Soil Carbon Sequestration by Biological Crusts in Photovoltaic Power Stations: Southern Tengger Desert and Artemisia ordosica Shrubland Restoration

Abstract

This study investigates the effects of different photovoltaic (PV) panel types on soil and biological soil crusts (BSCs) under vegetation restoration in sandy areas. A 150 MW PV power plant in Huanghuatan, located in the Tengger Desert, was selected as the research site. Soil and BSC properties, as well as carbon sequestration, were evaluated under three PV panel types: fixed-axis (FA); horizontal single-axis (HSA); and tilted single-axis (TSA). The objective was to clarify how these panel types influence soil quality and carbon storage during Artemisia ordosica Krasch. restoration in sandy environments and to explore the underlying mechanisms. The results showed that, compared with the surrounding pristine desert (PD), PV development significantly altered soil water content (WC), saturated water content (SWC), soil organic matter (SOM), and carbonate levels in soil and BSCs. Specifically: (1) FA and HSA panels increased WC in the BSCs and sub-crust soil, although water-holding capacity decreased in the HSA area; (2) SOM in the BSCs was notably lower under HSA and TSA panels; (3) HSA and TSA panels enhanced carbonate accumulation in non-crusted soil, while the lowest carbonate content in BSCs occurred under FA panels. The sub-crust soil in all PV areas had lower carbonate content than PD; and (4) Estimated carbon storage effectiveness was ranked as follows: HSA > TSA > PD > FA. This study provides theoretical support for ecological restoration in desert PV power plants.

1. Introduction

Global electricity generation relies heavily on fossil fuels, causing severe environmental problems. Solar energy, characterized by high efficiency and pollution-free operation [1,2,3], has become the most promising and fastest-growing renewable energy source [4,5,6]. By the end of 2022, China’s installed renewable energy capacity reached 1.27 billion kW. During the “14th Five-Year Plan” period, China’s newly added PV land area will be approximately 2400–3200 km²; the total PV land area will reach 585,000 km² by 2060. The Tengger Desert, China’s fourth-largest desert, covering approximately 42,700 km², is an ideal site for PV stations due to its abundant sunlight. However, the region has sparse vegetation and harsh environments, making it highly vulnerable to climate change and human activities. The construction of centralized PV stations alters land use types, causes drastic changes in surface morphology, and significantly affects regional microclimate, biological activities, and ecosystem functions such as carbon sequestration potential [7]. Therefore, studying the impacts of desert PV stations on soil environments is crucial for promoting the sustainable development of China’s PV industry and developing ecological restoration technologies.

Current research on the environmental impacts of PV stations has attracted significant attention. During construction, PV stations can damage soil structure and native vegetation, leading to losses of soil organic carbon (SOC), reduced nitrogen and phosphorus contents [8,9], and increased soil impoverishment. After construction, PV panels convert solar radiation into electricity, reducing solar radiation absorption by the surface, slowing the rate of temperature rise in the air (daily and annual temperature ranges) and on the surface [9,10]. Their shading effect reduces surface evaporation [10]; their barrier effect weakens wind speed and sediment transport, enhancing wind erosion prevention and sand fixation performance [11]. PV stations also mitigate surface salt accumulation, affecting SOC decomposition rates, soil respiration [9], and inorganic carbon transformation. As a dominant shrub in the Tengger Desert, Artemisia ordosica Krasch. enhances soil organic matter (SOM) through root sand fixation and litter input, forming a synergistic carbon sequestration effect with biological crusts. However, existing studies on soil carbon sequestration show inconsistent results. Research on how PV panel types affect interactions between Artemisia ordosica shrubland and biological crusts—especially how different bracket types (e.g., horizontal single-axis, tilted single-axis, fixed-axis) influence Artemisia growth and crust carbon sequestration by altering microenvironments—remains scarce, with no systematic conclusions.

Biological soil crusts (BSCs) are surface aggregates formed by cyanobacteria, lichens, mosses, and other microorganisms and cryptogams that bind surface soil particles via hyphae, rhizoids, and polysaccharides [12]. They cover 40%–70% of arid and semi-arid surfaces and are key indicators of desert ecosystem health. BSCs dominated by algae, lichens, and mosses are widely distributed in the Tengger Desert, effectively improving soil structure and function, enhancing stability and productivity, increasing surface water-holding capacity, promoting carbon and nitrogen cycles, and facilitating vascular plant colonization [13,14]. However, shading by PV panels and competition with Artemisia ordosica may inhibit BSC photosynthesis, affecting their carbon sequestration contribution [15,16,17]. Currently, research on how PV stations—which drastically alter desert environments—influence the properties of BSCs is limited. Thus, exploring the impacts of PV stations with Artemisia ordosica as the main restored vegetation on BSCs is important for evaluating the environmental effects of PV systems.

