Introduction

Alpine wetlands are fragile natural ecosystems on the Qinghai–Tibet Plateau; they play key roles in various ecological processes including water conservation, climate regulation, and carbon (C) storage. However, the alpine wetlands have deteriorated rapidly in recent years, especially in the Sanjiangyuan region (hosting the headwaters of the Yellow, Yangtze, and Lancang-Mekong rivers). The alpine wetland in this region have experienced widespread degradation, shrinking by 29% over the past decades, with the wetland ecosystems increasingly fragmented at an accelerated pace1. The degradation of alpine wetlands on the Qinghai–Tibet Plateau involve multiple drivers2,3, of which long-term overgrazing should be recognized as the main one4. The degradation of the alpine wetlands has deleterious effects on the ecological environment of the Qinghai–Tibet Plateau and thus implications for wetland management.

Biomass is an indicator of an ecosystem’s capacity to capture energy. Some researchers have suggested that aboveground biomass (AB) increases as degradation increases, and significant differences in AB have been observed in non-degraded and severely degraded sites5,6,7. However, Peng et al.8 reported that AB is only reduced when degradation is severe. The degradation of wetlands results in the loss of biodiversity, which poses a major threat to the sustainable development of agriculture and animal husbandry on the Qinghai–Tibet Plateau; wetland degradation also increases ecological vulnerability9. Previous studies have indicated that grazing-induced soil degradation reduces the dominance of dominant species, promotes the expansion of habitat conducive to the growth of less competitive species in the community, and alters species density and richness10,11. In addition, degradation often results in changes in plant functional groups, which can alter plant community structure. Peng et al.12 reported that the dominant plants shift from sedges to forbs in alpine wetlands as the severity of degradation increases. The native plant community gradually gives way to a secondary plant community dominated by non-native species as the severity of alpine wetlands degradation increases; this leads to the extinction of numerous native species and a decline in community diversity and stability, which ultimately results in a reduction in ecosystem services13,14. However, the “intermediate disturbance hypothesis” posits that plant diversity is highest at moderate levels of degradation15,16. The effects of degradation gradients on AB and plant community structure in alpine wetlands, as well as the underlying mechanisms, remain unclear.

Microbes are the predominant life forms in soil and play a key role in various soil processes such as nutrient cycling, organic matter decomposition, soil-borne plant pathogen suppression, soil structure formation, and plant growth17. Small changes in the diversity and structure of microbial communities can have a significant effect on organic matter decomposition, humus synthesis, and the transformation of soil nutrients18,19. Changes in the soil microenvironment can significantly affect microbial community composition and activity either directly or indirectly through changes in various soil factors such as soil organic C, pH, and nitrogen (N) availability20. For example, under the disturbance of grazing, animal trampling can result in soil compaction, altering various soil properties such as bulk density, aeration, redox conditions, and hydraulic conductivity21,22,23. The extra nutrients input through grazing, by altering the form of organic carbon input into the soil, regulates the structure and diversity of the microbial community24,25,26,27. The deposition of dung and urine by large animals may lead to a decrease in soil pH, an increase in soil N availability, and an acceleration of N cycling28,29, thereby influencing the structure of soil microbial communities23,30. Therefore, modifications in soil physicochemical properties during the degradation of alpine wetlands, especially caused by grazing, inevitably lead to shifts in the diversity and composition of the soil microbial community. Li et al.31 demonstrated that soil bacterial and fungal diversity is significantly higher in severely degraded soils than in less degraded soils, and the composition of fungal species is significantly altered in severely degraded wetlands compared with less degraded wetlands. Dong et al.32 found that the richness and diversity indices of microbial species were highest in lightly and moderately degraded alpine wetlands. Zhou et al.33 suggested that the diversity of soil microorganisms, bacteria, and fungi did not change significantly after degradation; however, they observed changes in the composition of the microbial community. Various mechanisms are thought to underlie changes in microbial diversity during the degradation of alpine wetlands. There is a consensus that the composition of the microbial community is altered, but the direction and magnitude of these changes remain uncertain. Moreover, these studies have demonstrated that soil microorganisms in alpine wetlands are sensitive to alterations in the soil microenvironment during degradation. This sensitivity results in an elevated utilization of C sources by microorganisms34, and this is accompanied by a notable decrease in soil microbial biomass C (MBC)35. Consequently, characterizing shifts in microbial activity in degraded wetlands is key for providing insights into global C source/sink dynamics36,37.

