A field experiment was conducted to study the changes in soil characteristics in cultivated grasslands in Qinghai Province, China. The experimental site was located near Xi Hai Town, Haibei Prefecture, Qinghai Province (elevation 3,156 m). The region has a plateau continental climate, with an average annual temperature of 0.9 °C. The average annual precipitation is 369.1 mm, and there is no absolute frost-free period. This location is representative of the Qinghai - Tibet Plateau. The high elevation and characteristic plateau continental climate, with its specific temperature and precipitation patterns, are typical of the broader region. Therefore, the selected site can effectively serve as a substitute for the Qinghai - Tibet Plateau in studying the changes in soil characteristics of cultivated grasslands.

Spatial - temporal substitution is a method that uses spatial differences to infer temporal changes. Zhao et al. used this approach to investigate the impact of different cultivation measures on soil quality over time. By comparing grasslands planted at different times but in similar locations, they deduced the temporal changes in soil quality. This method is beneficial because it can effectively eliminate the impact of inter - annual climate variations on soil quality, which are significant in the study region. In this study, a “spatial-temporal substitution” approach was employed. Cultivated grasslands established in 2014, 2018, and 2022 were selected as representative stages of the typical soil succession process (2014 represents the later stage of restoration, 2018 represents the middle stage of restoration, and 2022 represents the early stage of restoration). Additionally, un-degraded natural grassland (fenced and protected for 5 years) was selected as the control (CK) treatment. The slope direction, elevation, and precipitation conditions were similar among the different grassland plots ( Figure 1 ).

The cultivation of different grasslands was carried out using strip sowing with a row spacing of 30 cm and a seeding rate of 22.5 kg/hm². Before sowing, urea at a rate of 75 kg/hm² and diammonium phosphate at a rate of 150 kg/hm² were applied, and the cultivation was rainfed. Each plot had an area of 50 m² and was replicated three times. After seeding, one weed removal operation is conducted in the same year, followed by enclosure.

In 2023, rhizosphere soil samples were collected from cultivated grasslands at different restoration stages of Kentucky bluegrass and from undisturbed natural grasslands. For each treatment, five random replicate samples were collected, and then homogenized to provide a composite sample for each replicate site. Portions of soil were air-dried for characterization analysis, while subsamples for molecular and enzyme activity analysis were preserved with liquid nitrogen, transported in iceboxes, and stored at -80°C and 4°C, respectively.

Using the MagPure Soil DNA LQ Kit (Magan) kit, following the instructions, the genomic DNA of the samples was extracted. The concentration and purity of the DNA were determined using NanoDrop 2000 (Thermo Fisher Scientific, USA) and agarose gel electrophoresis, and the extracted DNA was stored at -20°C. The extracted genomic DNA was used as a template for PCR amplification of bacterial 16S rRNA genes and ITS genes. The V3-V4 variable region of the 16S rRNA gene was amplified using the universal primers 343F (5’-TACGGRAGGCAGCAG-3’) and 798R (5’-AGGGTATCTAATCCT-3’), while the ITS1 variable region of the ITS gene was amplified using the universal primers ITS1F (5’-CTTGGTCATTTAGAGGAAGTAA-3’) and ITS2 (5’-GCTGCGTTCTTCATCGATGC-3’).

The PCR amplification products were analyzed using agarose gel electrophoresis. The AMPure XP beads were used for purification, and the purified products were used as templates for the second-round PCR amplification. Subsequently, another round of purification with magnetic beads was conducted, and the purified second-round products were quantified using Qubit. The concentration was adjusted for sequencing. Sequencing was performed on the Illumina NovaSeq 6000 sequencing platform, generating 250 bp paired-end reads. The sequencing was conducted by Shanghai Oe Biotech Co., Ltd.

After the raw data is sequenced, the first step involves using the Cutadapt software to trim off the primer sequences from the raw data sequences. Subsequently, DADA2 is utilized for quality filtering, denoising, merging, and chimera removal of the paired-end raw data, conducting quality control analysis to obtain representative sequences and abundance tables. Finally, the QIIME 2 software package is employed to select representative sequences of each Amplicon Sequence Variant (ASV) and annotate all representative sequences by comparing them to the Silva database.

This study uses Principal Component Analysis (PCA) and correlation analysis to determine the Multidimensional Scaling (MDS) values for soil quality assessment (Zhang Z. et al., 2022). The Norm value is determined by the formula:

where Nik is the comprehensive load on the first k principal components with the characteristic value of the ith variable ≥1. Uik is the load of the i-th variable on the k-th principal component, which reflects the importance of the i-th variable in the k principal component. λk is the kth principal component eigenvalue.