Our study was conducted at the 150 MW PV power demonstration zone in Huanghuatan, Gulang County, Wuwei City, Gansu Province. Using the surrounding pristine desert as a control (PD), we investigated the physicochemical properties of soil and BSCs under three photovoltaic mounting configurations (horizontal single-axis (HSA), tilted single-axis (TSA), and fixed-axis (FA)) in a sand-fixing Artemisia ordosica-vegetated PV plant. Specifically, the present study tested the following hypotheses: (1) PV construction alters the characteristics of soils and BSCs within the power station; and (2) different PV panel types differentially affect the physicochemical properties and carbon storage capacity of soils and BSCs. The aim is to identify the optimal PV configuration for enhancing carbon fixation and soil quality in Artemisia ordosica-restored desert PV systems. The findings will provide critical scientific support for optimizing the “PV + Artemisia” ecological restoration model in desert PV parks.

2. Materials and Methods

2.1. Study Area Overview

The study area is the Xinhua Power Gulang Huanghuatan 150 MW PV Demonstration Zone, located at the northern edge of Huanghuatan Forest Farm, Huanghuatan Town, Gulang County, Wuwei City, Gansu Province, on the margin of the Tengger Desert (Figure 1 and Figure 2). It has a temperate continental climate with a multi-year average temperature of 6.7 °C, annual sunshine duration of 2693.5 h, average wind speed of 10.8 m/s, and annual precipitation of 288 mm. Gulang County has abundant sunlight, classified as a Grade I solar resource zone (annual radiation > 1750 kWh/m², GB/T 37526-2019 [18]) with an annual average total radiation of 5800–6200 MJ/m².

The station is equipped with FA, HSA, and TSA tracking PV panels (the PV zoning is shown in Figure 1b), and was constructed in November 2013. As of 2023, the power station has been in operation for 10 years. The station occupies a total area of 3.7 km², with the FA, TSA, and HSA zones covering 1.9 km², 0.6 km², and 1.2 km², respectively. The photovoltaic modules measured 1956 mm × 992 mm (FA: 30° tilt), 1960 mm × 996 mm (HSA: 0 ± 45° tracking rotation), and 2000 mm × 1000 mm (TSA: 20° fixed tilt ± 30° tracking rotation), respectively. Grass seeds were sown during construction without artificial irrigation; vegetation later grew naturally. After 10 years of operation, vegetation inside the station grew well, with an overall coverage of 65%–75% (highest in the TSA zone). The area has rich plant diversity, dominated by Artemisia ordosica, with other shrubs, such as Hedysarum mongolicum, reaching a maximum height of 2.1 m. Herbaceous plants include over 10 species, such as Stipa glareosa, Achnatherum splendens, Bassia dasyphylla, and Echinops gmelinii, with BSCs covering the surface. Outside the station, the native desert vegetation is relatively dense but less diverse, dominated by Artemisia ordosica with surface BSCs.

2.2. Sample Collection

In October 2023, sampling was conducted in areas with different PV panel types inside the station and in adjacent undisturbed desert. Using random sampling, pristine desert served as the control (three soil types were collected from each of the two control points, each with 3 replicates), with three treatments based on PV panel types (Figure 3): HSA, TSA, and FA (each group included three types of soil samples, with 3 replicates per type). A total of 15 sampling points were established, comprising 6 control and 9 treatment groups.

At each sampling point, three soil types were collected: non-crusted soil, BSCs, and sub-crust soil (all samples within the photovoltaic zone were collected from the areas between panels), resulting in a total of 45 samples (18 control samples and 9 samples of treatment groups). Non-crusted soil samples were collected by excavating three randomly selected 0–5 cm depth subsamples within 1 m × 1 m quadrats using a shovel, followed by homogenization, while BSCs (identified by hardened gray-brown surface morphology) within 3 m of corresponding quadrats were carefully excavated to a depth of 0.22 cm, with subsequent sub-crust soil sampling (0–5 cm depth) conducted at the exposed sites post-BSC removal. All non-crusted soil samples were composite samples (each comprising three subsamples), totaling 15 composite samples, with each bulk sample having a volume of approximately 400 cm³. The samples were immediately transported to the laboratory. All the samples were divided into two subsamples, one of which was sieved through 2 mm mesh, air-dried and used for measuring the soil physicochemical properties, whereas the other was employed to determine soil water content (WC) and saturated water content (SWC).