In this study, we investigated the ecological characteristics of alpine wetlands under varying degrees of degradation, including light, moderate, and severe degradation. Specifically, we evaluated the composition and diversity of plant species, the structure and diversity of soil bacterial and fungal communities, and various ecosystem indicators related to vegetation productivity and soil nutrient cycling. The aim of this study was to analyze correlations of environmental factors with the diversity of plant species, soil bacteria, and fungi, as well as clarify the multifaceted effects of degradation on alpine wetlands and the underlying environmental mechanisms driving these effects. The results of this study enhance our understanding of degradation and restoration processes in alpine wetlands on the Qinghai–Tibet Plateau.

Results

Variation in plant communities under different levels of wetland degradation

As shown in Fig. 1a, AB was highest in the non-degraded (WM) site (average of 487.0 ± 127.6 g m−2). In contrast, AB was lowest in the severely degraded (SD) site (average of 134.0 ± 26.6 g m−2). The AB of the WM site was significantly greater than that of the lightly degraded (LD), moderately degraded (MD), and SD sites. Members of the family Cyperaceae comprised 72.8% of the total AB of WM site. The percentage of Cyperaceae family members of AB decreased with the degree of degradation. Additionally, the percentage of Compositae family members was the lowest in the MD site. In the SD site, the percentage of Cyperaceae family members was low, and the percentage of other families such as Ephedraceae and Ranunculaceae was high (Fig. 1b). The importance value of sedge plants decreased with the severity of wetland degradation, indicating that their dominance gradually decreased as the severity of wetland degradation increased. This shift ultimately leads to a decrease or loss of wetland ecosystem functions, including maintaining water resources and supporting wildlife survival38,39. The Margalef index of the plant community in the MD site was significantly higher than that in other sites (Fig. 1c), while no significant differences were observed among the Shannon-Wiener diversity index of plant communities under different degrees of wetland degradation (Fig. 1d).

Fig. 1: Changes in aboveground biomass, plant composition, and diversity under different degradation levels.
Fig. 1: Changes in aboveground biomass, plant composition, and diversity under different degradation levels.
Full size image

Aboveground biomass (a), the percent of plant community families (b), the Margalef abundance index (c), and the Shannon-Wiener diversity index (d) under different degradation levels (mean ± SD). Different letters in the panel indicate significant differences among the treatments at a significance level of 0.05.

Variation in soil characteristics under different levels of wetland degradation

Soil water content (SWC) was significantly higher in different layers of WM soil than in degraded soils. Furthermore, SWC decreased consistently as the level of degradation increased (Fig. 2a). The pH of MD and SD soil was significantly higher than that of WM soil, and the pH of WM soil at a depth of 0–10 cm was significantly higher than that of LD soil at the same depth (Fig. 2b). The total organic C (TOC) content of WM soil in the three soil layers was higher than that in degraded soils, and the TOC content in WM soil was 128.7 ± 9.7 g kg−1, 92.9 ± 5.6 g kg−1, and 67.5 ± 3.9 g kg−1 in the 0–10 cm, 10–30 cm, and 30–50 cm soil layers, respectively. The TOC content of soil at various depths decreased with the level of degradation (Fig. 2c). The ammonium (NH4+) content was significantly higher in WM soil at a depth of 0–10 cm than in SD soil; no significant differences in the NH4+ content were observed between WM and SD soil in the other soil layers (Fig. 2d). Differences in soil nitrate (NO3-) content under different degradation levels were only observed in the 10–30 cm soil layer. The NO3− content was significantly higher in WM soil than in the other soils, and it was significantly lower in SD soil than in the other soils (Fig. 2e). The total N (TN) content of WM soil was higher than that of the degraded soils at all three soil depths. Additionally, soil TN decreased with the severity of degradation at all soil depths (Fig. 2f).

