barriers—polylactic acid (PLA) barriers, gauze barriers, and straw checkerboard barriers—deployed between solar panels in the Hobq Desert. The objective is to evaluate their effectiveness in altering soil texture and enhancing surface stability, thereby providing insights for sand hazard control in photovoltaic farms.
The research was conducted in the Hobq Desert, located in northern China, which features a typical temperate continental monsoon climate with significant diurnal temperature variations. The area receives an average annual precipitation of 258 mm, while evaporation rates can reach 2,400 mm. Strong winds, predominantly from the northwest, occur frequently between March and May, exacerbating sand movement. The local topography consists of mobile sand dunes with sparse vegetation cover. The photovoltaic farm where this study was carried out became operational in 2018, and prior to installation, the land was leveled to create a relatively flat surface. The soil is primarily composed of loose aeolian sand, rich in medium and fine sand fractions. For this investigation, we selected three parallel photovoltaic arrays with panels spaced 9 meters apart. Each array measured 15 meters in length, with the lower edge of the solar panels positioned 50 cm above the ground and the upper edge at 200 cm. Between these arrays, we installed three types of low vertical sand barriers in a 1 m × 1 m grid pattern: PLA barriers (10–15 cm height), gauze barriers (15–20 cm height), and straw checkerboard barriers (15–20 cm height). A control site with no sand barriers was also established for comparison.
Sampling was conducted in September 2023 during a period of stable weather with no rainfall or strong winds for seven consecutive days. We employed a five-point sampling method within the central grid of each sand barrier type and the control area. Soil samples were collected from four depth intervals: 0–2 cm, 2–10 cm, 10–20 cm, and 20–30 cm, using a layered soil sampler. Three replicates were taken for each sampling point, resulting in a total of 240 samples. After collection, the samples were air-dried in the laboratory and prepared for grain size analysis. Organic matter was removed by treating the samples with a mixture of 45 mL of ultrapure water and 30% hydrogen peroxide, followed by 24 hours of incubation. Carbonates were dissolved using hydrochloric acid, and the pH was neutralized through repeated dilution. Grain size distribution was determined using a laser diffraction particle size analyzer (Analysette22 MicroTec Plus, Germany) with a measurement range of 0.01–3,800 μm.
Soil particle size was classified according to the USDA system into clay (< 2 μm), silt (2–50 μm), very fine sand (50–100 μm), fine sand (100–250 μm), medium sand (250–500 μm), coarse sand (500–1,000 μm), and very coarse sand (1,000–2,000 μm). The grain size parameters were calculated using the Folk and Ward graphical method, based on the cumulative frequency percentiles. The following formulas were applied to derive key parameters:
$$ \phi = -\log_2 D $$
where \( D \) is the particle diameter in mm. The mean grain size (\( d_0 \)) was calculated as:
$$ d_0 = \frac{1}{3} (\phi_{16} + \phi_{50} + \phi_{84}) $$
The standard deviation (\( \sigma_0 \)), representing sorting, was computed as:
$$ \sigma_0 = \frac{\phi_{84} – \phi_{16}}{4} + \frac{\phi_{95} – \phi_{5}}{6.6} $$
Skewness (\( S_0 \)), indicating the symmetry of the distribution, was given by:
$$ S_0 = \frac{\phi_{16} + \phi_{84} – 2\phi_{50}}{2(\phi_{84} – \phi_{16})} + \frac{\phi_{5} + \phi_{95} – 2\phi_{50}}{2(\phi_{95} – \phi_{5})} $$
Kurtosis (\( K_0 \)), describing the peakedness of the distribution, was derived as:
$$ K_0 = \frac{\phi_{95} – \phi_{5}}{2.44(\phi_{75} – \phi_{25})} $$
Additionally, the fractal dimension (\( D_A \)) of the soil particles was calculated using the volume-based model:
$$ \left( \frac{R_i}{R_{\text{max}}} \right)^{3 – D_A} = \frac{V(r < R_i)}{V_T} $$
where \( R_i \) is the particle diameter, \( R_{\text{max}} \) is the maximum particle size, \( V(r < R_i) \) is the volume percentage of particles smaller than \( R_i \), and \( V_T \) is the total volume percentage. Data analysis was performed using Excel and SPSS, with one-way ANOVA and LSD tests for multiple comparisons. Graphs were generated using Origin software.