2.3. Estimation of Physicochemical Properties

WC was immediately determined by the oven-drying method (105 °C for 48 h) [19]. SWC was determined by placing 100 g of fresh soil in a perforated PVC tube (bottom-sealed with gauze to prevent particle loss), recording its initial mass (m₁), saturating it via 12 h water immersion (with the water level maintained above the soil surface), allowing for free drainage until the gravitational water ceased (natural state), measuring the saturated mass (m₂), and finally oven-drying the sample at 105 °C to a constant weight (m₃). SWC was calculated as follows [20]:

$$SWC(\%)=(m_2-m_1)/m_3\times 100\%$$

(1)

pH was measured using a 1:2.5 soil-deionized water mixture (HI 98311, HANNA Instruments, Padua, Italy). Electrical conductivity (EC) was measured using a 1:5 soil–deionized water mixture (pHS-3E, Rex, Shanghai INESA Scientific Instrument Co., Limited company, Shanghai, China). SOC was measured using the potassium dichromate volumetric-dilution heat method [19]. The conversion between SOM and SOC followed the equation: SOM (g/kg) = SOC (g/kg) × 1.724 (the van Bemmelen factor) [19].

Carbonate content was determined by neutralization titration [19]. Briefly, air-dried soil (<0.149 mm) was accurately weighed (5.000 ± 0.005 g) and reacted with 20.00 mL of 1 mol/L HCl in a 250 mL conical flask for 30 min with intermittent shaking. The residual acid was titrated against 0.5 mol/L NaOH standard solution (standardized with potassium hydrogen phthalate) using phenolphthalein as the indicator (endpoint at a pH of 8.3). The carbonate content was calculated as:

$${CaCO}_3(\%)=\left[\left(V_0-V\right)\times c\times 0.050\right]/m\times 100$$

(2)

where V₀ and V are NaOH volumes (mL) consumed by the blank and sample, respectively; c is the NaOH concentration (mol/L); and m is soil mass (g).

2.4. Estimation of Soil and Crust Carbon Stocks

To systematically reveal changes in SOC and inorganic carbon (SIC) contents in shallow soils and BSCs under different PV installation modes and their relationships, PV installation modes were categorized into four types: PD, HSA, TSA, and FA. The soil type in the study area is Aeolian Sandy Soil (per the Chinese soil classification system) [21]. For the 0–5 cm soil layer, SOC and SIC (represented by carbonate content) stocks were estimated using measured data, combined with bulk densities from regional studies in the literature [22,23]: 1.60 g/cm³ for non-crusted soil; 1.45 g/cm³ for sub-crust soil; 1.32 g/cm³ for BSC; and a BSC thickness of 2.2 mm.

SOC and SIC stocks were calculated using the following formulas [24]:

$${SOCS}_j=10\times 0.58\times H_j\times O_j\times W_j$$

(3)

$${SICS}_j=10\times H_j\times I_j\times W_j$$

(4)

$${SCS}_j={SOCS}_j+{SICS}_j$$

(5)

where j denotes sample type (non-crusted soil, BSCs, and sub-crust soil); SOCS_j is the SOC density of the j-th soil sample (kg C/m²); H_j is the average thickness (cm); O_j is the average organic matter content (%); W_j is the average bulk density (g/cm³); 0.58 is the conversion factor between SOM and SOC; SICS_j is the inorganic carbon density of the j-th soil sample (kg C/m²); I_j is the average carbonate content (%); and SCS_j is the total carbon density of the j-th soil sample (kg C/m²).

2.5. Data Processing

Data were organized and statistically analyzed using Microsoft Excel 2019 (Microsoft Corporation, Redmond, WA, USA) and IBM SPSS Statistics 26.0 (IBM Corp., Armonk, NY, USA). One-way analysis of variance (ANOVA) was employed to compare soil properties among groups. Fisher’s least significant difference (LSD) test was used for mean separation at significance levels of p < 0.05. Pearson correlation coefficients were used to analyze relationships between indicators (significance level p = 0.05). OriginPro 2021 (OriginLab Corporation, Northampton, MA, USA) was used for graphing. Data are presented as mean ± standard deviation.

3. Results

3.1. Soil and Crust Physicochemical Properties

Analyses of WC, pH, EC, SWC, SOM, and carbonate content (CO₃²⁻) in soils and BSCs are shown in

Table 1. Physical properties of soils and BSCs.

