Abstract
Background:
In the severely eroded red soil regions of southern China, characterized by significant ecological degradation, there is an urgent need for effective vegetation restoration techniques.
Method:
A space-for-time substitution approach was employed, wherein three 20 m × 20 m quadrats were established at each of the Paspalum wettsteinii restoration sites (1-, 6-, and 12-year-old) and a control bare land. Soil samples were collected for physicochemical characterization and 16S rDNA sequencing analysis.
Results:
The result showed that soil structure improved rapidly within the first 6 years, followed by a 357% increase in soil organic matter in the later 6 years. Bacterial α-diversity initially rose and then declined, while the microbial network shifted from loose to tightly reconnected between 6 and 12 years, with Thaumarchaeota emerging as a key stability-indicating taxon. Total nitrogen (TN), organic matter (OM), and available potassium (AK) collectively explained >75% of the community variation, indicating that carbon-nitrogen synergy drives community assembly, with available potassium regulating the secondary axis.
Conclusion:
Vegetation restoration using P. wettsteinii significantly improved soil physicochemical properties and drove bacterial diversity and functional succession, exhibiting a distinct "stage-specific synergistic recovery" pattern in which the 6–12 years period represents a stage for soil bacterial communities, shifting from structural restoration to functional stabilization.
1 Introduction
Soil degradation has emerged as a critical challenge for global terrestrial ecosystems (; ). In the red soil region of southern China, intense weathering, high rainfall, and historical human disturbance have led to severe soil exposure, depletion of nutrient pools, and disruption of microbial networks, thereby constraining regional ecological security and sustainable land use (Zhang et al., 2014; Meng et al., 2021). Vegetation restoration is a primary strategy to reverse such degradation. By enhancing soil organic carbon (SOC) and nutrient storage through litter input, root turnover, and mycorrhizal associations, restoration triggers cascading effects on the structure and function of soil microbial communities (Qi et al., 2017). Microorganisms, however, are not merely passive responders. Their rapid turnover and metabolic diversity make them a key driver of soil quality, with the trajectory of bacterial community assembly determining nutrient cycling rates, carbon sequestration potential, and ecosystem resilience (Sun et al., 2018). Therefore, clarifying the temporal dynamics and critical transition points in the co-evolution of vegetation, soil properties, and microbial communities is essential for accurately assessing restoration outcomes and informing management strategies.
Advances in high-throughput sequencing have greatly expanded knowledge of soil bacterial diversity during restoration. Many studies report a general increase in bacterial α-diversity over time, often marked by a shift from oligotrophic taxa (e.g., Acidobacteria) to copiotrophic groups (e.g., Actinobacteria) as the dominant populations (). Yet, increased diversity does not necessarily reflect functional recovery, as critical processes such as nitrification and organic phosphorus mineralization depend on tightly connected co-occurrence networks (Wu et al., 2024). Network analyses suggest that microbial interactions are initially sparse after restoration begins, then increase significantly in node number and connectivity during mid-term stages (5–10 years), forming a “small-world” structure that enhances stability and functional redundancy (Miao et al., 2021; ; Wang W. et al., 2025). However, multi-time-point data are still scarce regarding whether such network transitions follow a consistent threshold and whether they coincide with or lag behind changes in soil physicochemical properties.
Moreover, the influence of soil factors on bacterial communities appears stage-specific. In degraded bare soil, pH and bulk density typically dominate community assembly. As restoration proceeds, resource constraints shift from “environmental filtering” toward a dual control by “substrate quality and quantity,” with organic matter and total nitrogen becoming central explanatory variables (Wang F. et al., 2021). Most previous studies, however, have relied on one or two sampling points, making it difficult to identify the transition between these regulatory phases. In addition, key synergistic indicators-such as available nutrients (e.g., available potassium) and soil aggregate structure (>0.25 mm)-have received less attention, limiting mechanistic insight into the recovery process.
To address these gaps, this study examined a chronosequence of Paspalum wettsteinii restoration (1, 6, and 12 years) alongside bare-land controls. By integrating soil physicochemical trajectories, 16S rDNA gene sequencing, redundancy analysis (RDA), and co-occurrence network analysis, we systematically investigated the following questions: What are the stage-specific changes and turning points in soil physicochemical properties across the restoration chronosequence? How do bacterial α-diversity, β-diversity, and co-occurrence network complexity respond to increasing restoration duration? Which soil factors dominate shifts in bacterial community structure, and how does their explanatory power vary over time?
The results are expected to identify ecological thresholds during vegetation restoration in degraded red soils and provide a theoretical and indicator-based framework for evaluating restoration effectiveness from a microbial perspective.
2 Materials and methods
2.1 Study area and sample collection
The study site is located in Hetian Town, Changting County, Fujian Province, China (25°33′-25°48′N, 116°18′-116°31′E). It receives an annual average of 1924.6 h of sunshine, with total precipitation amounting to 1737.1 mm. The temperature ranges from 17.5 °C to 18.8 °C, peaking at 38.9 °C during the summer and dropping to −4.9 °C in the winter. Additionally, the frost-free period lasts for approximately 260 days.
To intuitively illustrate the geographical location of the research area and the visual differences among experimental plots, a schematic diagram of the study area and field photographs showing the plot characteristics were provided in this section (Figure 1).
FIGURE 1
In August 2019, soil samples were collected from three plots (20 × 20 m) in areas where P. wettsteinii had been replanted at different intervals (1, 6, and 12 years prior to sampling; planted in March 2007, 2013, and 2018, respectively). Control samples were obtained from adjacent unplanted (bare) soils. To ensure the validity of the space-for-time substitution, the selected plots were carefully chosen to have consistent baseline conditions, specifically exhibiting similar slope gradients (averaging 18°), soil types (red soil), and initial degradation levels (severely eroded bare land) prior to the restoration project. P. wettsteinii seeds were initially sown in strip-shaped ditches dug along contour lines, covering an area of 0.33 ha at a density of 30 kg ha−1.
In each plot, five soil samples were collected from a depth of 0–20 cm using a cutting ring arranged in a quincunx pattern, employing the five-point sampling method (Sudrajat et al., 2019). Additionally, five soil profiles (0–20 cm) were excavated according to the quartering method, with 1 kg of soil collected from each profile. Soil samples from the same square were mixed evenly following the five-point sampling method; this composite sample was then returned to the laboratory, dried, ground, and passed through a 2 mm nylon sieve for subsequent soil chemical analysis. Furthermore, five additional soil samples (0–20 cm) were collected using a 10 cm × 10 cm × 10 cm tin box for soil water stability aggregate analysis, adhering to the five-point sampling method. Lastly, five soil samples (0–20 cm) were collected in high-temperature sterilized tissue culture bottles, following the five-point sampling method. Each trial site contains three plots, and the mean value of soil physical and chemical properties measured from five samples within each plot is considered as one replicate. Each site comprises three plots, i.e., three replicates.
Soil samples from the same plot were mixed uniformly, and 100 g was collected and stored in dry ice for soil bacterial 16S rDNA sequencing, after the removal of gravel and plant roots on a super-clean work table. Three replicates for each trial site.
2.2 DNA extraction and 16S rDNA sequencing
Soil DNA was extracted using the MagPure Stool DNA KF Kit B and Magnetic Bead Fecal and Soil Genome Extraction Kit, following the manufacturer’s protocol. Briefly, a sample (100–200 mg) was weighed and added to a tube containing grinding beads. After adding 1 mL of Buffer ATL/PVP - 10 and grinding, the mixture was incubated at 65 °C for 20 min and then centrifuged. The supernatant was mixed with Buffer PCI and centrifuged again. Then it was transferred to a deep well plate with magnetic bead binding solution, including Proteinase K and RNase A. Following several washing steps, DNA was eluted with 100 μL of Elution Buffer using a Kingfisher device.
For library construction, the V4 region of 16S rDNA was targeted using Phanta Max Master Mix with degenerate primers. PCR was performed in a 50 μL system with 30 ng template DNA, including initial denaturation, 30 cycles of denaturation, annealing, extension and a final extension. PCR products were purified and sequenced on the Illumina MiSeq platform, generating 2 × 250 bp paired - end reads.
2.3 Data processing methods
Soil physical-chemical data analysis was conducted using R software (version 3.1.1). For the classification of soil bacterial species, OTU clustering of assembled sequence tags was performed using USEARCH software (v7.0.1090), enabling preliminary taxonomic assignment. To assess species richness and evenness within samples, Alpha diversity indices-including the observed species count, Chao1 index, ACE index, Shannon index, and Simpson index-were calculated using mothur software (v1.31.2). Detailed formulas for these indices are available on the mothur official website (http://www.mothur.org/wiki/Calculators).
Beta diversity analysis was conducted with QIIME software (v1.80) to examine differences in species composition between samples. A heatmap visualizing beta diversity and sample similarities was generated using the aheatmap function from the NMF package in R software (v3.1.1). Furthermore, LEfSe analysis was performed via the online LEfSe tool (https://huttenhower.sph.harvard.edu/galaxy/) to identify significant biomarkers reflecting species differences across groups.