The Soil Quality Index (SQI) is calculated through an equation that determines the weights and scores of soil evaluation indicators.

where N is the number of indicators. S_i is the membership value of the soil index. P_i represents the contribution rate of the i-th comprehensive index for each treatment. W_i is the indicator weight (Refer to Tian et al., 2018 for calculation method). The SQI ranges from 0 to 1 (Zhang F. et al., 2022).

Due to the varying impact of different soil physicochemical properties on soil quality, the membership functions for different soil indicators also differ. Membership functions can be classified as ascending and descending types. The types of membership functions for different soil physicochemical indicators are shown in Table 1 .

The Mothur software package (v.1.30.1) was used to estimate community diversity using the Shannon index and calculate classification α diversity. Non-metric Multidimensional Scaling (NMDS) was employed to describe the clustering of different samples, further reflecting the changes in microbial community structure under intercropping patterns, i.e., microbial diversity. Redundancy Analysis (RDA) was conducted to study the correlations between soil microbial composition, soil nutrients, soil pH, soil bulk density and soil enzyme activities. The RDA analysis was performed using the CANOCO 4.5 software package. Spearman correlation analysis was conducted using IBM SPSS Statistics 19.0 to determine the relationships between soil microbial characteristics (abundance, α diversity, and β diversity) and soil properties (Soil nutrients, soil pH, soil bulk density) and enzyme activities. One-way Analysis of Variance (ANOVA) was used to analyze differences in soil properties (Soil nutrients, soil pH, soil bulk density), soil enzyme activities, and microbial characteristics under different intercropping patterns (P< 0.05). Mean differences were assessed using the LSD test (P< 0.05), where different letters indicate significance. Origin 9.1 was used for data visualization, including soil nutrients and enzyme activities.

The cultivated grassland in different management stages significantly influenced the content of SOC, C stock, TN, and TP (P<0.05) ( Figures 2a–f ). In the second year of grassland cultivation (early restoration stage), the content of SOC, C stock, TN, and TP in the cultivated grassland was significantly lower compared to the CK treatment. With the increase in restoration time, the SOC, C stock, TN, and TP of the cultivated grassland showed an increasing trend ( Figures 2c–f ). Gradually, the differences compared to the CK treatment became non-significant (P>0.05) ( Figures 2a–d ).

For soil enzyme activity, the activity of SALPT (alkaline protease), SBAX (asic xylanase), and SUE in cultivated grasslands at different management stages significantly influenced (P<0.05), with similar trends of changes in activity. They all showed an increase followed by a decrease with increasing years of management ( Figures 2i–k ). In the second year of grassland cultivation, the activity of SALPT (alkaline protease) did not differ significantly from the CK treatment (P>0.05). With the increase in management time, the activity of SALPT (alkaline protease) significantly increased in the sixth and tenth year compared to the CK treatment (P<0.05) ( Figure 2i ). As for SBAX, the activity did not differ significantly from the CK treatment in the second and tenth year of management, but the activity in the sixth year of cultivated grassland significantly increased compared to the CK treatment (P<0.05) ( Figure 2j ). Regarding SUE, the activity did not differ significantly from the CK treatment in the sixth and tenth year of management, but in the second year of cultivated grassland, the activity of SUE significantly decreased compared to the CK treatment (P<0.05) ( Figure 2k ).

In addition, the activity of SCL (cellulase) and SSC (sucrase) in cultivated grasslands with different planting years significantly influenced (P<0.05). The activities of SCL (cellulase) and SCAT (catalase) decreased with the increase in management years ( Figures 2g, h ). In the second year of grassland cultivation, the activity of SCL (cellulase) was significantly higher than the CK treatment (P<0.05), but with the increase in management time, the activity of SCL (cellulase) did not differ significantly from the CK treatment (P>0.05) ( Figures 2g, h ). The activities of SSC (sucrase) and SAKP (alkaline phosphatase) showed an increasing trend with the increase in planting years, but the impact on SAKP (alkaline phosphatase) activity was not significant (P>0.05) ( Figures 2l, m ). The activity of SSC (sucrase) decreased the most with the increase in management time, and in the tenth year of cultivated grassland, the activities of SCL (cellulase) and SSC (sucrase) were gradually significantly higher than the CK treatment (P<0.05) ( Figure 2l ).