Fig. 2: Variation in soil characteristics and stoichiometry at different soil depths under different degradation levels.
Fig. 2: Variation in soil characteristics and stoichiometry at different soil depths under different degradation levels.
Full size image

Soil water content (a), soil pH (b), total organic carbon content (c), ammonium concentration (d), nitrate concentration (e), total nitrogen (f), dissolved organic carbon content (g), microbial biomass carbon content (h), microbial biomass nitrogen content (i), total phosphorus content (j), available phosphorus concentration (k), carbon to nitrogen ratio (l), nitrogen to phosphorus ratio (m), and carbon to phosphorus ratio (n) under different degradation levels (mean ± SD). Different letters in each panel indicate significant differences among the treatments at a significance level of 0.05.

The dissolved organic C (DOC) content was highest in LD soils at a 0–10 cm depth (127.6 ± 9.1 mg kg−1 on average), and it was significantly higher in LD soil than in WM and SD soils (Fig. 2g). However, no significant differences in the DOC content between WM, LD, and MD soils were observed at a 10–30 cm soil depth, but the DOC content was significantly higher in SD soil (average of 27.8 ± 4.6 mg kg−1) than in WM, LD, and MD soils. The DOC content was significantly lower in SD and MD soils than in WM and LD soils at a 30–50 cm depth (Fig. 2g). The MBC content of WM soil was significantly higher than that of degraded soils in all three soil layers, and there was no significant difference in MBC among degraded soils (Fig. 2h). Microbial biomass N (MBN) content was highest in WM soil at 0–10 cm and 10–30 cm depths (Fig. 2i). The total phosphorus (TP) and available phosphorus (AP) contents were highest in WM soil at a 0–10 cm soil depth, and they were significantly higher in WM soil than in degraded soil; the TP and AP contents decreased with the severity of degradation. However, the TP content in soil at a depth of 30–50 cm increased with the severity of degradation (Fig. 2j, k). The degradation of wetlands significantly affected soil stoichiometry. No significant differences in soil C to N ratios were observed between soil layers with varying degrees of degradation (Fig. 2l). The ratios of N to P and C to P in non-degraded soils were notably greater than those in degraded soils at various depths (Fig. 2m, n). Furthermore, the contribution of geographic distances to the differences in environmental conditions between sampling sites accounted for approximately 7.4% (Supplementary Fig. 1b). Thus, the differences in soil properties between sites were mainly due to wetland degradation.

Structure of bacterial and fungal communities at different soil depths under different stages of wetland degradation

Wetland degradation had a significant effect on bacterial α-diversity indices at 0–10 cm soil depth, showing a decrease followed by an increase (Fig. 3a–c). The microbial community’s α-diversity reached a relatively high level in the MD soil. For example, the Chao1 index of the bacterial community was significantly higher in MD soil (17377.3 ± 517.7) at a 0–10 cm depth than in LD soil (15855.4 ± 1249.8) (Fig. 3a). The same pattern also applied to the changes in the ACE index and Simpson index across different stages of wetland degradation (Fig. 3b, c). On the contrary, no significant differences in the Chao1 index and ACE index of the fungal community at 0–10 cm soil depth were observed under different levels of degradation. The ACE index and Chao1 index of fungal communities in the 10–30 cm and 30–50 cm soil layers demonstrated a gradual increase with the intensifying wetland degradation (Fig. 3d, e). However, no significant differences in the Simpson index of the fungal community in all three soil layers were observed under different levels of degradation (Fig. 3f). In addition, the correlation between geographical distance and the Bray–Curtis dissimilarities among microbial communities was not significant (p = 0.640, Supplementary Fig. 1a). The contribution of spatial heterogeneity to the differences in microbial communities accounted for approximately 6.0%, indicating that the differences in microbial communities were mainly due to wetland degradation (Supplementary Fig. 1a).