Installation Types	Sample Types	pH	EC (μS/cm)	WC	SWC
FA	1	9.10 ± 0.02 a	60.70 ± 5.47 a	0.49% ± 0.15% a	25.54% ± 0.25% a
	2	8.39 ± 0.06 b	84.40 ± 5.63 b	0.62% ± 0.08% ab	49.42% ± 0.72% b
	3	9.07 ± 0.03 a	55.73 ± 0.67 a	2.44% ± 0.45% b	25.17% ± 3.91% a
HSA	1	9.13 ± 0.06 a	58.05 ± 2.19 a	0.64% ± 0.12% a	18.68% ± 2.55% a
	2	8.69 ± 0.01 b	94.55 ± 0.21 b	0.36% ± 0.01% a	38.00% ± 0.64% b
	3	9.03 ± 0.05 a	56.03 ± 1.32 a	1.43% ± 0.31% b	15.81% ± 6.11% a
TSA	1	9.08 ± 0.02 a	62.70 ± 7.88 a	0.87% ± 0.01% a	25.10% ± 3.50% a
	2	8.44 ± 0.05 b	89.97 ± 10.89 b	0.37% ± 0.10% b	47.29% ± 2.35% c
	3	9.08 ± 0.03 a	57.02 ± 4.38 a	0.84% ± 0.01% a	33.17% ± 3.07% b
PD	1	9.11 ± 0.05 a	56.53 ± 2.12 a	1.08% ± 0.46% a	27.23% ± 3.54% a
	2	8.48 ± 0.08 b	105.30 ± 8.13 b	0.28% ± 0.14% b	50.07% ± 0.56% b
	3	9.13 ± 0.01 a	56.20 ± 2.27 a	0.72% ± 0.13% ab	30.39% ± 1.73% a

Note: PD represents pristine desert, HSA represents horizontal single-axis, TSA represents tilted single-axis, and FA represents FA. Sample types: 1 indicates non-crusted soil, 2 indicates BSC, and 3 indicates sub-crust soil. The same applies hereinafter. Different lowercase letters denote significant differences among treatments (p < 0.05). The same applies hereinafter.

and

Table 2. Organic matter and carbonate content of samples.

Installation Types	Sample Types	SOC	SOM (g/kg)	CO32−
FA	1	0.43% ± 0.01% a	7.36 ± 0.17 a	10.05% ± 0.50% a
	2	1.07% ± 0.14% b	18.52 ± 2.36 b	10.40% ± 1.32% a
	3	0.25% ± 0.09% a	4.27 ± 1.57 a	10.95% ± 0.70% a
HSA	1	0.22% ± 0.19% a	3.72 ± 3.22 a	14.55% ± 2.99% a
	2	0.49% ± 0.18% b	8.45 ± 3.08 b	13.20% ± 1.10% a
	3	0.21% ± 0.10% a	3.68 ± 1.73 a	11.11% ± 1.03% b
TSA	1	0.30% ± 0.09% a	5.18 ± 1.50 a	12.23% ± 0.92% a
	2	0.57% ± 0.05% b	9.81 ± 0.94 b	12.05% ± 0.84% a
	3	0.31% ± 0.15% a	5.36 ± 2.52 a	13.80% ± 2.22% a
PD	1	0.30% ± 0.16% a	5.18 ± 2.83 a	9.41% ± 1.93% a
	2	1.21% ± 0.20% b	20.80 ± 3.46 b	12.74% ± 0.60% ab
	3	0.52% ± 0.36% a	8.99 ± 6.13 a	15.10% ± 2.71% b

Note: For installation types (PD, HSA, TSA, FA) and sample types (1, 2, 3), refer to Table 1. Different lowercase letters denote significant differences among treatments (p < 0.05).

and Figure 4.

3.2. pH Changes

BSC presence did not significantly alter soil pH, regardless of PV installation. Compared to the control, PV installation did not significantly change pH in non-crusted soil but significantly reduced pH in sub-crust soil. FA, TSA, and HSA zones decreased by 0.62%, 0.51%, and 0.71% on average (Table 1), respectively, with no significant differences between installation types (Figure 4a). Compared to the control, FA and TSA PV brackets did not significantly affect BSC pH, while HSA brackets significantly increased the BSC pH by an average of 2.4% (Figure 4a).

3.3. EC Changes

BSC presence did not significantly alter soil EC, regardless of PV installation (Table 1). PV installation had no significant effect on EC in non-crusted or sub-crust soil compared to the control. However, PV installation significantly reduced EC in BSC: FA, TSA, and HSA zones decreased by 19.85%, 14.56%, and 10.21% on average, respectively, with no significant differences between PV types (Figure 4b).