In the second year of grassland cultivation, the ACE index of bacterial and fungal communities was significantly lower than the CK treatment (P<0.05). However, with prolonged management, the ACE index gradually increased. By the tenth year, the ACE index of the bacterial community no longer showed significant differences compared to the CK treatment (P > 0.05), while the fungal community exhibited a significantly higher ACE index than the CK treatment (P< 0.05) ( Figure 3a ). Similar trends were observed for the Pielou and Shannon indices. In the second year, both indices were significantly lower than those of the CK treatment (P< 0.05). But with increasing management duration, their values gradually increased, and by the tenth year, no significant differences were observed compared to the CK treatment (P > 0.05) ( Figures 3b, c ).

Non-metric multidimensional scaling (NMDS) analysis revealed distinct clustering patterns in bacterial and fungal communities over different cultivation durations. The bacterial communities in the CK treatment, sixth-year, and tenth-year cultivated grasslands were relatively similar, while those in the second-year cultivated grasslands differed significantly from the CK treatment and later successional stages ( Figure 3d ). For the fungal community, the CK treatment and sixth-year cultivated grasslands displayed similarity, whereas the fungal communities in the second and tenth years exhibited significant differences compared to both the CK treatment and the sixth-year cultivated grasslands ( Figure 3e ).

Across different management durations, Acidobacteriota (23.82%), Proteobacteria (16.11%), Planctomycetota (12.06%), Actinobacteriota (10.06%), and Bacteroidota (9.91%) were the dominant bacterial phyla. The relative abundance of Acidobacteriota initially increased and then declined with extended management duration, while Proteobacteria showed a continuous decline over time. In contrast, Actinobacteriota and Planctomycetota increased with longer cultivation years. The relative abundance of Bacteroidota initially decreased but later increased with cultivation time ( Figure 3f ).

Regarding fungal communities, the dominant phyla were Ascomycota (56.93%), Mortierellomycota (14.54%), and Basidiomycota (9.31%). The relative abundance of Ascomycota initially decreased but increased with prolonged cultivation. Mortierellomycota exhibited an initial increase followed by a decrease, whereas Basidiomycota steadily increased with extended cultivation time ( Figure 3G ).

LEfSe analysis further highlighted the impact of cultivation duration on microbial communities. In the early stages, cultivated grasslands significantly reduced the relative abundance of various bacterial phyla compared to the CK treatment. However, long-term cultivation resulted in minimal differences between cultivated and CK treatments. For fungal communities, the initial establishment of cultivated grasslands had a lesser impact compared to bacterial communities. Over time, multi-year cultivation increased the relative abundance of various fungal phyla compared to the CK treatment ( Supplementary Figures S1 , S2 ).

Principal Component Analysis (PCA) was performed, and the results are shown in Table 2 . For each principal component with eigenvalues ≥1, a set of soil indicators was selected based on factor loadings with absolute values >0.5. Indicators with loadings >0.5 in different components were grouped based on low correlation with other indicators ( Figure 4a ). Finally, total nitrogen (TN), total phosphorus (TP), soil organic carbon (SOC), and alkaline protease (SALPT) were selected as the final indicators for soil quality assessment, and PCA was conducted on these indicators ( Table 2 ). The calculation of TN, TP, SOC, and SALPT shared common factor variances and weights ( Supplementary Table S1 ).

According to the Soil Quality Index (SQI) calculation method, the SQI of different cultivated grasslands and control sites was calculated. The contributions of various evaluation indicators in the Multidimensional Scaling (MDS) analysis were analyzed based on the SQI of different treatments. The study reveals that the SQI of cultivated grasslands shows an increasing trend with the growth of management time, with the SQI of the grassland cultivated in the second year being the lowest. However, as the management time increases, the SQI of the cultivated grasslands gradually increases and in the tenth year (0.69 ± 0.06), significantly higher than the CK treatment ( Figure 4b ). From Figure 4c , it can be observed that the soil index of the entire dataset has a good linear relationship with the soil index from the Multidimensional Scaling (MDS) analysis (R² = 0.957, P< 0.01). This indicates that the soil quality assessment indicators based on MDS can serve as a substitute for the entire dataset in evaluating the soil quality of different cultivated grasslands in the study area.