The grain size composition of sediments under the different sand barriers revealed significant variations compared to the control. Overall, the soil texture across all barrier types was dominated by fine and medium sand, but the installation of sand barriers led to a noticeable increase in finer particles such as clay, silt, and very fine sand. The following table summarizes the average content of each grain size fraction across the 0–30 cm depth profile:
| Grain Size Fraction | Control (%) | PLA Barrier (%) | Gauze Barrier (%) | Straw Checkerboard Barrier (%) |
|---|---|---|---|---|
| Clay (< 2 μm) | 0.45 | 0.98 | 1.52 | 0.87 |
| Silt (2–50 μm) | 1.32 | 2.15 | 3.25 | 2.41 |
| Very Fine Sand (50–100 μm) | 2.89 | 3.67 | 4.72 | 3.51 |
| Fine Sand (100–250 μm) | 45.63 | 46.89 | 44.12 | 45.97 |
| Medium Sand (250–500 μm) | 48.71 | 46.38 | 44.14 | 46.24 |
| Coarse Sand (500–1,000 μm) | 0.95 | 0.82 | 1.15 | 0.92 |
| Very Coarse Sand (1,000–2,000 μm) | 0.05 | 0.11 | 0.10 | 0.08 |
The total content of clay, silt, and very fine sand was highest under the gauze barrier, reaching 8.43%, which was 123.02% higher than the control (3.78%). This indicates a significant refinement of soil texture due to the sand barriers, with the most pronounced effects observed in the 20–30 cm layer. For instance, the gauze barrier showed a 28.60% and 34.47% increase in very fine sand content compared to the PLA and straw checkerboard barriers, respectively. In contrast, the content of medium sand decreased under all barrier types, with reductions of 2.33%, 4.54%, and 2.41% for PLA, gauze, and straw checkerboard barriers, respectively. Coarse sand content was generally low but varied with depth, peaking in the 0–2 cm layer under the PLA barrier at 0.23% and decreasing with increasing depth.
The grain size parameters provided further insights into the distribution characteristics. The mean grain size (\( d_0 \)) increased under all sand barriers compared to the control, reflecting a shift towards finer particles. The average \( d_0 \) values over the 0–30 cm profile were 2.19φ for PLA barriers, 2.21φ for gauze barriers, and 2.18φ for straw checkerboard barriers, compared to 2.15φ for the control. This represents increases of 1.86%, 2.79%, and 1.39%, respectively. The standard deviation (\( \sigma_0 \)), which measures sorting, was significantly higher under the barriers, indicating poorer sorting compared to the control. The control had an average \( \sigma_0 \) of 0.67 (moderately well sorted), while the barriers ranged from 1.07 to 1.23 (poorly sorted). The fractal dimension (\( D_A \)), which describes the complexity of particle size distribution, also increased, with values of 2.14, 2.22, and 2.15 for PLA, gauze, and straw checkerboard barriers, respectively, versus 1.66 for the control. This suggests a more heterogeneous grain size distribution under the barriers.
Skewness (\( S_0 \)) and kurtosis (\( K_0 \)) values revealed differences in the symmetry and peakedness of the distributions. The control exhibited near-symmetrical distributions with \( S_0 \) around 0.06 and \( K_0 \) around 0.97. The gauze barrier showed positive skewness (0.14) and sharp peaks (\( K_0 = 1.11 \)), indicating a dominance of finer particles and a concentrated distribution. In contrast, the PLA and straw checkerboard barriers had near-symmetrical distributions (\( S_0 \) around 0.09) with medium kurtosis (\( K_0 \) around 1.02–1.03), approximating a normal distribution. The following table summarizes the grain size parameters for each barrier type across depths:
| Barrier Type | Depth (cm) | Mean Grain Size (\( d_0 \), φ) | Standard Deviation (\( \sigma_0 \)) | Skewness (\( S_0 \)) | Kurtosis (\( K_0 \)) | Fractal Dimension (\( D_A \)) |
|---|---|---|---|---|---|---|
| Control | 0–2 | 2.15 | 0.67 | 0.06 | 0.97 | 1.65 |
| 2–10 | 2.16 | 0.69 | 0.06 | 0.97 | 1.68 | |
| 10–20 | 2.11 | 0.68 | 0.06 | 0.97 | 1.68 | |
| 20–30 | 2.18 | 0.65 | 0.05 | 0.97 | 1.64 | |
| PLA Barrier | 0–2 | 2.15 | 1.07 | 0.09 | 1.00 | 2.14 |
| 2–10 | 2.17 | 1.05 | 0.09 | 1.01 | 2.14 | |
| 10–20 | 2.18 | 1.07 | 0.08 | 1.02 | 2.14 | |
| 20–30 | 2.27 | 1.10 | 0.11 | 1.03 | 2.16 | |
| Gauze Barrier | 0–2 | 2.16 | 1.26 | 0.13 | 1.13 | 2.22 |
| 2–10 | 2.21 | 1.25 | 0.14 | 1.09 | 2.22 | |