Fig. 3: α-diversity indices of the bacterial and fungal communities at different soil depths under different levels of degradation.
Fig. 3: α-diversity indices of the bacterial and fungal communities at different soil depths under different levels of degradation.
Full size image

Chao1 index, ACE index, and Simpson index of the bacterial community (a–c) and fungal community (d–f) at different soil depths under different levels of degradation. The letters marked the difference in the degrees of degradation of the same soil layer (p < 0.05).

The predominant bacterial classes in the soil, with a relative abundance exceeding 2.5%, included Gammaproteobacteria, Alphaproteobacteria, Bacteroidia, Blastocatellia, Vicinamibacteria, Verrucomicrobiae, Actinobacteria, and Thermoleophilia (Fig. 4a–c). The relative abundance of Gammaproteobacteria was highest in all soil layers, and the abundance of Gammaproteobacteria communities was significantly higher in WM soil than in LD soils in all three soil layers (Fig. 4a–c, Supplementary Fig. 2a–c). The relative abundance of Gammaproteobacteria increased with the severity of degradation. Alphaproteobacteria communities were the second most abundant across the various soil layers. Specifically, the relative abundance of Alphaproteobacteria in the 0–10 cm soil layer was significantly higher in WM soil than in SD soil (Fig. 4a, Supplementary Fig. 2a). Furthermore, the relative abundance of Verrucomimicrobiae communities peaked in lightly degraded soils, and no disparities were observed in other soil types such as LD, MD, and SD soils (Fig. 4a–c, Supplementary Fig. 2a–c). The dominant fungal communities were Agaricomycetes and Dothideomycetes, which comprised an average of 31% and 2.4% of the fungi in all soils, respectively (Supplementary Fig. 2d–f). Significant differences in the relative abundance of Dothideomycetes were only observed in the 10–30 cm soil layer, indicating that the relative abundance of Dothideomycetes was significantly higher in SD soil than in LD and WM soils (Supplementary Fig. 2e). The relative abundance of oligotrophs was higher in LD and MD soils than in SD soils at a soil depth of 0–10 cm (Fig. 4d). The relative abundance of oligotrophs initially increased and then decreased as degradation intensifies (Fig. 4d). The ratio of oligotrophs to copiotrophs (O/C) was significantly higher in the LD soil at different depths than in other soils (p < 0.05, Fig. 4e).

Fig. 4: Composition of bacterial communities and the relative abundance of bacterial groups classified according to different life strategies at different soil depths under varying degradation levels.
Fig. 4: Composition of bacterial communities and the relative abundance of bacterial groups classified according to different life strategies at different soil depths under varying degradation levels.
Full size image

Relative abundance of the main classes of bacteria (a–c), the relative abundance of copiotrophs (d), and oligotrophs/copiotrophs (O/C, e) of bacterial communities at different soil depths under different degradation levels (mean ± SD). Different letters in the panel indicate significant differences among the treatments at a significance level of 0.05. The classification of the oligotrophic and copiotrophic taxa is shown in Supplementary Table 4.

Relationships among soil, plants, and microbial factors at different soil depths under different stages of wetland degradation