3.4. Water Content Changes

Under FA and HSA PV panels, BSCs significantly increased soil water content, while no significant effect was observed in the control or TSA zones (Table 1). PV installation affected soil and BSC properties. For non-crusted soil, PV presence reduced WC, with FA panels showing a significant 54.63% decrease compared to the control, while TSA and HSA panels showed no significant differences. No significant differences in non-crusted soil WC were observed between PV types, although the FA panels generally had lower WC than the other two types. For sub-crust soil, FA and HSA panels significantly increased WC in sub-crust soil by 238.9% and 98.61%, on average, compared to the control, while TSA panels showed no significant difference. FA panels had a significantly stronger moisturizing effect than tilted and HSA panels (190.5% and 70.63% higher, respectively). The installation of FA PV panels significantly affected the WC of BSC. Compared with the control, FA, TSA, and HSA PV panels increased BSC WC by 121.4%, 67.57%, and 72.22%, respectively. This indicates that FA PV panels had a significantly stronger moisturizing effect on BSCs than TSA and HSA panels (Figure 4c). The SWC of all sample types was lowest in HSA zones (Figure 4d).

3.5. Organic Matter Content

BSC presence did not significantly affect SOM content, regardless of PV installation. In the control, organic matter in sub-crust soil was higher than in non-crusted soil, while in PV zones, sub-crust SOM was slightly lower than in non-crusted soil (Table 2). For non-crusted soil, HSA panels significantly reduced organic matter by 28.19% compared to the control, while FA and TSA panels showed no significant differences. PV installation had no significant effect on organic matter in sub-crust soil. For BSC, FA panels did not significantly affect organic matter content, while TSA and HSA panels significantly reduced it by 52.84% and 59.38% on average, respectively, compared to the control (Figure 4e).

3.6. Carbonate Content

The carbonate content in soils and BSCs under different PV types is shown in Figure 4f and Table 2. In the control, the average carbonate contents in non-crusted soil, BSC, and sub-crust soil were 9.41%, 12.74%, and 15.10%, respectively. BSCs significantly increased soil carbonate in the control by 60.47%, on average. In PV zones, this effect varied by type: HSA panels significantly increased carbonate in sub-crust soil via BSCs, while FA and TSA panels showed no significant effect. Compared to the control, PV installation increased the carbonate content in non-crusted soil, with tilted and HSA zones showing significant increases of 29.97% and 54.62% on average, respectively. PV installation significantly reduced carbonate in sub-crust soil: FA, TSA, and HSA zones decreased by 27.48%, 8.61%, and 26.42% on average, respectively. FA panels significantly reduced BSC carbonate by 18.37% compared to the control.

3.7. Correlation Between Soil Physicochemical Properties and Carbon Composition

Pearson correlation analysis (Figure 5) showed significant negative correlations between pH and SOC. EC, WC, and SWC each exhibited a significant positive correlation with SOC. This indicates that a higher pH (stronger alkalinity) was associated with lower SOC, while higher EC, WC, and SWC were associated with higher SOC. The carbonate content showed no significant correlation with physicochemical indicators but showed a weak negative correlation with SOC.

3.8. Carbon Density of Biological Crusts and Soil

Carbon densities in soils and BSCs (

Table 3. Carbon density of soil and BSC.

Installation Types	Sample Types	SOCS (kg C/m2)	SICS (kg C/m2)	SCS (kg C/m2)
FA	1	0.34	8.04	8.38
	2	0.03	0.30	0.33
	3	0.18	7.94	8.12
HSA	1	0.18	11.64	11.82
	2	0.01	0.38	0.40
	3	0.15	8.05	8.21
TSA	1	0.24	9.78	10.02
	2	0.02	0.35	0.37
	3	0.22	10.01	10.23
PD	1	0.24	7.53	7.77
	2	0.04	0.37	0.41
	3	0.38	10.95	11.32

Note: For installation types (PD, HSA, TSA, FA) and sample types (1, 2, 3), refer to Table 1. SOCS represents SOC density, SICS represents inorganic carbon density, SCS represents total carbon density.

) were calculated using bulk densities (1.60 g/cm³ for non-crusted soil, 1.45 g/cm³ for sub-crust soil, 1.32 g/cm³ for BSC) and depth (5 cm for non-crusted and sub-crust soil, 0.22 cm for BSC), combined with data from Table 2. Based on field observations (25% crust coverage in each plot), average carbon densities (

Table 4. Average carbon density of soils under different types of PV.