From the perspective of soil total nitrogen content and nitrogen-cycling related enzyme activities, cultivated grassland succession exhibited distinct patterns across different stages. In the early stages (2nd year), soil total nitrogen content, urease, and protease activities were significantly lower than in undegraded natural grasslands (P< 0.05). This decline is attributed to reduced vegetation cover and aboveground biomass, which limit nitrogen input and accelerate microbial nitrogen decomposition (Xu et al., 2025; Wu et al., 2024; Beniston et al., 2014). These effects are exacerbated by cultivation practices and soil erosion, further reducing soil nitrogen content (Ma Y. et al., 2024, von Lützow et al., 2006; Zhao Y. et al., 2024). However, as succession progresses, soil nitrogen content and enzyme activities (urease and protease) significantly recovered by the 6th and 10th years. This recovery is likely driven by increased nitrogen availability from plant residue decomposition, which enhances soil nitrogen content (De Rosa et al., 2024; Zhu et al., 2023). Additionally, improved soil microbial communities in cultivated grasslands may enhance nitrogen fixation (Xu R. et al., 2023; Ma R. et al., 2024). The increased enzyme activities indicate accelerated nitrogen cycling, benefiting plant nitrogen uptake and soil fertility (Feng et al., 2019; Xue et al., 2018).

Regarding total phosphorus, no significant changes in soil total phosphorus content were observed in the early stages compared to undegraded grasslands (P< 0.05). This is likely due to a balance between phosphorus input and consumption during the initial stages of Kentucky bluegrass cultivation (Sun X. et al., 2024). Despite the potential for phosphorus depletion due to cultivation practices and soil erosion, phosphorus levels were maintained through fertilization and soil weathering (Xu et al., 2024). Over time, the decomposition of plant residues continues to release mineral nutrients, increasing soil phosphorus content.

Soil organic carbon (SOC) and carbon stock showed significant reductions in the early stages of grassland succession. This change can be attributed to lower plant diversity, reduced plant input, and higher microbial decomposition rates of organic carbon (Samson et al., 2024). However, in the mid- and late stages of succession, as invasive species increased and plant diversity improved, underground carbon input also increased. This led to the accumulation of soil organic carbon, facilitated by root exudates and microbial activity (Wang A. et al., 2024; Lange et al., 2015). Additionally, changes in sucrase and basic xylanase activities indicate that, over time, the rate of organic matter decomposition in the soil accelerated, promoting soil fertility and microbial activity (Wang A. et al., 2020; Su et al., 2024).

In summary, the succession of cultivated grasslands significantly influences soil nutrients, carbon stock, and enzyme activities. Early succession is characterized by nutrient depletion and low enzyme activity, while longer recovery periods lead to increases in soil nitrogen, phosphorus, and carbon, alongside enhanced enzyme activity.

Regarding microbial diversity, soil microbial richness, evenness, and Shannon diversity were lowest in the early stages of cultivated grassland succession. However, with increased recovery time, fungal communities showed significant improvements in richness, evenness, and Shannon diversity, reaching levels comparable to those of undegraded grasslands. In contrast, bacterial communities did not exhibit significant differences compared to natural grasslands. These findings suggest that long-term cultivation of Kentucky bluegrass can restore microbial community diversity and stability, particularly for fungal communities (Sun H. et al., 2024, Li et al., 2024).

Regarding bacterial community changes, in late-stage cultivated grasslands, the relative abundance of specific bacterial taxa shifted significantly. The abundance of Acidobacteriota and Proteobacteria decreased, while Actinobacteriota increased. This shift aligns with the hypothesis that increased soil nitrogen levels favor fast-growing bacterial groups such as Actinobacteriota, while slower-growing groups like Acidobacteriota and Proteobacteria decline (Chen et al., 2024; Pan et al., 2024; Coonan et al., 2020; Zhang et al., 2019). Our results support this trend, with soil nitrogen content increasing over time, accompanied by a decrease in Acidobacteriota and Proteobacteria and an increase in Actinobacteriota (Kielak et al., 2016; Li S. et al., 2024). Additionally, LEfSe analysis revealed that late-stage cultivated grasslands decreased the relative abundance of RB41, Chthoniobacteraceae, and Candidatus Udaeobacter, while increasing AKYG587. RB41 is involved in nitrogen fixation, and Chthoniobacteraceae plays a role in soil polysaccharide decomposition (Zou et al., 2024; Yang et al., 2024). AKYG587 is linked to organic matter decomposition and soil microbial community maintenance (Xing et al., 2020; Arellano-García et al., 2024). These changes suggest that multi-year cultivation of Kentucky bluegrass enhances organic matter decomposition and bacterial community stability (Lin M. et al., 2023; Yusuf et al., 2021).