Non-metric multidimensional scaling (NMDS) based on Bray–Curtis distance was carried out for soil bacterial phyla using an online NMDS tool (https://www.bioincloud.tech/standalone-task-ui/bray_nmds). Co-occurrence network analysis at the phylum level was performed with the online propr tool (https://www.bioincloud.tech/standalone-task-ui/co_gene_net). Additionally, redundancy analysis (RDA) between soil bacterial genera and soil physicochemical properties was conducted using an online RDA platform (https://www.bioincloud.tech/standalone-task-ui/rda).
3 Results
3.1 Soil physicochemical restoration trajectory
With the progression of vegetation restoration, soil physicochemical properties showed significant and continuous improvement (Figures 2, 3), following a distinct pattern of “phased transition followed by stabilization.” This reflects the gradual recovery of degraded bare land toward a healthy soil ecosystem. Soil bulk density decreased from 1.56 g/cm3 in bare land to 1.16 g/cm3 after 12 years of restoration, a reduction of 25.6%. About 80% of this decrease occurred within the first 1–6 years, indicating rapid early-stage structural loosening and subsequent stabilization. Non-capillary porosity increased from 8.56% to 18.15%, and capillary porosity rose from 31.69% to 49.58%, leading to a marked increase in total porosity. Non-capillary porosity in particular exhibited an exponential growth trend (R2 = 0.91; Supplementary Figure S1), characterized by a “slow initial increase followed by rapid growth”, suggesting more pronounced recovery of soil aeration in later stages.
FIGURE 2
FIGURE 3
The content of water-stable aggregates (>0.25 mm) increased from 12.31% to 19.94%, a rise of 62%, and was strongly positively correlated with organic matter content (r = 0.98; Supplementary Figure S1), confirming that organic matter accumulation is a key driver of soil structure improvement. Soil fertility indicators displayed a nonlinear recovery pattern: “slow initially–accelerated in mid-term–stabilized later.” Organic matter content increased from 3.78 g/kg to 17.29 g/kg, a 357% increase, with 52% of this gain occurring between 6 and 12 years, highlighting the cumulative effect of litter input on carbon sequestration in later stages. Total nitrogen rose from 0.26 g/kg to 1.39 g/kg, alkali-hydrolyzable nitrogen from 72.49 μg/g to 259.38 μg/g, available potassium from 9.87 mg/kg to 93.55 mg/kg, total phosphorus from 32.47 μg/g to 135.77 μg/g, and total potassium from 0.41% to 1.64%, all showing significant upward trends. Synergistic recovery was observed among nutrient elements—for example, the growth-rate ratio of N, P, and K approximated 1:1.2:1.5-reflecting the progressive restoration of soil nutrient pools and the reactivation of biogeochemical cycles.
3.2 Temporal variations in soil bacterial diversity
3.2.1 α-diversity
The richness and diversity indices of soil bacterial communities exhibited an increasing trend from the initial to middle stages of restoration, followed by a significant decline after long-term recovery, with certain regular differences observed among various indices. Specifically, Observed species, Chao index, and ACE index all displayed a pattern of initial increase followed by decrease with restoration time (Figures 4a–c): after 1 and 6 years of restoration, soil bacterial species richness was significantly higher than that of the control group, indicating that vegetation restoration effectively promoted the increase of soil bacterial species in the early stage; however, after 12 years of restoration, all the above richness indices showed significant decreases, suggesting that environmental or biological factors inhibiting the maintenance or growth of bacterial species richness may have emerged during the long-term restoration process.
FIGURE 4
The Shannon index (Figure 4d) showed a similar trend, increasing in the early restoration stage but declining significantly after 12 years, reflecting that species diversity of bacterial communities did not continue to improve but instead regressed during the long-term restoration phase. Notably, the Simpson index (Figure 4e) displayed an opposite pattern compared to the above indices. This index showed an upward trend after 12 years of restoration, indicating increased relative abundance of dominant bacterial species in the community at this stage, with community structure tending toward dominance by a few species and reduced evenness. This result corroborates the decline in Shannon index, further suggesting that soil bacterial communities may have undergone processes such as niche compression, intensified competition, or enhanced environmental filtering after long-term restoration, leading to decreased species diversity but increased dominance. This pattern is consistent with findings from native species reseeding in the desert steppe of Northwest China, where long-term vegetation restoration was found to reduce niche overlap, strengthen interspecific competition, and gradually increase the dominance of core constructive species (e.g., Agropyron mongolicum, Lespedeza potaninii) while decreasing the abundance and survival space of rare species, ultimately resulting in reduced community diversity but enhanced structural stability ().
Furthermore, Good’s coverage remained at high levels with minimal fluctuation across all treatments (Figure 4f), indicating that the sequencing depth employed in this study adequately covered the soil bacterial community, ensuring the reliability of diversity analysis results. Based on the above results, it can be preliminarily concluded that the impact of vegetation restoration on soil bacterial diversity is not linear growth, but rather reaches a peak in the middle stage followed by diversity decline during long-term recovery. This change may be associated with alterations in soil organic matter composition, root exudates, microenvironment, and interspecific competition relationships during vegetation succession.
3.2.2 β-diversity
The Figure 5 presents a weighted UniFrac distance clustering heatmap of soil bacterial communities across varying durations of vegetation restoration, illustrating the differences in community structure among the samples. The clustering results reveal a clear separation between the control group (CK) and the vegetation restoration groups (1a, 6a, 12a), indicating that vegetation restoration significantly impacts the structure of soil bacterial communities, thereby altering their composition and relative abundance. Within the vegetation restoration groups, the similarity in community structure among samples initially decreases with increasing restoration duration, followed by a subsequent increase. The three replicate samples (1a.1, 1a.2, 1a.3) from the treatment group exhibited relatively minor differences in community structure and were closer to the CK group, suggesting that the soil bacterial community structure was not yet stable during the initial stages of vegetation restoration and remained influenced by the original soil environment. In the 6a and 12a treatment groups, the differences in community structure among samples increased, particularly in the 12a treatment group, where the three replicate samples (12a.1, 12a.2, 12a3) were more distanced from the other treatment groups. This indicates that long-term vegetation restoration has resulted in more profound changes in the soil environment, fostering the development of a unique bacterial community structure. This phenomenon may be attributed to the combined effects of root exudates during vegetation growth and alterations in soil physicochemical properties, which have directed the succession of bacterial communities in a specific trajectory, ultimately leading to significant differentiation in soil bacterial communities across various restoration periods. Overall, the weighted UniFrac distance clustering heatmap distinctly demonstrates the influence of vegetation restoration duration on the structure of soil bacterial communities. As the process of vegetation restoration advances, the composition of soil bacterial communities undergoes gradual changes and approaches a state of stability. Notably, significant differences in community characteristics emerge at various stages of restoration, offering a crucial foundation for a more profound understanding of the relationship between vegetation restoration and the succession of soil microbial communities.
FIGURE 5
3.3 NMDS analysis of soil bacterial community succession
NMDS ordination (Bray-Curtis distance; stress = 0.033) (Figure 6) showed that P. wettsteinii restoration significantly altered soil bacterial community structure. This exceptionally low stress value strongly reinforces how distinctly different the bacterial community structures are across the restoration stages. Sequences were clearly separated along the first axis (NMDS1), forming a gradient from the control (CK) through the 1-year (1a), 6-year (6a) and 12-year (12a) restoration stages. PERMANOVA confirmed that restoration duration significantly explained the observed community differences (F = 10.589, P = 0.001). The 95% confidence ellipses revealed complete spatial separation between the CK community and those of the restored sites (1a, 6a, 12a). The bacterial community at 1a remained close to CK, indicating an approximately 1-year lag in microbial response after vegetation establishment. After 6a, the community shifted markedly to the right along NMDS1, suggesting that cumulative effects of root exudates and litter inputs began to dominate ecological filtering. By 12a, the community showed strong convergence (smallest ellipse), reflecting the selective establishment of dominant taxa and the onset of a relatively stable state. This temporal threshold (6–12 years) aligns with the previously observed window of decelerating soil-nutrient accumulation, supporting the “soil-microbe synergistic recovery” hypothesis.
FIGURE 6
3.4 Chronological succession of soil bacterial indicator taxa during vegetation restoration
LEfSe analysis (LDA > 3.0, P < 0.05) revealed a clear chronological succession of soil bacterial indicator taxa with increasing restoration time of P. wettsteinii (Figure 7). The bare land control was characterized by Chloroflexi and Corynebacteriales, indicating carbon source scarcity and low pH stress resulting from long-term degradation. At the 1-year restoration stage, the community shifted to dominance by Chitinophagia, Cytophagia, and Proteobacteria, reflecting both the demand for cellulose/chitin degradation and a surge of fast-growing (r-strategist) bacteria following fresh litter input. By year 6, co-enrichment of Acidobacteria (especially Subgroup 6), Flavobacteriia, Sphingobacteriia, Gemmatimonadetes, Betaproteobacteria, and Verrucomicrobia indicated that the soil had entered a synergistic phase of organic carbon accumulation, nitrification, and oligotrophy. At year 12, the indicator taxa were reduced to Thaumarchaeota (Nitrososphaeraceae), suggesting that ammonia-oxidizing archaea had become the dominant functional group, with the system approaching stability. This successional pathway followed the sequence: “poor nutrition-Chloroflexi” → “r-strategist-degraders” → “oligotrophic-Acidobacteria” → “ammonia-oxidizing archaea dominance,” which aligns with the physicochemical inflection point where the increase rates of soil SOC and TN began to slow after 6 years. These results suggest that vegetation restoration drives functional upgrading of the bacterial community through substrate and microhabitat filtering. Thaumarchaeota enrichment only at 12 years can serve as a microbial marker for the P. wettsteinii restoration system entering a “nitrification steady state.” The co-occurrence of multiple indicator taxa at the 6-year stage reflects intense niche differentiation, representing a critical window for soil-vegetation co-recovery. At this stage, introducing nitrogen-fixing shrubs could further strengthen the carbon-nitrogen cycling network established by Acidobacteria and Betaproteobacteria.