Based on the Redundancy analysis (RDA), the eigenvalues for the first and second RDA axes in the 0–10 cm soil layer were 0.745 and 0.057, respectively. These two axes explained 80.2% of the variation in soil physicochemical properties and soil microorganisms (Fig. 5a). The first RDA axis was correlated with the extent of degradation, with the degree of degradation increasing from right to left. Significant differences were observed among various degrees of degradation, and this was closely associated with changes in SWC, MBC, and TOC. In the RDA of the 10–30 cm soil layer, the eigenvalues of the first and second RDA axes were 0.701 and 0.045, respectively. These two axes explained a total of 74.6% of the variation in soil physicochemical properties and soil microorganisms (Fig. 5b). A notable difference was observed between WM and degraded soils (LD, MD, and SD), and this was primarily associated with SWC, the concentration of AP and TOC, and the relative abundance of Gammaproteobacteria communities. No distinct separation was observed between MD and LD soils. In the RDA of soil layers at 30–50 cm, minimal variation was observed among soils under different degradation levels. However, clear separation was observed between WM soil and the degraded soils. This separation was mainly attributed to differences in SWC, the concentrations of TN and TOC, and the relative abundance of Gammaproteobacteria (Fig. 5c). Obviously, the sensitivity of bacterial community to wetland degradation intensity decreased as soil depth increased. Whereas, the fungal communities remained relatively stable under wetland degradation, as there was no clear separation among soils based on variation in fungal communities (Supplementary Fig. 3). When all data points were analyzed together, we found that among microbial-related factors, the Chao1 index of fungi and bacteria were significantly negatively correlated with most edaphic factors, including the C/P and N/P ratios, SWC, TOC, and TN, and only positively correlated with soil pH (Fig. 6). The environmental factors related to soil nutrition-related factors are negatively correlated with soil pH, while positively correlated with plant-related factors (i.e., AB) (Fig. 6). Among the plant-related factors, the Margalef index was positively correlated with the Chao1 index of the fungal community (Fig. 6).

Fig. 5: Relationships between soil properties and the main microbial classes of bacterial community at different soil depths under different degradation levels.
Fig. 5: Relationships between soil properties and the main microbial classes of bacterial community at different soil depths under different degradation levels.
Full size image

Redundancy analysis (RDA) of the multivariate relationships between soil properties and the main microbial classes of bacterial community at soil depths of 0–10 cm (a), 10-30 cm (b), and 30-50 cm (c) under different degradation levels.

Fig. 6: Mantel tests of the composition of copiotrophic and oligotrophic bacterial communities for all environmental factors and pairwise comparisons of environmental factors.
Fig. 6: Mantel tests of the composition of copiotrophic and oligotrophic bacterial communities for all environmental factors and pairwise comparisons of environmental factors.
Full size image

Edge width corresponds to the Mantel’s r statistic for the corresponding distance correlations, and edge color denotes the statistical significance based on 9999 permutations. The color gradient denotes Spearman’s correlation coefficients. SWC soil water content, pH soil pH, TOC total organic carbon content, DOC dissolved organic carbon content, MBC microbial biomass carbon content, TN total nitrogen content, NH4 ammonium concentration, NO3 nitrate concentration, MBN microbial biomass nitrogen content, TP total phosphorus content, AP available phosphorus content, C/N carbon to nitrogen ratio, N/P nitrogen to phosphorus ratio, C/P carbon to phosphorus ratio, O/C oligotrophs abundance to copiotrophs abundance, chao1_F Chao1 index of fungal community, chao1_B Chao1 index of bacterial community, Simpson_F Simpson index of fungal community, Simpson_B Simpson index of bacterial community, AB aboveground biomass, Margalef Margalef abundance index, Compositae relative abundance of Compositae.

The effects of environmental factors on copiotrophic and oligotrophic bacteria under wetland degradation