Installation Types	Non-Crusted Soil (kg C/m2)	BSCs (kg C/m2)	Sub-Crust Soil (kg C/m2)	Average Carbon Densities (kg C/m2)
FA	8.38	0.33	8.12	8.40
HSA	11.82	0.40	8.21	11.01
TSA	10.02	0.37	10.23	10.17
PD	7.77	0.41	11.32	8.76

Note: For installation types (PD, HSA, TSA, FA) refer to Table 1.

) were calculated assuming a 1:3 crust-to-non-crust ratio.

4. Discussion

4.1. Impact of PV Station Construction on Soil Water Content

The study area is a typical arid region where limited water is a key constraint on biological growth, making soil WC critical to ecosystem function [25]. Soil evaporation depends on near-surface air humidity; lower humidity accelerates evaporation. The impacts of the PV station on air humidity and soil WC vary with local environmental conditions and surface characteristics. Previous studies indicate that PV stations in desert areas generally increase air temperature and reduce humidity [26]. PV panels block direct sunlight, reducing surface heat absorption and soil temperature, which affects plant transpiration and evaporation [26]. However, they also absorb or reflect longwave radiation, hindering nighttime surface cooling [27]. Their low heat capacity causes faster air warming and lower humidity nearby [28], with heat radiation increasing the temperature inside the PV station compared to outside [7], thereby promoting evaporation and transpiration.

Current research on PV impacts on soil WC remains inconsistent, possibly due to differences in location, station age, and sampling times. In this study, non-crusted soil WC inside the station was lower than in the surrounding desert, consistent with the results of a study by Tanner et al. [29] in the Mojave Desert. This may be because sparse precipitation and high evaporation potential in the region make the moisture-preserving effect of PV shading negligible, resulting in lower WC. WC reduction varied by PV type. No significant differences were observed between the three installation modes; only the FA panels showed a significant difference from the control. This may be due to the more flexible sunlight angle adjustment in tilted and HSA panels, enhancing shading and evaporation inhibition, which offsets temperature-induced evaporation promotion. In contrast, FA panels have a fixed angle and weaker shading.

BSCs significantly regulate ecohydrological processes in desert ecosystems [13,30]. In this study, BSCs inside the station were collected between panels, as PV shading and rain interception hinder BSC growth, leaving few BSCs under panels. The effects of BSCs on shallow soil WC (0–5 cm) differed between the PV power station and the control area: BSCs in the surrounding desert had no significant effect on shallow soil WC, whereas BSCs within the station increased soil WC, particularly in the FA and HSA PV zones. This finding aligns with previous research on BSCs [31]. This is primarily due to the rain interception function of BSCs, which significantly alters precipitation infiltration processes and the redistribution pattern of soil WC. Under certain conditions, this reduces the effective recharge of precipitation to deep soil WC. Additionally, algal BSCs exhibit an evaporation-inhibiting effect after small rainfall events (precipitation < 5 mm), prolonging the retention of WC in shallow soil [7,32]. Higher WC in sub-crust soil and BSCs inside the station—especially in FA zones—may be due to reduced surface wind speed by large-scale PV arrays [10]. Wind increases BSC WC loss; in bare desert, wind keeps BSCs dry, while lower amounts of wind inside the station maintain higher WC, facilitating shallow soil WC retention. FA panels more effectively block wind than flexible TSA/HSA panels, explaining their stronger moisturizing effect.

4.2. Impact of PV Station Construction on Soil pH and EC

pH critically affects soil nutrient availability, microbial activity, and plant growth. This study confirmed alkaline soils in the region. The pH of BSCs was significantly lower than in underlying and surrounding non-crusted soils, consistent with that reported by Du et al. [13]. PV installation did not affect non-crusted soil pH, consistent with that reported by Zhou et al. [33]. BSCs did not alter soil pH in the control but slightly reduced it in PV zones (non-significant). PV installation significantly reduced pH in sub-crust soil, possibly due to higher WC in the BSCs and sub-crust soil inside the station. BSC development refines soil and reduces pH [34]; lower WC in the control BSCs limited their activity and pH regulation, while higher WC in the PV zones enhanced BSC activity, reducing alkalinity more effectively.

EC indicates salt content and fertility. No significant EC differences were observed between PV and control soils, regardless of BSC presence. Previous studies on PV impacts on EC are inconsistent. Luo et al. [30] reported 27.8% higher EC inside PV stations, while Wang et al. [35] found lower EC inside. PV panels significantly reduced BSC EC, possibly due to altered BSC types under PV, with PV sand barriers inhibiting dust and sediment capture by rough BSC surfaces [13], reducing salt accumulation.