FIGURE 7
3.5 Co-occurrence network analysis reveals dynamic restructuring of soil bacterial interactions during vegetation restoration
Empirical co-occurrence network analysis (similarity threshold set at 0.6) revealed that the structure of soil bacterial interactions underwent a dynamic shift from “initially loose to later reorganized” as the duration of P. wettsteinii restoration increased (Table 1; Supplementary Figure S2). Networks for each stage were constructed based on three biological replicates (n = 3). The bare land control (CK) exhibited the smallest network, with only 15 nodes. However, it contained 79 edges, resulting in a high average connectivity of 10.5 and a very short average path length of 1.0. This pattern characterizes a “small yet tightly connected” network with high linkage density. After 1 year of restoration (1a), the node count increased to 18, but the number of edges decreased to 55. Consequently, the average connectivity dropped to 6.1, and the average path length increased to 2.30. This indicates a temporary dilution or weakening of the original community associations during the initial phase of vegetation establishment. By the sixth year (6a), the node count slightly decreased to 17, and edges further declined to 48. The network became sparser, with an average connectivity of 5.6 and an average path length of 1.98. At 12 years of restoration (12a), the network consisted of 16 nodes, but the number of edges recovered to 51. The average degree increased to 6.375, and the average path length shortened to 1.825. This suggests that bacterial interactions became tighter again and tended towards a new stable configuration.
TABLE 1
| Treatment | Similarity threshold | Number of nodes | Number of edges | Average degree | Average path length |
|---|---|---|---|---|---|
| CK | 0.6 | 15 | 79 | 10.533 | 1 |
| 1a | 0.6 | 18 | 55 | 6.111 | 2.301 |
| 6a | 0.6 | 17 | 48 | 5.647 | 1.978 |
| 12a | 0.6 | 16 | 51 | 6.375 | 1.825 |
Empirical co-occurrence network characteristics of soil bacterial communities under different restoration stages.
3.6 Key factors driving soil bacterial community assembly
The RDA-envfit analysis of bacterial communities at the genus level in relation to soil factors (Table 2; Supplementary Figure S3) revealed that total nitrogen (TN), organic matter (OM), and available potassium (AK) exhibited highly significant explanatory power (P < 0.01), with r2 values of 0.946, 0.790, and 0.750, respectively. Together, these three factors explained over 75% of the community variation. The vectors of OM and TN both pointed toward the positive direction of RDA1 (OM: RDA1 = 0.71, RDA2 = 0.71; TN: RDA1 = 0.72, RDA2 = 0.69), indicating that they jointly drive the distribution of the bacterial community along the positive axis of RDA1. The AK vector aligned almost completely with RDA2 (RDA2 = 0.998), mainly explaining the variation along the second axis. In addition, pH and bulk density (BD), also reached significant levels (P < 0.05), with r2 values ranging from 0.535 to 0.523. The remaining factors (e.g., OP, TP, TK, N/P, AG) showed no significant effects (P > 0.05).
TABLE 2
| Factors | RDA1 | RDA2 | r2 | P value |
|---|---|---|---|---|
| HN | 0.000 | 0.000 | 0.000 | 1.000 |
| AK | 0.067 | 0.998 | 0.750 | 0.003 |
| OP | 0.461 | 0.887 | 0.231 | 0.311 |
| TN | 0.723 | 0.691 | 0.946 | 0.000 |
| TP | −0.849 | −0.528 | 0.338 | 0.166 |
| TK | −0.529 | −0.849 | 0.444 | 0.068 |
| OM | 0.708 | 0.706 | 0.790 | 0.001 |
| pH | −0.755 | −0.656 | 0.535 | 0.035 |
| BD | −0.692 | −0.722 | 0.523 | 0.037 |
| NCP | −0.399 | −0.917 | 0.324 | 0.160 |
| CP | −0.675 | −0.738 | 0.559 | 0.033 |
| AG | −0.729 | −0.684 | 0.458 | 0.067 |
Results of RDA-envfit analysis by category.
The high explanatory power of TN and OM confirms that “carbon-nitrogen synergy” serves as the primary control axis for bacterial community assembly in the restored P. wettsteinii grassland. As litter accumulates, soil organic carbon (SOC) and TN increase simultaneously, directly selecting for oligotrophic (e.g., Acidobacteria) and copiotrophic (e.g., Proteobacteria) dominant genera and promoting community succession along the positive direction of RDA1. The independent contribution of AK to the second axis (r2 = 0.75) suggests that potassium availability acts as a secondary yet critical sorting pressure, potentially influencing bacterial growth rates by regulating cellular K+-ATPase activity, thereby shaping genus-level assemblages with distinct ecological strategies. The significance of pH, BD, and CP further indicates that soil acidity, compaction and porosity collectively form a “secondary filter.” Here, the vectors of pH and BD both lie in the negative direction and form an obtuse angle with TN/OM, implying that increased acidity and higher bulk density suppress the dominance of carbon- and nitrogen-utilizing bacteria, steering community turnover toward acid-tolerant and compaction-resistant lineages.
4 Discussion
This study systematically reveals the co-evolution patterns of soil physicochemical properties and bacterial communities during the vegetation restoration of P. wettsteinii, highlighting the “stage-specific restoration” effect of vegetation recovery on soil structure and function. It further points out the critical role of soil carbon-nitrogen synergistic mechanisms in driving the reconstruction of microbial communities. The findings not only provide empirical support for understanding the restoration mechanisms of degraded ecosystems but also offer a theoretical basis for optimizing regional ecological restoration strategies.
4.1 Characteristics and ecological significance of "stage-specific recovery" in soil physicochemical properties
This study reveals that soil nutrients follow a distinct three-stage recovery trajectory: “rapid initial improvement, accelerated mid-term accumulation, and eventual stabilization.” The three successional stages (1-year, 6-year, and 12-year) were not defined by a unified and universal scientific classification standard, but artificially divided based on the actual field conditions and vegetation succession characteristics of the study area. In the selection of time gradients, we fully considered the actual restoration duration of the experimental plots, site accessibility and long-term field management conditions, and the selected time gradients can effectively represent the natural succession process of local vegetation after disturbance. In the initial stage, the rapid establishment of above-ground cover improves soil physical structure, aligning with previous findings (; ). This structural improvement facilitates water retention and alleviates compaction, creating favorable microhabitats for soil fauna and microorganisms (Raupp et al., 2024). As restoration progresses, organic matter input from litter and root turnover stimulates microbial activity, driving nutrient sequestration (). By the later stage (12 years), the ecosystem approaches a steady state where organic matter input and output reach equilibrium.
4.2 Mechanisms of dynamic changes in soil pH during vegetation restoration
Soil pH exhibited a unique “increase-then-decrease” dynamic, reflecting a shift in dominant ecological drivers. The initial rise in pH aligns with observations by Zheng et al. (2024) and Zhao et al. (2024), attributed to the release of alkaline substances during organic matter decomposition and the accumulation of base cations following vegetation establishment. However, the subsequent decline in pH at 12 years is particularly noteworthy. Given the phosphorus-deficient nature of the red soils in Fujian (), the observed acidification is likely driven by the active secretion of organic acids by roots and microbes to solubilize insoluble phosphorus, a mechanism consistent with the “phosphorus limitation hypothesis” in mature ecosystems (). Therefore, the pH decline does not indicate degradation; rather, it signals the ecosystem’s transition into a mature stage characterized by intense biological interactions and active phosphorus cycling. Managers should recognize this acidification as a sign of functional maturity and monitor potential phosphorus deficiencies.
4.3 Mechanisms of response of soil bacterial community α- and β-diversity
4.3.1 α-diversity
In the early stages of vegetation restoration (1–6 years), the species richness (Observed species, Chao, ACE) and diversity indices (Shannon, Simpson) of bacterial communities show an increasing trend. However, there is a tendency for both species richness and diversity indices to significantly decline when vegetation restoration reaches 12 years. The underlying reason for this phenomenon is that vegetation restoration is a long - term process that can fundamentally transform soil ecosystems. Throughout this lengthy process, bacterial communities undergo a series of predictable successional stages, with their diversity change curve exhibiting a pattern of first increasing, then decreasing, and finally stabilizing, which is actually a very common ecological pattern. In the initial phase of restoration (1–6 years), the introduction of vegetation greatly improves the soil micro - environment, such as the rapid increase in the input of organic matter and nutrients, providing highly favorable conditions for the colonization and growth of a variety of bacteria, thereby enabling rapid increases in diversity (Luo et al., 2018; Ren et al., 2018; ). However, as the restoration process continues and the recovery time extends to 12 years, the ecosystem enters a new developmental stage, with significant enhancement in the intensity of environmental filtering and biotic interactions, which may lead to a decline in diversity from the previously reached peak (; Wang X. et al., 2021).