For edaphic factors, the composition of copiotrophic bacteria were significantly correlated with soil pH, which indicated that differences in copiotrophic bacterial community composition increased as the difference in soil pH between plots increased. In addition, the composition of oligotrophic bacteria was significantly correlated with SWC, TOC, TN, and soil stoichiometry (i.e., C/N, C/P, and N/P) according to the Mantel tests (Fig. 6). This indicates that the composition of oligotrophic bacteria was more sensitive to changes in edaphic factors during wetland degradation compared to copiotrophs. Furthermore, the variation in TOC among soils was closely associated with shifts in the composition of oligotrophic bacteria, indicating its negative effect on carbon sequestration under wetland degradation. The relative abundance of oligotrophs was closely related to changes in soil pH. When the pH was low, the relative abundance of oligotrophs decreased rapidly as the pH increases. However, when the soil pH exceeds approximately 7.0, the relative abundance of oligotrophs decreases at a much slower rate with increasing pH (R2 = 0.132, p < 0.01, Fig. 7a, Supplementary Tables 1, 2). On the contrary, the relative abundance of oligotrophs exhibited a unimodal response to changes in soil C/P (Supplementary Tables 1, 2). Specifically, when the C/P ratio was low, the relative abundance of oligotrophs increased with an increase in C/P. Conversely, when the C/P ratio was high, the relative abundance of oligotrophs decreased with an increase in C/P (R2 = 0.268, p < 0.01, Fig. 7b, Supplementary Tables 1, 2). The microbial Chao1 diversity index demonstrates a quadratic polynomial relationship with the O/C ratio, showing a trend of increasing and then decreasing as the O/C ratio gradually increases (R2 = 0.225, p = 0.015, Fig. 7c, Supplementary Tables 1, 2).

Fig. 7: Relationships between the relative abundance of oligotrophic bacteria and soil pH and C/P, and relationship between bacterial O/C and the Chao1 index under wetland degradation.
Fig. 7: Relationships between the relative abundance of oligotrophic bacteria and soil pH and C/P, and relationship between bacterial O/C and the Chao1 index under wetland degradation.
Full size image

The relative abundance of oligotrophic bacteria is plotted against soil pH (a) and soil C/P (b), and the bacterial O/C is plotted against the Chao1 index (c). The desirable regression curve was selected based on a smaller Akaike’s information criterion (AIC) and a relatively larger R2 simultaneously. The lower the AIC and the higher the R2, the better the model. The analysis results and the equations of the fitted curves are shown in Supplementary Table 1 and Supplementary Table 2, respectively.

Discussion

Biomass is a crucial metric for evaluating the productivity and growth of grassland plants and for assessing the overall health of ecosystems40. The results of this study revealed a significant decrease in AB as wetland degradation intensifies (Fig. 1a). Specifically, the AB of WM soil was significantly higher than that of SD soil, which is consistent with the results of previous studies of alpine wetlands41. This difference might be attributed to the decrease in soil fertility and water-holding capacity as degradation intensifies (Figs. 2, 6). This shift would have negative effects on ecosystem services provided by alpine wetlands, including stabilizing ecosystem balance, providing wildlife habitat, and maintaining water resources38,39. In addition, the plant species richness index of the MD site was greater than that of the WM site (Fig. c, d). This finding is consistent with the “intermediate disturbance hypothesis,” which posits that plant species diversity is highest under conditions of moderate disturbance42,43. In the LD site, the dominant plant communities primarily comprised sedge plants that propagate asexually. Forage grasses, mainly Cyperaceae, are more susceptible to fluctuations in water and nutrients compared with miscellaneous grasses, which potentially alters the competitive dynamics among vegetation species in degraded wetlands44. The complementary effect of the ecological niche suggests that a decrease in the abundance of a functional group in a community provides space and resources for other functional groups, which facilitates their growth45. Consequently, the consumption of large numbers of sedge plants by grazing livestock allows miscellaneous grasses, such as Compositae species, to invade, and this explains the high diversity and abundance of Compositae in moderately degraded alpine wetlands. Further degradation of wetlands leads to decreases in soil moisture and nutrients, which might result in the elimination of most plant species46. When degradation is severe, only plants adapted to growth in arid and barren environments are capable of surviving; this explains the decline in both the Shannon-Wiener diversity index and the Margalef abundance index of severely degraded wetlands.