| 10–20 | 2.20 | 1.20 | 0.14 | 1.12 | 2.23 | |
| 20–30 | 2.28 | 1.19 | 0.13 | 1.09 | 2.22 | |
| Straw Checkerboard Barrier | 0–2 | 2.16 | 1.08 | 0.09 | 1.03 | 2.15 |
| 2–10 | 2.19 | 1.12 | 0.08 | 1.03 | 2.15 | |
| 10–20 | 2.13 | 1.15 | 0.09 | 1.04 | 2.15 | |
| 20–30 | 2.24 | 1.02 | 0.11 | 1.02 | 2.15 |
The frequency distribution curves of soil particle size under all sand barriers exhibited a unimodal pattern, with peaks occurring between 200 and 250 μm. However, the position and height of these peaks varied with depth and barrier type. For example, in the 0–2 cm layer, the peak volume percentages were 11.6% for PLA barriers, 10.92% for straw checkerboard barriers, and 10.68% for gauze barriers. As depth increased, the straw checkerboard barriers showed a gradual rise in peak height, while the gauze barriers maintained a relatively stable distribution. The cumulative frequency curves demonstrated that all barriers had steeper slopes in the finer particle range (0–100 μm) compared to the control, indicating better sorting of fine particles. The gauze barrier consistently showed the most pronounced changes, with cumulative frequencies slightly higher than the other barriers, particularly in the 0–10 cm layer. This suggests that the gauze barrier is more effective at trapping and retaining fine sediments.
The discussion of these findings centers on the interaction between solar panels, wind dynamics, and sand barriers. The deployment of photovoltaic arrays alters local wind fields, reducing near-surface wind speeds and promoting the deposition of fine particles. The sand barriers enhance this effect by disrupting airflow and creating turbulence that facilitates the settling of clay, silt, and very fine sand. The gauze barrier, with its porous structure, acts as a three-dimensional filter, efficiently capturing fine particles and leading to the highest refinement of soil texture. This aligns with previous studies that highlight the role of barrier porosity in modulating sediment transport. The positive skewness and sharp kurtosis under the gauze barrier indicate a selective accumulation of fine materials, which can improve soil stability and nutrient retention over time. In contrast, the PLA and straw checkerboard barriers, while effective, show less pronounced changes due to their different structural properties. The straw checkerboard, for instance, allows for some airflow, which may result in a more mixed grain size distribution. The increase in fractal dimension under all barriers reflects greater soil heterogeneity, which is beneficial for forming stable surface crusts and supporting vegetation establishment in photovoltaic farms.
From a practical perspective, these results have important implications for sand hazard management in photovoltaic projects. The refinement of soil texture under sand barriers reduces wind erosion potential and enhances the long-term sustainability of solar panels. Specifically, the gauze barrier emerges as the most effective type for promoting soil development and mitigating sand accumulation around photovoltaic installations. However, factors such as cost, durability, and environmental impact should also be considered when selecting sand barriers for large-scale applications. Future research could explore the combined effects of sand barriers and vegetation restoration in photovoltaic farms to further optimize ecological and energy benefits.
In conclusion, this study demonstrates that low vertical sand barriers significantly influence sediment grain size characteristics between solar panels in desert environments. All three barrier types—PLA, gauze, and straw checkerboard—promote soil texture refinement by increasing the content of fine particles and improving surface stability. The gauze barrier shows the most substantial effects, with higher clay, silt, and very fine sand contents, positive skewness, and sharp kurtosis. These changes contribute to reduced wind erosion and enhanced soil formation processes, supporting the integration of photovoltaic energy production with desert ecosystem management. As the global demand for renewable energy grows, such strategies will be crucial for the sustainable development of photovoltaic projects in arid regions.