The stage - dependent changes in key environmental drivers are an important factor influencing the changes in bacterial diversity. There is a close coupling between the changes in bacterial diversity and the evolution of soil physicochemical properties. In different stages of recovery, the key environmental factors that dominate the assembly of bacterial communities are likely to change. In the early stage of recovery (1–6 years), the input of plant root exudates and litter significantly increases the content of soil organic carbon, nitrogen, phosphorus and other nutrients. The sudden enrichment of these resources can support the growth of a more diverse range of microbial communities, thereby enhancing diversity (Zheng et al., 2021; ; ). However, in the middle and late stages of recovery (e.g., 12 years), soil properties may stabilize or undergo qualitative changes. For example, soil pH may continue to change, or soil organic matter may transform from easily decomposable components to more recalcitrant ones (such as lignin and tannins). These changes create new environmental selection pressures. Bacterial groups that are not adapted to the new environment may be eliminated, while those that are more adapted to stable or specific conditions may become dominant. This may lead to a decline in overall diversity from the initial high level, but at the same time, the functions of the community may become more specialized (Wang X. et al., 2021; Zhou et al., 2024).
In addition, the shift from “r - strategy” to “K - strategy” groups is an important aspect of the systematic changes in bacterial community composition during ecological restoration. In the early stages of ecological restoration, “r - strategy” bacteria (e.g., certain groups of Proteobacteria and Firmicutes) that can rapidly utilize simple carbon sources and nutrients may proliferate in large numbers, contributing to higher diversity (; Singh et al., 2022). However, as restoration progresses, more complex and recalcitrant organic matter accumulates in the soil. The relative abundance of “K - strategy” bacteria (e.g., Acidobacteria and Actinobacteria) that are better at utilizing these refractory carbon sources and grow more slowly will gradually increase (; Zhou et al., 2024). This succession of dominant groups, even if there is a slight decline in total species richness, signifies the maturation of ecosystem functions and the enhancement of soil carbon sequestration capacity.
The feedback between vegetation and soil and the intensification of competition also affect bacterial diversity. As the vegetation community develops and matures (e.g., 12 years), the root network and litter layer of vegetation are fully established, and competition among plants for soil resources (water, nutrients) will also intensify. This competition can indirectly affect microbial communities through changes in the quality and quantity of rhizodeposits. In addition, stable plant communities will shape more specialized and tighter microbial interaction networks, such as symbiotic and competitive relationships. This may lead to the exclusion of some less competitive bacterial species, thereby reducing the α - diversity (local diversity) of the community. However, at the same time, it may increase β - diversity (differences between sites) or network complexity (Wang X. et al., 2021; Wu et al., 2023).
In long - term recovery trajectories, fluctuations and convergence are also very common. Numerous long - term sequential studies have confirmed that the recovery of soil microbial communities requires several decades, and various fluctuations may occur during this recovery process. Some studies have pointed out that it takes 15–20 years for bacterial diversity in mine restoration to approach the level of undisturbed soil. Another study lasting 52 years found that even after half a century of recovery, the structure of bacterial communities still differs from that of the original soil. Therefore, the decline in diversity at 12 years of recovery is likely just a fluctuation stage in the long - term recovery trajectory. The community structure is still constantly adjusting in order to eventually reach a new steady - state that matches the restored plant community and soil properties. The diversity level of this steady - state may not be the same as the initial peak or the original soil (; Schmid et al., 2020; Liddicoat et al., 2022).
In summary, the decline in soil bacterial diversity at 12 years of vegetation restoration compared to 6 years is the result of the combined effects of ecosystem succession and stage - dependent changes in soil environmental conditions. This does not mean a regression in restoration, but rather more likely reflects the transition of bacterial communities from the “pioneer phase” of rapid response to resource inputs to the “mature phase” of functional specialization and structural stability. In this stage, although species richness may slightly decrease, the community composition is often more adapted to the restored ecological environment, and its ecological functions, such as carbon and nitrogen cycling and the decomposition of recalcitrant organic matter, may be more efficient and stable (Wang X. et al., 2021; ; Zhou et al., 2024). When evaluating the effectiveness of restoration, it is not appropriate to rely solely on the single index of diversity. Instead, a comprehensive assessment should be conducted based on multiple - dimension indicators, including community structure, functional genes, network complexity, and soil ecological functions (; Liddicoat et al., 2022).
4.3.2 Beta diversity responses of the soil bacterial community
This study investigated the successional trends of community structure through β-diversity analysis, revealing the key role of restoration duration in explaining community differences. NMDS and PERMANOVA results clearly indicated that bacterial communities exhibited significant shifts during the early restoration stage (1 year). This stage represents a period of active response to external changes (e.g., restoration measures, natural colonization). During this phase, the community rapidly reorganizes, with a high rate of change, and differences among samples may be further amplified due to initial conditions, random colonization events, or microenvironmental variations (Shabarova et al., 2021; Shi et al., 2022; ). Subsequently, during the intermediate transitional stage (6 years), environmental conditions gradually stabilize, with competition and niche differentiation beginning to dominate. The rate of community change correspondingly slows, showing a tendency to converge toward a target state (; Ramineh et al., 2023; Wang et al., 2023). By the late stage (12 years), the community structure stabilizes, reaching or approaching a “stable state.” At this point, internal regulatory mechanisms (e.g., negative feedback) dominate, the ability to resist external disturbances (resistance) is significantly enhanced, overall structural changes are minimal, and compositions among different samples become highly similar (; Lu et al., 2023). This phenomenon of “structural convergence” may reflect the continuous role of ecological filtering mechanisms throughout the restoration process. Existing studies have also pointed out that after long-term natural restoration, a new stable state may form, although it may differ from the undisturbed original state (Lu et al., 2023).
From an ecological perspective, the findings of this study have important implications for active management interventions. During the early restoration stage, introducing key species with strong adaptability can effectively accelerate the convergence of the community toward an ideal stable state, thereby significantly shortening the time required for ecological restoration (Lopez et al., 2024). However, caution is warranted regarding the potential risks posed by “transient invaders,” which may inadvertently steer the community toward undesirable alternative stable states. Therefore, in ecological restoration practices, it is essential to comprehensively consider species selection and the complexity of community dynamics to achieve more efficient and precise ecological restoration goals (Ostertag et al., 2015; Ramírez Soto et al., 2018).
4.4 Succession of bacterial indicator taxa and functional recovery stages
This study revealed a clear successional trajectory of bacterial indicator taxa along the restoration chronosequence through LEfSe analysis. This finding not only reflects changes in soil nutrient status but also indicates a functional shift in the system from “carbon degradation” towards “stable nitrogen cycling.” Specifically, the successional pathway progresses from dominance by “oligotrophic-Chloroflexi” to “r-strategist degraders (e.g., Chitinophagia, Cytophagia)” and finally to dominance by “oligotrophic-Acidobacteria” and “ammonia-oxidizing archaea (Thaumarchaeota)”. This process illustrates the dynamic response of the microbial community to the stages of ecosystem recovery.
In the early restoration stage, many members of the Chloroflexi phylum, as typical oligotrophic bacteria, exhibit characteristics such as low metabolic activity, slow growth, but strong tolerance to harsh environments (e.g., high temperature, drought, low nutrients) (). They can survive under conditions of extremely limited resources, laying the groundwork for the initiation of subsequent ecological processes. Similarly, in the early remediation of mine tailings, taxa tolerant to heavy metals and extreme pH (e.g., certain Actinobacteria and Firmicutes) have also been observed to dominate (Singh et al., 2022). At this stage, the microbial community structure is primarily driven by strong environmental filtering (deterministic processes), meaning harsh abiotic conditions (e.g., low total carbon, sandy texture) strictly select for species capable of survival (Samia et al., 2016). Community assembly may be more influenced by stochastic processes (e.g., dispersal limitation, ecological drift) due to the scarcity and patchiness of suitable habitats (Yang et al., 2025).
With the colonization and growth of vegetation, plants begin to input available organic carbon and energy into the soil through litter inputs, root exudates (e.g., proteinaceous substances, low-molecular-weight humic substances), etc. (Zhang et al., 2021; ). The soil environment starts transitioning from an “oligotrophic” state to one with increased resource availability. This change attracts “opportunistic” or “copiotrophic” bacteria characterized by r-strategist traits. Chitinophagia and Cytophagia (belonging to the Bacteroidetes phylum) are typical representatives, known for their rapid growth rates and high efficiency in utilizing newly input, readily degradable organic matter (e.g., proteins, chitin, cellulose) (). Concurrently, the relative abundances of common copiotrophic or facultative copiotrophic phyla such as Actinobacteria and Proteobacteria increase significantly, becoming dominant groups at this stage (). These groups actively participate in the initial degradation and transformation of organic matter. For example, some members of the Proteobacteria can not only utilize small-molecule organic matter but also participate in the transformation of humic substance components. They engage in close rhizosphere interactions with plants, accelerating nutrient cycling and providing utilizable resources for plants and other microorganisms. At this stage, microbial community assembly is still strongly influenced by resource availability (e.g., dissolved organic carbon), but the role of deterministic selection may increase (; Li et al., 2024).