Alpine wetland degradation involves the deterioration of vegetation and soil and disrupts vegetation–soil relationships. Soil organic C, N, and P are key indicators of soil fertility47. TOC plays a key role in soil structure formation and directly affects soil fertility, erosion resistance, and bulk density48. The results of our study revealed a significant decrease in the concentration of soil TOC, TN, and AP as the severity of wetland degradation increased (Fig. 2). These findings are consistent with the results of previous studies31,49. The decrease in the content of soil nutrients can be attributed to several factors. The decrease in soil TOC can be attributed to the sparse vegetation resulting from degradation, which leads to a significant reduction in AB and the accumulation of organic matter. Consequently, the soil’s capacity for nutrient enrichment weakens, which results in a decrease in the content of organic C. In addition, the decrease in vegetation coverage caused by wetland degradation can lead to an increase in the exposure of the soil surface and the vulnerability of the soil to erosion and weathering, which would ultimately result in the loss of soil nutrients50,51. Furthermore, the decrease in SWC inhibits microbial activity and microbial biomass, which results in reduced mineralization; the conversion of N and P under such conditions is insufficient for meeting the growth requirements of plants52,53, and this ultimately limits plant growth and slows the return of organic matter to the soil. The decline in microbial activity and abundance decreases the N-fixing ability of the soil, which leads to a substantial loss of N and thus soil TN54. In contrast to C and N, soil P is primarily derived from the parent material, which makes degradation-induced soil loss a key factor underlying the reduction in the soil AP content55. Additionally, soil stoichiometry is often used as an indicator of the nutrient cycling status within soil35. A C:N:P ratio of approximately 60:7:1 was reported in a global assessment of soil stoichiometry conducted by Cleveland and Liptzin56. The results of this study indicated that the C:N:P ratios in alpine wetlands under natural conditions were lower than global averages, suggesting that nutrients are highly limiting (especially P) in alpine wetland soils on the Qinghai–Tibet Plateau57. Furthermore, the limitation of N and P in soil microbial communities is expected to increase as the severity of wetland degradation increases, which is consistent with the results of previous studies35. The degradation of alpine wetlands slowed the replenishment of soil nutrients, disrupted the equilibrium between soil C, N, and P reserves, and increased the vulnerability of alpine wetland ecosystems to degradation55.

Soil pH has a significant effect on soil fertility, the decomposition of organic matter, N mineralization, and plant growth, and this is attributed to its effect on the conversion of soil nutrients and the functions of soil microorganisms58,59. In our study, the pH was kept at a relatively low level under lightly degradation compared to under moderately and severely degradation, which could be related to the phosphorus starvation responses. The ecosystem begins to deteriorate as degradation intensifies; under these conditions, soil phosphorus becomes an increasingly important factor affecting plant characteristics60. During the phosphorus starvation response, plants initially detect the absence of phosphorus, then activate signal cascades to initiate and coordinate adaptive responses, ultimately leading to the release of organic acids from the plant roots into the soil, which intensifies soil acidification61. As the intensity of degradation continues to increase, the plants gradually die, the secretion of root system decreases, and the soil pH gradually increases. Moreover, soil phosphorus exhibited high sensitivity to both acidic and alkaline environments62. In the case of higher soil alkalinity, soil phosphorus will gradually transform into a form that is difficult for plants to absorb and utilize. This conversion not only hinders the uptake and utilization of phosphorus by plants but may eventually lead to more severe phosphorus limitation63. Therefore, as degradation intensifies, an increase in pH could make it more difficult for the plant to absorb phosphorus, exacerbate plant death, and ultimately create a vicious cycle.