4.3. Impact of PV Station Construction on Soil Organic Matter

Soil carbon pools indicate land degradation. In arid western China, vegetation degradation and soil erosion due to climate, topography, and human activity cause significant loss of SOC [36,37].

Regarding the impact of PV panel installations on organic matter in non-crusted soil, this study found that there was no significant difference in SOM content between the PV zone and the control area, which is consistent with the findings of Wang et al. [35]. This is probably because significant damage was caused to the original environment during PV construction, resulting in a substantial loss of organic matter. However, through subsequent vegetation restoration via seeding, well-grown Artemisia ordosica shrubland with high diversity had developed in the power station by the time of sampling, with an overall vegetation coverage of 65%~75%. Vegetation growth and litterfall thus replenished organic matter in the soil within the station.

In this study, sub-crust soils in natural deserts exhibited higher organic matter content compared to non-crusted soils, demonstrating that BSC development enhances sub-crust SOM accumulation, consistent with findings by Li et al. [32] and Hu et al. [38]. Arid-region BSCs accumulate nutrients via photosynthesis, becoming major sources of SOC [39]. However, BSCs in PV zones did not enhance sub-crust soil SOC, possibly due to: (1) construction disturbance damaging original BSCs, with recovery still incomplete [16]; (2) PV shading inhibiting BSC photosynthesis and SOC synthesis; and (3) competition between BSCs and plants—the arid-region shrub cover negatively correlates with BSC abundance [17], and higher Artemisia cover in restored PV zones intensifies competition for light, water, and nutrients, limiting BSC growth and SOC input [16].

Overall, the organic matter content of BSCs in the PV zone was lower than that in the control area, which may share similar reasons to those responsible for the aforementioned differences in sub-crust soil. Significantly lower organic matter in TSA and HSA zones than in the control and FA zones likely reflects stronger shading by these panels, further inhibiting BSC photosynthesis and organic matter synthesis.

4.4. Impact of PV Station Construction on Soil Carbonate

Soil inorganic carbon—dominated by CaCO₃—forms via weathering, with slower turnover, higher accumulation, greater stability, and larger stocks than SOC [36]. In the control, BSCs significantly increased sub-crust carbonate, possibly due to enhanced SOC via BSC photosynthesis. Inorganic carbon sequestration—represented by carbonate content—correlates positively with SOC in arid soils [40]. SOC influences the soil adsorption of CO₂. High mineralization rates in alkaline, calcium-rich arid soils generate more CO₂, which reacts to form dissolved and particulate inorganic carbon, increasing stocks [40]. The absence of this effect in the BSCs of PV zones may relate to their failure to increase sub-crust organic matter.

Higher carbonate in non-crusted PV soil—significantly in TSA and HSA zones—may result from higher silt and clay content. The study area comprises a PV power station where artificial vegetation restoration has been implemented, dominated by Artemisia ordosica shrubland. Studies by Meng et al. [41] and Cai et al. [42] demonstrated that artificially revegetated photovoltaic desert soils exhibit significantly higher silt and clay contents compared to non-restored areas, with this textural modification showing a positive correlation with enhanced inorganic carbon sequestration capacity [40]. Non-significant carbonate increases in FA zones may result from lower WC, as higher WC enhances inorganic carbon sequestration in alkaline arid soils [40], with lower WC in FA zones offsetting the effect of higher silt and clay content.

The lower amount of carbonate in sub-crust PV soil compared to the control may result from lower organic matter and a lower pH. Higher SOC enhances inorganic sequestration [40], while a lower pH (higher H⁺) promotes carbonate dissolution, reducing content.

4.5. Relationship Between Soil Physicochemical Properties and Carbon Composition

SOC dynamics are strongly constrained by physicochemical properties. pH significantly affects SOC via microbial activity and organic-mineral complexation [43]. Consistent with previous studies [43,44,45], this study found a significant negative correlation between pH and SOC, likely due to: (1) alkaline conditions accelerating microbial decomposition of organic matter, promoting SOC mineralization [46] and impeding SOC accumulation [44,45]; and (2) the pH increase limiting plant micronutrient uptake, reducing biomass input to soil [45].

WC also significantly positively correlated with SOC. Higher WC benefits Artemisia growth and photosynthesis, increasing root and litter input; it regulates microbial activity to affect SOC mineralization. Optimal moisture promotes clay–organic matter aggregation, forming water-stable macroaggregates that protect SOC [45].