When vegetation restoration reaches the late stage (12 years), the ecosystem tends to stabilize. Soil organic matter accumulation reaches a high level, but most of it may have transformed into more complex, recalcitrant forms (e.g., stable humus). Nutrient forms also shift from readily available pools to more complex cycling pools. Acidobacteria are well-known oligotrophic bacteria in soils. They grow slowly but have high energy use efficiency, being particularly adept at competing in environments rich in organic matter but with low availability, decomposing complex organic compounds (; Wang W. et al., 2025). Their dominance marks the mature stage where the ecosystem shifts from rapid resource exploitation towards efficient resource conservation and recycling. Similarly, in forest succession chronosequences, the dominant soil bacteria in later stages may shift from copiotrophic (r-strategists) to oligotrophic (K-strategists) (; Wang et al., 2023; Sun et al., 2025). Simultaneously, the emergence of ammonia-oxidizing archaea (Thaumarchaeota) holds special functional significance. They are key drivers of the ammonia oxidation process, converting ammonium (NH4+) to nitrite, thereby initiating nitrification. In the late restoration stage, with the accumulation and transformation of the nitrogen pool, nitrogen cycling processes become particularly important. Their proliferation indicates that nitrogen cycling, especially nitrification, has become one of the core components of the ecosystem’s biogeochemical cycling (). Studies have also found that the dominant taxa (e.g., Actinobacteria, Rhizobia) within microbial communities harboring phosphorus cycling-related functional genes (e.g., phoD) are coupled with nitrogen fixation processes, reflecting the coordinated development of nutrient cycling (Wang F. et al., 2021). At this stage, microbial succession is more strongly driven by soil properties (e.g., total carbon, total nitrogen, C/N ratio) (Liu et al., 2020; Xu et al., 2024), plant community composition, and microbial interactions (e.g., symbiosis, competition) (). The microbial co-occurrence network may become more complex and tightly connected, reflecting more mature ecological interactions (Li et al., 2024). The role of deterministic processes (e.g., niche-based selection) in community assembly may become prominent again (Wang H. et al., 2025).
It is particularly noteworthy that Thaumarchaeota (e.g., Nitrososphaeraceae) were significantly enriched only in the 12-year restoration stage, indicating the system’s entry into a state of “nitrification homeostasis,” which can serve as a microbial marker for the functional maturity of the P. wettsteinii restoration system. This finding aligns with the previously proposed concept of “microbial functional climax,” where soil microbial communities, through long-term succession, ultimately form a functional state highly adapted to the vegetation and environmental conditions (; Wang X. et al., 2021; ).
The soil bacterial succession pathway following vegetation restoration discovered in this study finds support in various ecological restoration contexts, including arid deserts (), karst forests (), mine tailings (; Yang et al., 2025), and alpine mining areas (Wang et al., 2023). This reveals a common ecological logic underlying microbial community succession: through the replacement of their nutritional strategies and functional traits, microorganisms dynamically adapt to and drive the development of the soil habitat from a simple, harsh state towards a complex, functionally integrated one. Specifically, our analysis identified Total Nitrogen (TN) as a primary driver of bacterial community assembly, and the later stage was characterized by the emergence of nitrifying taxa (Thaumarchaeota), indicating a heightened demand for nitrogen cycling. While P. wettsteinii effectively supplies carbon via litter, introducing nitrogen-fixing shrubs during the 6- to 12-year window would biologically supplement the nitrogen pool. This creates a complementary “grass-shrub” system that optimizes the C:N balance, sustaining the “carbon-nitrogen synergy” identified in this study and accelerating the transition to a functionally mature ecosystem. This finding not only provides a new perspective for understanding the ecological mechanisms of microbial community succession but also offers significant guidance for ecological restoration practices. In future restoration efforts, greater attention should be paid to the dynamic changes and functional roles of microbial communities to achieve sustainable recovery and stable development of ecosystems.
4.5 Network structure reconstruction and community stability enhancement
In the field of ecological restoration, soil microbial communities are not only key drivers of ecosystem functions, but also serve as important indicators for assessing restoration progress and ecosystem stability due to their intricate internal interaction networks (). This study, through co-occurrence network analysis, reveals the dynamic restructuring process of bacterial community network structure during vegetation restoration, which follows an evolution from “tight–loose–reorganization.” This process profoundly reflects the response and adaptation of microbial communities to different stages of ecosystem restoration.
In the early stages of restoration, the connectivity of the bacterial community network weakens, which may be related to niche differentiation promoted by newly input organic matter. The establishment of new vegetation typically introduces fresh, readily decomposable organic matter (such as litter and root exudates) into the soil. This input of resources alters the soil’s carbon source structure and nutrient availability, thereby diminishing the interactions among existing populations. In impoverished or degraded soils, microorganisms may form extensive, tightly connected networks based on competition or mutualism to survive under resource scarcity (; ). However, the input of fresh organic matter creates new and diverse niches. Different microbial groups can differentiate and occupy distinct resource utilization pathways based on their metabolic traits (e.g., oligotrophic or copiotrophic), thereby reducing direct competition for limited common resources. This niche differentiation process weakens the obligatory interactions upon which the original populations relied for survival, leading to a decline in overall network connectivity, potentially increased path lengths, and a relatively “looser” network structure. Concurrently, rapid changes in soil physicochemical properties (e.g., pH, nutrient content, moisture) during the early restoration phase impose strong environmental filtering (). Only those microbial taxa capable of adapting to the new environment can survive and proliferate. This deterministic process selects for specific microbial populations while eliminating non-adapted ones, thereby disrupting the equilibrium of the original community, restructuring interspecies associations, and further promoting the deconstruction of the network from its original tight state (; Zhang et al., 2023; Newton et al., 2024).
As vegetation restoration progresses to later stages (e.g., 12 years), the ecosystem gradually stabilizes, and microbial community succession enters a new phase of self-organization. The emergence of keystone taxa (key nodes) is a significant feature of this stage. Following the initial environmental filtering and niche differentiation, certain microbial groups with broad adaptability, strong metabolic capabilities, or critical ecological functions become prominent, serving as keystone taxa or key nodes within the community (Yuan et al., 2023; Oggerin et al., 2024; Xie et al., 2025). For instance, in karst region vegetation restoration, genera like Bryobacter and Rhizomicrobium were identified as keystone taxa in co-occurrence networks; in the restoration process of Robinia pseudoacacia forests on the Loess Plateau, abundant microbial taxa were found to possess wider niche widths and higher network centrality (e.g., degree centrality, betweenness centrality) (Li et al., 2024). These keystone taxa act as “hubs” within the network, connecting different functional modules and significantly enhancing the overall connectivity and efficiency of the network. Simultaneously, as the plant community and soil organic matter quality stabilize (e.g., an increased proportion of recalcitrant lignin-like compounds), the functional demands of the microbial community shift from rapid utilization of simple carbon sources to cooperative degradation of complex organic matter. This fosters the establishment of symbiotic or mutualistic relationships based on functional complementarity. For example, in the direct revegetation of a zinc smelting slag site, selective associations were observed between the bacterial community and specific dissolved organic matter molecules, forming complex interaction networks to adapt to the harsh environment (Zhou et al., 2024). Such close cooperation based on functionality leads to new, more efficient interaction patterns within the microbial community, manifested as a recovery in network connectivity, shortened average path lengths, and more efficient flow of information and materials (; ; Li et al., 2024). The reorganized microbial network often exhibits higher complexity and stability characteristics. Studies generally find that successful vegetation restoration promotes an increase in the number of nodes and connections in soil microbial (especially bacterial) co-occurrence networks, accompanied by modular structures and “small-world” properties (). Modularity implies the formation of several functional subgroups within the network, with tight internal connections within subgroups and connections between subgroups maintained by a few key nodes (). This structure ensures specialization and efficiency of local functions while maintaining global connectivity and robustness, making the ecosystem more resistant to external disturbances.