Regarding fungal diversity, Mortierellomycota initially increased but later decreased in abundance. This fungus plays a critical role in mineral phosphorus solubilization and nutrient availability (Liu et al., 2024; Baćmaga et al., 2024). The observed decline may result from a transition from sexual to asexual reproduction, which reduces phosphorus demand over time (Kruczyńska et al., 2023; Wang X. et al., 2024). Consistent with these findings, LEfSe analysis revealed that late-stage cultivated grasslands had lower relative abundances of Mortierella, Actinomucor, Chaetomiaceae, and Fusicolla compared to natural grasslands. Mortierella is involved in soil phosphorus cycling, while Fusicolla may form mycorrhizal relationships to enhance nutrient uptake (Ozimek and Hanaka, 2020; Shen et al., 2022).

In summary, the succession of cultivated grasslands promotes the diversity and stability of fungal and bacterial communities, with changes in community composition closely linked to soil nitrogen dynamics and organic matter decomposition.

Current soil quality evaluation methods, such as CASH (Comprehensive Assessment of Soil Health), HSHT (Haney Soil Health Test), and M-SQR (Muencheberg Soil Quality Rating), are widely used for assessing arable soil quality but have limitations when applied to grassland soils (Chang et al., 2022) (.Li et al., 2023; Zhao B. et al., 2024; Yu et al., 2018). Unlike croplands, the grasslands of the Qinghai-Tibet Plateau serve not only to meet livestock production demands but also to achieve ecological value. Therefore, this study assessed the soil quality of cultivated grasslands dominated by Kentucky bluegrass using a range of indicators that reflect ecosystem service functions, including soil bulk density, pH, carbon stock, soil organic carbon (SOC), total nitrogen (TN), total phosphorus (TP), nitrogen and carbon cycling enzymes, and microbial communities.

Results indicate that soil quality in cultivated grasslands improved with increasing recovery time, with late-stage grasslands showing significantly higher soil quality than natural grasslands. This suggests that multi-year cultivation substantially enhances soil quality compared to undegraded natural grasslands (climax communities). Analysis of the Soil Quality Index (SQI) revealed that total nitrogen content was the key factor influencing soil quality, contributing 33.3%–43.3% to the SQI ( Figure 4d ). During the early stages of succession, total phosphorus also contributed significantly to the SQI. These findings highlight the importance of nitrogen and phosphorus fertilization in improving soil quality in both natural and cultivated grasslands. Additionally, the proportion of SOC and TN in the SQI decreased with increasing management time, likely due to their gradual accumulation in the soil, which alleviates their limitations on Kentucky bluegrass growth.

This study utilized Nature-based Solutions (NbS) to investigate the effects of Kentucky bluegrass-dominated cultivated grasslands on soil quality, nutrient cycling, and microbial community dynamics. Cultivated grasslands, as an NbS, have demonstrated their ability to restore degraded ecosystems while providing multiple benefits, such as enhanced biodiversity and improved soil fertility. However, several limitations of this study highlight the need for further research to optimize the application of NbS in grassland restoration (Bardgett et al., 2021).

Firstly, this study employed a monoculture approach with Kentucky bluegrass. While this method demonstrated substantial improvements in soil quality over longer time scales, monoculture can lead to reduced plant community diversity and compromised stability. Future research should explore mixed sowing strategies that incorporate Kentucky bluegrass with other forage grasses, such as legumes, to enhance plant community diversity and stability while reducing the need for fertilizers.

Secondly, although the study initially explored the impact of Kentucky bluegrass on soil microbial community structure, revealing significant increases in fungal richness and stability, the specific mechanisms by which these microbial changes influence nutrient cycling and ecosystem functions remain unclear. Future research should delve deeper into the functional dynamics of soil microorganisms and the changes in soil metabolites to elucidate their direct and indirect impacts on soil quality.

Lastly, while the study identified soil total nitrogen and phosphorus as key indicators of soil quality, it did not validate the specific effects of fertilization on soil quality. Future research should further investigate the long-term impacts of different fertilization strategies on soil quality in grasslands to optimize management practices and enhance soil health.

In summary, although this study offers preliminary insights into the application of Kentucky bluegrass for grassland restoration, its limitations should not be overlooked. Future research should delve deeper into mixed sowing approaches, soil microbial functions, and the long-term impacts of fertilization strategies. This will help refine the management practices for using Kentucky bluegrass to restore degraded grasslands and enhance overall ecosystem health and sustainability.