SOC also regulates physicochemical properties; significant positive correlations were observed between SOC, EC, and SWC. SOC possesses intrinsic high water-absorbing properties, facilitates soil aggregate formation, optimizes soil pore structure, and enhances water-holding capacity. Microbial decomposition of SOC releases substantial soluble organic and inorganic ions, consequently elevating soil EC.

The soil carbonate content showed no significant correlation with other physicochemical properties. This likely occurs because carbonates are primarily derived from the parent bedrock, while limited moisture restricts dissolution–reprecipitation translocation processes. Carbonate accumulation is predominantly driven by allochthonous inputs.

4.6. Restoration Effect of PV Station Construction and Ecological Restoration Measures

This 10-year vegetation restoration study showed PV station impacts on soils, BSCs, and crust-mediated ecological processes. PV reduced non-crusted soil WC, did not improve salinity–alkalinity or organic matter, but significantly increased carbonate, promoting soil inorganic carbon sequestration.

The implementation of vegetation restoration measures dominated by Artemisia ordosica in PV power stations exerts a significant impact on the formation of BSCs. Currently, findings regarding the relationship between BSCs and higher plants in arid and semi-arid ecosystems remain inconsistent. On the one hand, the cultivation of shrubs, such as Artemisia ordosica, can effectively enhance soil stability and erosion resistance [47]; their litter can increase SOM content and improve soil nutrient status [41], as well as alter the structure and function of soil microbial communities [48], thereby influencing the formation of BSCs. Meanwhile, the growth of Artemisia ordosica can reduce surface temperature, decrease water evaporation, and increase soil WC by intercepting precipitation [49,50], providing suitable temperature and humidity conditions for the growth of BSCs. On the other hand, with the continuous restoration of vegetation, plant coverage and biomass increase accordingly, intensifying the competition between BSCs and plants for natural resources (light, water, and nutrients), which exerts stress on the growth of BSCs [16].

In the present study, the growth of BSCs inside the station was generally inferior to that outside; some impacts of BSCs on soil also varied according to PV panel types. The artificial disturbance during PV station construction caused extensive mortality of BSCs, leading to the re-development of surface BSCs, which had not yet recovered to the natural levels outside the station. The sand barrier effect of the PV station provides an environment with reduced wind-blown sand for BSCs, which is conducive to BSCs increasing soil WC in the shallow layer. Simultaneously, the planting of shrubs, such as Artemisia ordosica, also promotes the growth of BSCs. However, the shading effect inside the station and competitive stress from plants restrict the growth of BSC, resulting in the lower salt content and carbon sequestration capacity of BSCs within the station. This has affected inorganic carbon pools, such as carbonates, exerting adverse impacts on the carbon sequestration of BSCs and the sub-crust soil.

Evaluated by carbon sequestration, vegetation-restored PV stations enhanced non-crusted soil sequestration, with higher average carbon density inside—especially in horizontal and TSA zones. However, the damage to BSCs and the inhibition of their growth caused by PV installations resulted in lower average carbon density in both BSCs and sub-crust soil inside the station, with their carbon sequestration effect being significantly lower than that outside. Overall, PV construction and vegetation restoration still promoted inorganic carbon sequestration inside the station soil—markedly, in HSA and TSA zones. Therefore, for future desert PV construction, priority can be given to the installation of HSA and TSA systems to further enhance the soil carbon sequestration effect within photovoltaic power stations.

5. Conclusions

This study aimed to explore the optimal PV configuration for enhancing carbon fixation and soil quality in Artemisia ordosica-restored desert PV systems. The results demonstrated that: (1) Compared to PD, FA and HSA increased WC in both BSCs and sub-crust soil. However, HSA reduced water-holding capacity in these layers. SOM in sub-crust soils decreased by 52.84%–59.38% in HSA and TSA areas, while carbonate content in non-crusted soil increased by 29.97%–54.62% in these areas. Notably, the carbonate content in BSCs was significantly reduced in FA areas. All PV installation types demonstrated a reduction in carbonate content in sub-crust soils; (2) PV affected SOC sequestration mainly via WC regulation, improving physicochemical properties, while inorganic carbon was minimally affected by this process; and (3) PV construction and vegetation restoration promoted soil inorganic carbon sequestration inside the PV station. The effectiveness of the different PV panel types in improving the carbon sequestration effect was ranked HSA > TSA > PD > FA. Overall, implementing vegetation restoration measures during photovoltaic construction effectively improves soil quality, with HSA and TSA emerging as the most promising types for enhancing soil carbon sequestration effects within PV power stations.