In summary, the dynamic restructuring of the bacterial community network along a “tight–loose–reorganization” trajectory during vegetation restoration represents a typical microcosm of ecosystem succession. In the early restoration stage, the “disturbance” from resource input and environmental filtering trigger niche differentiation, dissolving the original tight network built upon competition for scarce resources. As restoration progresses into the mid to late stages, with stabilizing environmental conditions and changing resource quality, the microbial community undergoes self-organization. This process selects keystone taxa with critical ecological functions as new hubs and rebuilds a novel interaction network based on functional complementarity, characterized by greater complexity, efficiency, modularity, and stability. This process profoundly reveals that soil microbial communities are not passive responders to environmental changes but are dynamic life-support networks capable of actively adjusting their internal association structures to enhance ecosystem functions (such as carbon sequestration, nutrient cycling) and resilience. Understanding this dynamic pattern holds crucial theoretical value for evaluating the effectiveness of ecological restoration and predicting the long-term trajectory of ecosystem evolution. Understanding this dynamic pattern holds crucial theoretical value for evaluating the effectiveness of ecological restoration and predicting the long-term trajectory of ecosystem evolution. However, it is important to acknowledge that the relatively small sample size (n = 3 per stage) used for network construction may limit the statistical power to detect weak correlations, and future studies with larger sample sizes are warranted to validate these network topologies.
4.6 Regulatory mechanisms of soil factors on bacterial community assembly
During ecological restoration, the assembly of soil microbial communities is influenced by a combination of various environmental factors, among which soil total nitrogen (TN), organic matter (OM), and available potassium (AK) are key drivers. This study, through RDA analysis, reveals the important roles and synergistic regulatory mechanisms of these factors in the process of vegetation restoration, providing a new perspective for understanding the dynamic changes in microbial communities.
Soil total nitrogen (TN) is one of the key factors affecting the composition of rhizosphere soil bacterial communities. In studies of Ammodendron bifolium, soil total nitrogen is one of the key soil factors influencing its rhizobacterial community (Zhu et al., 2023). Similarly, in the short-term restoration process of mine-reclaimed soil, TN was also identified as an important environmental variable affecting bacterial community structure. This suggests that changes in TN levels can select for bacterial taxa with different nitrogen metabolic capacities, thereby altering community composition.
Soil organic matter (OM), as the primary source of energy and carbon for heterotrophic microorganisms, directly determines microbial activity and community structure through its content and composition. Studies have confirmed that OM, together with TN, TP, and others, explains most of the variation in bacterial community composition during vegetation restoration (Xu et al., 2019). By promoting soil aggregate formation and improving water and nutrient retention capacity, OM creates a more suitable microhabitat for microorganisms (Liu et al., 2024). In the vegetation restoration chronosequence of degraded granite red soil, soil organic carbon (SOC, the main component of OM) content showed a significant positive correlation with the abundance of bacteria and archaea, and together they promoted the enhancement of soil aggregate stability. This illustrates the pathway through which OM indirectly drives microbial community assembly by improving the physical environment.
Available potassium (AK), in its readily utilizable form for plants and microorganisms, significantly affects the metabolic activities of microbial communities. AK often forms strong coupling relationships with other factors (such as TN, OM) to jointly drive microbial community assembly. Studies have shown that AK often jointly explains the variation in microbial community structure alongside other nutrient indicators such as TN and OM. In a temporal study of larch forests, soil bacterial community structure was significantly influenced by AK, soil organic matter (SOM), and TN (Zhang et al., 2023). Similarly, in soils subjected to long-term polycyclic aromatic hydrocarbon (PAH) contamination, AK, along with total organic carbon (TOC), pH, and other factors, was identified as a primary environmental driver of microbial community changes.
The RDA analysis in this study indicates that total nitrogen (TN), organic matter (OM), and available potassium (AK) are key drivers of bacterial community assembly, collectively explaining over 75% of the community variation. TN and OM primarily drive community distribution along the first axis, reflecting the central role of “carbon-nitrogen synergy” in community assembly; while AK independently acts on the second axis, likely influencing bacterial growth strategies by regulating cell K+-ATPase activity. This “resource-environment” dual-axis regulatory mechanism reveals that vegetation restoration guides the microbial community towards functional maturity by simultaneously enhancing resource supply and optimizing habitat conditions. This finding holds significant practical guidance. In ecological restoration projects, attention should not only be paid to increasing vegetation coverage but also to promoting the virtuous cycle and accumulation of key soil nutrient elements through reasonable vegetation configuration and management measures (such as selecting species with nitrogen-fixing capabilities or well-developed fine roots), thereby positively guiding the restoration of soil microbial communities and accelerating the holistic recovery of the structure and function of degraded ecosystems.
This finding holds significant practical guidance. In ecological restoration projects, attention should not only be paid to increasing vegetation coverage but also to promoting the virtuous cycle and accumulation of key soil nutrient elements through reasonable vegetation configuration and management measures (such as selecting species with nitrogen-fixing capabilities or well-developed fine roots), thereby positively guiding the restoration of soil microbial communities and accelerating the holistic recovery of the structure and function of degraded ecosystems. It is important to acknowledge that the soil samples in this study were collected at a single time point (August), representing a “summer snapshot” of the bacterial community. Since bacterial communities can exhibit significant seasonal fluctuations driven by changes in temperature and precipitation, future research should incorporate multi-season sampling to fully elucidate the seasonal dynamics of microbial succession during vegetation restoration.
5 Conclusion
In summary, the vegetation restoration using P. wettsteinii has significantly improved soil structure and nutrient status through litter input, root activity, and microbial processes, thereby driving an increase in bacterial diversity and functional succession. The soil physicochemical and microbial systems exhibit a distinct characteristic of “stage-specific synergistic recovery,” with the period of 6–12 years serving as a critical transition phase where the system gradually shifts from “structural recovery” to “functional stabilization.” Future research could further integrate metagenomic and functional gene microarray technologies to deeply analyze the succession mechanisms of microbial functional genes during the restoration process. Concurrently, it is recommended to introduce nitrogen-fixing shrubs or leguminous plants during this critical phase to strengthen the carbon-nitrogen coupling network and enhance the efficiency of system restoration as well as its ecological service functions.
Statements
Data availability statement
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.
Author contributions
YW: Data curation, Funding acquisition, Investigation, Project administration, Visualization, Writing – original draft. XL: Formal Analysis, Investigation, Writing – review and editing.
Funding
The author(s) declared that financial support was received for this work and/or its publication. This work was supported by the Longyan Science and Technology Project (grant numbers: 2018LYF5010).
Acknowledgments
The authors would like to express their gratitude to EditSprings (https://www.editsprings.cn) for the expert linguistic services provided.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fenvs.2026.1808649/full#supplementary-material
SUPPLEMENTARY FIGURE S1Correlation matrix (Pearson) of soil physicochemical properties.
SUPPLEMENTARY FIGURE S2Co-occurrence network of soil bacterial communities.
SUPPLEMENTARY FIGURE S3Redundancy analysis (RDA) ordination diagram between soil physicochemical properties and bacterial communities at the genus level.
References
1
AnF.NiuZ.LiuT.SuY. (2022). Succession of soil bacterial community along a 46-year choronsequence artificial revegetation in an arid oasis-desert ecotone. Sci. Total Environ.814, 152496. 10.1016/j.scitotenv.2021.152496
2
BarberN. A.KlimekD. M.BellJ. K.SwingleyW. D. (2023). Restoration age and reintroduced bison may shape soil bacterial communities in restored tallgrass prairies. FEMS Microbiol. Ecol.99, fiad007. 10.1093/femsec/fiad007
3
BhattacharyyaS. S.RosG. H.FurtakK.IqbalH. M. N.Parra-SaldívarR. (2022). Soil carbon sequestration – an interplay between soil microbial community and soil organic matter dynamics. Sci. Total Environ.815, 152928. 10.1016/j.scitotenv.2022.152928
4
ChaerG. M.FernandesM. F.MyroldD. D.BottomleyP. J. (2009). Shifts in microbial community composition and physiological profiles across a gradient of induced soil degradation. Soil Sci. Soc. Am. J.73, 1327–1334. 10.2136/sssaj2008.0276
5
ChenW.ZhaoJ.CaoC.TianH. (2018). Shrub biomass estimation in semi-arid sandland ecosystem based on remote sensing technology. Glob. Ecol. Conserv.16, e00479. 10.1016/j.gecco.2018.e00479
6
CongJ.YangY.LiuX.LuH.LiuX.ZhouJ.et al (2015). Analyses of soil microbial community compositions and functional genes reveal potential consequences of natural forest succession. Sci. Rep.5, 10007. 10.1038/srep10007
7
DayA. M.NjeruE. M.DavidJ. (2024). Soil health and ecological restoration. Burleigh Dodds Ser. Agric. Sci., 163–196. 10.19103/as.2024.0134.06
8
DongX.ChenM.ChenQ.LiuK.LongJ.LiY.et al (2024). Rare microbial taxa as the major drivers of nutrient acquisition under moss biocrusts in karst area. Front. Microbiol.15, 1384367. 10.3389/fmicb.2024.1384367
9
EichorstS. A.BreznakJ. A.SchmidtT. M. (2007). Isolation and characterization of soil bacteria that define terriglobus gen. nov., in the phylum acidobacteria. Appl. Environ. Microbiol.73, 2708–2717. 10.1128/aem.02140-06
10
FanM.LiJ.TangZ.ShangguanZ. (2020). Soil bacterial community succession during desertification in a desert steppe ecosystem. Land Degrad. Dev.31, 1662–1674. 10.1002/ldr.3545
11
FanL.ZhaoW.FengW.MoP.ZhaoY.YangG.et al (2021). Insight into the characteristics of soil microbial diversity during the ecological restoration of mines: a case study in dabaoshan mining area, China. Sustainability13, 11684. 10.3390/su132111684
12
FengS.ChenM.GaoM.LiuM.WangK.WangJ.et al (2024). Soil microbial community construction under revegetation in newly created land. Sci. Total Environ.954, 176496. 10.1016/j.scitotenv.2024.176496
13
GrahamE. B.KnelmanJ. E. (2023). Implications of soil microbial community assembly for ecosystem restoration: patterns, process, and potential. Microb. Ecol.85, 809–819. 10.1007/s00248-022-02155-w
14
GuoA.ZhaoZ. (2022). Microbial community response to soil degradation with different geographical distribution. SSRN Electron. J. 10.2139/ssrn.4231444
15
GuoY.ChenX.WuY.ZhangL.ChengJ.WeiG.et al (2018). Natural revegetation of a semiarid habitat alters taxonomic and functional diversity of soil microbial communities. Sci. Total Environ.635, 598–606. 10.1016/j.scitotenv.2018.04.171
16
GuptaR.SinghA.SrivastavaM.ShankerK.PandeyR. (2019). Plant-microbe interactions endorse growth by uplifting microbial community structure of bacopa monnieri rhizosphere under nematode stress. Microbiol. Res.218, 87–96. 10.1016/j.micres.2018.10.006
17
HaoC.DuP.RenJ.HuL.ZhangZ. (2023). Halophyte elymus dahuricus colonization regulates microbial community succession by mediating saline-alkaline and biogenic organic matter in bauxite residue. Sci. Total Environ.905, 167140. 10.1016/j.scitotenv.2023.167140
18
HeG.PengT.GuoY.WenS.JiL.LuoZ. (2022). Forest succession improves the complexity of soil microbial interaction and ecological stochasticity of community assembly: evidence from phoebe bournei-dominated forests in subtropical regions. Front. Microbiol.13, 1021258. 10.3389/fmicb.2022.1021258
19
HeY.WenY.LiK.YeS.ZhangH.HeF.et al (2025). Responses of soil multifunctionality, microbial diversity, and network complexity to tree species mixing in eucalyptus plantations. Ind. Crop. Prod.225, 120575. 10.1016/j.indcrop.2025.120575
20
HuL.LiQ.YanJ.LiuC.ZhongJ. (2022). Vegetation restoration facilitates belowground microbial network complexity and recalcitrant soil organic carbon storage in southwest China karst region. Sci. Total Environ.820, 153137. 10.1016/j.scitotenv.2022.153137
21
HuT.XiongK.YuY.WangJ.WuY. (2023). Ecological stoichiometry and homeostasis characteristics of plant-litter-soil system with vegetation restoration of the karst desertification control. Front. Plant Sci.14, 11. 10.3389/fpls.2023.1224691
22
IshimotoC. K.AonoA. H.NagaiJ. S.SousaH.MirandaA. R. L.MeloV. M. M.et al (2021). Microbial co-occurrence network and its key microorganisms in soil with permanent application of composted tannery sludge. Sci. Total Environ.789, 147945. 10.1016/j.scitotenv.2021.147945
23
JiaT.LiangX.GuoT.WuT.ChaiB. (2022). Bacterial community succession and influencing factors for imperata cylindrica litter decomposition in a copper tailings area of China. Sci. Total Environ.815, 152908. 10.1016/j.scitotenv.2021.152908
24
JiangS.SongM.DuH.WangF.SongT.ChenH.et al (2024). Soil properties regulate soil microbial communities during forest succession in a karst region of southwest China. Microorganisms12, 2136. 10.3390/microorganisms12112136
25
JiangH.ZhouY.LiW.LuQ.XuD.MaH.et al (2025). Reseeding native species promotes community stability by improving species diversity, niche, and interspecific relationships in the desert steppe of northwest China. Ecol. Evol.15, e70929. 10.1002/ece3.70929
26
JingZ.LiQ.WeiY.DongB.YuanP.LiuR. X.et al (2023). Mechanistic insights into dissolved organic matter-driven protistan and bacterial community dynamics influenced by vegetation restoration. Environ. Res.227, 115710. 10.1016/j.envres.2023.115710
27
KongL.ZhangL.WangY.HuangZ. (2023). Impact of ecological restoration on the physicochemical properties and bacterial communities in alpine mining area soils. Microorganisms12, 41. 10.3390/microorganisms12010041
28
KumarA.DasA.SinghD.DasM. K.SrivastavaG. P.SinghJ. P.et al (2023). Soil health restoration in degraded lands: a microbiological perspective. Land Degrad. Dev.34, 5155–5170. 10.1002/ldr.4851
29
LawesM. J.MooreA. M.AndersenA. N.PreeceN. D.FranklinD. C. (2017). Ants as ecological indicators of rainforest restoration: community convergence and the development of an ant forest indicator index in the australian wet tropics. Ecol. Evol.7, 8442–8455. 10.1002/ece3.2992
30
LiY.WenH.ChenL.YinT. (2014). Succession of bacterial community structure and diversity in soil along a chronosequence of reclamation and re-vegetation on coal mine spoils in China. PLoS One9, e115024. 10.1371/journal.pone.0115024
31
LiJ.FanM.WeiZ.ZhangK.MaX.ShangguanZ. (2024). Broad environmental adaptation of abundant microbial taxa in robinia pseudoacacia forests during long-term vegetation restoration. Environ. Res.242, 117720. 10.1016/j.envres.2023.117720
32
LiddicoatC.KraussS. L.BissettA.BorrettR. J.DuckiL. C.PeddleS. D.et al (2022). Next generation restoration metrics: using soil eDNA bacterial community data to measure trajectories towards rehabilitation targets. J. Environ. Manage.310, 114748. 10.1016/j.jenvman.2022.114748
33
LiuJ.JiaX.YanW.ZhongY.ShangguanZ. (2020). Changes in soil microbial community structure during long‐term secondary succession. Land Degrad. Dev.31, 1151–1166. 10.1002/ldr.3505
34
LiuS.LinZ.DuanX.DengY. (2024). Effects of soil microorganisms on aggregate stability during vegetation recovery in degraded granitic red soil areas. Appl. Soil Ecol.204, 105734. 10.1016/j.apsoil.2024.105734
35
LopezK. P.HovickS. M.MattinglyK. Z.WestonL. M.HoffordN. P.FinleyL.et al (2024). Suppression of reed canarygrass by assisted succession: a sixteen‐year restoration experiment. J. Appl. Ecol.61, 1805–1816. 10.1111/1365-2664.14705
36
LuD.MaoZ.TangY.FengB.XuL. (2023). Driving factors influencing soil microbial community succession of coal mining subsidence areas during natural recovery in inner Mongolia grasslands. Microorganisms12, 87. 10.3390/microorganisms12010087
37
LuoY.WuY.WangH.XingR.ZhengZ.QiuJ.et al (2018). Bacterial community structure and diversity responses to the direct revegetation of an artisanal zinc smelting slag after 5 years. Environ. Sci. Pollut. Res.25, 14773–14788. 10.1007/s11356-018-1573-6
38
MengX.CaoJ.WangX.ZhangC.LvJ. (2021). Spatial characteristics of the human factors of soil erosion at the boundary of political divisions: a spatial approach. Catena201, 105278. 10.1016/j.catena.2021.105278
39
MiaoL.WangC.AdyelT. M.ZhaoJ.YanN.WuJ.et al (2021). Periphytic biofilm formation on natural and artificial substrates: comparison of microbial compositions, interactions, and functions. Front. Microbiol.12, 684903. 10.3389/fmicb.2021.684903
40
NewtonD.HoP.HuangK. C. (2024). The network structure of cross-feeding impacts microbial community diversity under growth-inhibiting stresses. Biorxiv Cold Spring Harb. Lab. 10.1101/2024.10.29.620519
41
OggerinM.ViverT.BrüwerJ.VoßD.García-LlorcaM.ZielinskiO.et al (2024). Niche differentiation within bacterial key-taxa in stratified surface waters of the southern pacific gyre. ISME J.18, wrae155. 10.1093/ismejo/wrae155
42
OstertagR.WarmanL.CordellS.VitousekP. M. (2015). Using plant functional traits to restore hawaiian rainforest. J. Appl. Ecol.52, 805–809. 10.1111/1365-2664.12413
43
QiQ.ZhaoM.WangS.MaX.WangY.GaoY.et al (2017). The biogeographic pattern of microbial functional genes along an altitudinal gradient of the tibetan pasture. Front. Microbiol.8, 976. 10.3389/fmicb.2017.00976
44
RaminehA.JourgholamiM.EtemadV.JafariM.PicchioR. (2023). Effect of different vegetation restoration on recovery of compaction-induced soil degradation in hyrcanian mixed forests: influence on soil c and n pools and enzyme activities. Forests14, 603. 10.3390/f14030603
45
Ramírez SotoA.Lucio PalacioC. R.Rodríguez MesaR.Sheseña HernándezI.FarhatF. N.Villa‐BonillaB.et al (2018). Restoration of tropical montane cloud forests: a six‐prong strategy. Restor. Ecol.26, 206–211. 10.1111/rec.12660
46
RauppP.CarrilloY.NielsenU. N. (2024). Soil health to enhance ecological restoration and conservation. J. Sustain. Agric. Environ.3, e70022. 10.1002/sae2.70022
47
RenQ.LiC.YangW.SongH.MaP.WangC.et al (2018). Revegetation of the riparian zone of the three gorges dam reservoir leads to increased soil bacterial diversity. Environ. Sci. Pollut. Res.25, 23748–23763. 10.1007/s11356-018-2333-3
48
SamiaI.PhilippeL.AbdelkaderB.ZohraI.ElisabethN. (2016). Effect of anthropisation and revegetation efforts on soil bacterial community structure in terga sandpit, Algeria. J. Biodivers. Environ. Sci.9, 283–296.
49
SchmidC. A. O.ReichelR.SchröderP.BrüggemannN.SchloterM. (2020). 52 years of ecological restoration following a major disturbance by opencast lignite mining does not reassemble microbiome structures of the original arable soils. Sci. Total Environ.745, 140955. 10.1016/j.scitotenv.2020.140955
50
ShabarovaT.SalcherM. M.PorcalP.ZnachorP.NedomaJ.GrossartH. P.et al (2021). Recovery of freshwater microbial communities after extreme rain events is mediated by cyclic succession. Nat. Microbiol.6, 479–488. 10.1038/s41564-020-00852-1
51
ShiW.XingY.ZhuY.GaoN.YingY. (2022). Diverse responses of pqqc ‐ andphod ‐harbouring bacterial communities to variation in soil properties of moso bamboo forests. Microb. Biotechnol.15, 2097–2111. 10.1111/1751-7915.14029
52
SinghP.JainK. R.LakhmapurkarJ.GavaliD.DesaiC.MadamwarD. (2022). Microbial community structure and functions during chronosequence-based phytoremediation programme of lignite tailing soil. Environ. Technol. Innov.27, 102447. 10.1016/j.eti.2022.102447
53
SudrajatS.WidhayasaB.RusdiansyahR.SusantoD. (2019). Rhizosphere fungal community, soil physicochemical properties, understorey vegetation and their relationship during post-coal mining reclamation in East Kalimantan, Indonesia. Biodivers. J. Biol. Divers.20. 10.13057/biodiv/d200723
54
SunW.XiaoE.KruminsV.HäggblomM. M.DongY.PuZ.et al (2018). Rhizosphere microbial response to multiple metal(loid)s in different contaminated arable soils indicates crop-specific metal-microbe interactions. Appl. Environ. Microbiol.84, e00701–e00718. 10.1128/aem.00701-18
55
SunH.SunF.DengX.StornN. (2025). Soil carbon fractions drive microbial community assembly processes during forest succession. J. Environ. Manage.373, 123638. 10.1016/j.jenvman.2024.123638
56
WangX.LiuJ.HeZ.XingC.ZhuJ.GuX.et al (2021). Forest gaps mediate the structure and function of the soil microbial community in a castanopsis kawakamii forest. Ecol. Indic.122, 107288. 10.1016/j.ecolind.2020.107288
57
WangH.LiuH.YangT.LvG.LiW.ChenY.et al (2023). Mechanisms underlying the succession of plant rhizosphere microbial community structure and function in an alpine open-pit coal mining disturbance zone. J. Environ. Manage.325, 116571. 10.1016/j.jenvman.2022.116571
58
WangH.CrowtherT. W.IsobeK.WangH.TatenoR.ShiW. (2025). Niche conservatism and community assembly reveal microbial community divergent succession between litter and topsoil. Mol. Ecol.34, e17723. 10.1111/mec.17723
59
WangW.BiS.LiF.DegenA. A.LiS.HuangM.et al (2025). Soil organic matter composition affects ecosystem multifunctionality by mediating the composition of microbial communities in long-term restored meadows. Environ. Microbiome20, 22. 10.1186/s40793-025-00678-6
60
WangF.ZhangY.XiaY.CuiZ.CaoC. (2021). Soil microbial community succession based on PhoD and gcd genes along a chronosequence of sand-fixation forest. Forests12, 1707. 10.3390/f12121707
61
WuF.WangY.SunH.ZhouJ.LiR. (2023). Reforestation regulated soil bacterial community structure along vertical profiles in the loess plateau. Front. Microbiol.14, 1324052. 10.3389/fmicb.2023.1324052
62
WuL.RenC.JiangH.ZhangW.ChenN.ZhaoX.et al (2024). Land abandonment transforms soil microbiome stability and functional profiles in apple orchards of the chinese losses plateau. Sci. Total Environ.906, 167556. 10.1016/j.scitotenv.2023.167556
63
XieP.WangQ.QuW.ChenX.FengY.MaJ.et al (2025). Revealing real impact of microalgae on seasonal dynamics of bacterial community in a pilot-scale microalgal-bacterial consortium system. Water Res.274, 123145. 10.1016/j.watres.2025.123145
64
XuH.QuQ.LiP.GuoZ.WulanE.XueS. (2019). Stocks and stoichiometry of soil organic carbon, total nitrogen, and total phosphorus after vegetation restoration in the loess hilly region, China. Forests10, 27. 10.3390/f10010027
65
XuM.XiJ.LiuY.LiS. (2024). Adaptation strategies of soil microorganisms in resource changes and stoichiometric imbalances induced by secondary succession on the loess plateau. J. Environ. Manage.370, 122668. 10.1016/j.jenvman.2024.122668
66
YangS.LiuB.WangL.DuranR. (2025). Dispatched microbial community assembly processes driving ecological succession during phytostabilization of mercury-rich tailings. Environ. Pollut.365, 125376. 10.1016/j.envpol.2024.125376
67
YuanS.DuH.ZhaoD.QiaoZ.ZhengJ.YuX.et al (2023). Stochastic processes drive the assembly and metabolite profiles of keystone taxa during chinese strong-flavor baijiu fermentation. Microbiol. Spectr.11, e05103-22. 10.1128/spectrum.05103-22
68
ZhangH.YuD.DongL.ShiX.WarnerE.GuZ.et al (2014). Regional soil erosion assessment from remote sensing data in rehabilitated high density canopy forests of southern China. Catena123, 106–112. 10.1016/j.catena.2014.07.013
69
ZhangX.WangL.ZhouW.HuW.HuJ.HuM. (2021). Changes in litter traits induced by vegetation restoration accelerate litter decomposition in robinia pseudoacacia plantations. Land Degrad. Dev.33, 179–192. 10.1002/ldr.4136
70
ZhangJ.LiuQ.WangD.ZhangZ. (2023). Soil microbial community, soil quality, and productivity along a chronosequence of larix principis-rupprechtii forests. Plants12, 2913. 10.3390/plants12162913
71
ZhaoZ.RenK.GaoY.ZhaoM.ZhouL.HuoS.et al (2024). Changes in soil inorganic carbon following vegetation restoration in the cropland on the loess plateau in China: a meta-analysis. J. Environ. Manage.372, 123412. 10.1016/j.jenvman.2024.123412
72
ZhengX.YanM.LinC.GuoB.DingH.YuJ.et al (2021). Vegetation restoration types affect soil bacterial community composition and diversity in degraded lands in subtropical of China. Restor. Ecol.30, e13494. 10.1111/rec.13494
73
ZhengW.GuoX.ZhouP.TangL.LaiJ.DaiY.et al (2024). Vegetation restoration enhancing soil carbon sequestration in karst rocky desertification ecosystems: a meta-analysis. J. Environ. Manage.370, 122530. 10.1016/j.jenvman.2024.122530
74
ZhouD.LuoY.LuoY.HeY.ChenY.WanZ.et al (2024). Chemodiversity of dissolved organic matter and its association with the bacterial community at a zinc smelting slag site after 10 years of direct revegetation. Sci. Total Environ.950, 175322. 10.1016/j.scitotenv.2024.175322
75
ZhuY.HuangY.NuerhamantiN.BaiX.WangH.ZhuX. Y.et al (2023). The composition and diversity of the rhizosphere bacterial community of ammodendron bifolium growing in the takeermohuer desert are different from those in the nonrhizosphere. Microb. Ecol.87, 2. 10.1007/s00248-023-02320-9
Summary
Keywords
bacteria, diversity, Paspalum wettsteinii, soil ecology, vegetation restoration
Citation
Wang Y and Lin X (2026) Temporal dynamics of soil physicochemical properties and bacterial community assembly during 12-year vegetation by planting Paspalum wettsteinii on degraded mountainous. Front. Environ. Sci. 14:1808649. doi: 10.3389/fenvs.2026.1808649
Received
11 February 2026
Revised
03 May 2026
Accepted
05 May 2026
Published
29 May 2026
Volume
14 - 2026
Edited by
Chong Jiang, Guangdong Academy of Science (CAS), China
Reviewed by
Umar Daraz, Gansu Agricultural University, China
Jawdat Bakr, Sulaimani Polytechnic University, Iraq
Updates
Copyright
© 2026 Wang and Lin.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Xiufeng Lin, 307052955@qq.com
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.