The diversity and composition of soil microorganisms are closely linked to their resilience and ability to adapt to changes in environmental conditions64. Highly diverse soil microbial communities are more stable and resilient in the face of environmental disturbances compared with less diverse soil microbial communities65,66. Microorganisms can adapt to the soil environment by reversibly adjusting their physiology to maximize resource utilization67. The “intermediate disturbance hypothesis” suggests that species richness at the local scale peaks at intermediate values of disturbance frequency or spatial extents of disturbance68. In our study, the α-diversity indices for bacteria first decreased and then increased as the severity of degradation increased, reaching their peak under moderately wetland degradation, which provided support for the “intermediate disturbance hypothesis.” This could be closely link to the altered bacterial community composition by favoring oligotrophic dominance under lightly wetland degradation and the nonlinear relationship between the relative abundance of oligotrophs and soil pH and C/P. Elemental ratios (i.e., resource stoichiometry) are reported to affect microbial community composition through r- or K- strategies69. Generally, oligotrophs exhibit a high affinity for substrates and may outcompete copiotrophs when available C and P are low70,71,72. Consequently, in the early stages of wetland degradation, the decrease in soil P impedes the proliferation of copiotrophs, thereby creating conducive environments for the growth of oligotrophs. However, as degradation intensifies and substantial soil P depletion occurs, the activity of oligotrophs is ultimately suppressed (Fig. 7b). Additionally, we found that the relative abundance of certain keystone taxa, such as Gammaproteobacteria, Alphaproteobacteria, and Bacteroidia, increased under water-deficit and nutrient-deficient conditions induced by light degradation. The relative abundance of oligotrophic, slow-growing K-strategists increased, and the abundance of copiotrophic, fast-growing r-strategists decreased. This is consistent with the results of previous studies showing that oligotrophic organisms may outcompete copiotrophic organisms under sustained stress when resources are scarce73,74. Copiotrophic bacteria exhibit elevated rates of growth, increased population size variability, and reduced substrate affinities compared with oligotrophic bacteria. In environments where microorganisms are exposed to sustained environmental stress, particularly stress stemming from limited resources, oligotrophs are expected to have a competitive advantage over copiotrophs70. Although a slight decrease in nutrients may lead to the enhancement of carbon utilization by slow-growing microorganisms (i.e., r-strategists). However, K-strategist microorganisms also have higher extinction rates when the environment experiences a significant deficit. This phenomenon explains the decrease in the relative abundance of oligotrophs under severe degradation. Significant nonlinear correlations of soil pH with oligotrophic communities were observed. High pH has the potential to enhance C use efficiency, which results in the breakdown of resistant organic compounds and facilitates the proliferation of oligotrophic microorganisms75. High pH can also impede microbial activity76. This could explain the increase in the relative abundance of oligotrophs when soil pH decreases significantly under light degradation, and the subsequent decrease in the relative abundance of oligotrophs when soil pH increases under moderate degradation.

Methods

Study site

The study area is located in Maqin County (34°23'27.6“N, 100°15'28.8“E) of Guoluo Tibetan Autonomous Prefecture, Qinghai Province, on the Qinghai–Tibet Plateau at an altitude of 3780 m (Fig. 8), which belongs to the Sanjiangyuan region. This region experiences a continental plateau climate with two seasons: a cold season and a warm season. The mean annual temperature is −0.4 °C; the lowest temperature is observed in January (−12.4 °C), and the highest temperature is observed in July (9.8 °C). The average annual precipitation is 508.5 mm, with 60% of the annual precipitation falling from June to August. The soil type within the study area was classified as alpine meadow soil according to the Chinese Soil Taxonomy, which corresponds to the cambisols in the World Reference Base (WRB) for soil resources77. The area is characterized by low-lying vegetation and fragile ecosystems that are susceptible to disturbances. According to the comprehensive grassland survey conducted by Maqin County Bureau of Agriculture and Animal Husbandry in 1980, the study area was typically undegraded wetland. The vegetation distribution across the study site was relatively uniform with little activity by local herdsmen, and Kobresia was identified as the dominant species78. However, over the past decades, this region has undergone varying degrees of disturbance, with overgrazing by domesticated livestock being the most significant issue (yak is the main livestock in the study area)4,79. Due to varying degrees of grazing, this region exhibits a gradient of degradation statuses, characterized by different stages of vegetation succession43.

Fig. 8: Sample locations and photos in Maqin County on the Tibetan Plateau.
Fig. 8: Sample locations and photos in Maqin County on the Tibetan Plateau.
Full size image

The land use data utilized was extracted from the 1 km grid dataset of land use in China in 2015, which is openly accessible from the National Earth System Science Data Center (https://www.geodata.cn).

Experimental design and sampling

Field sampling was carried out in July 2021. The severity of degradation was determined following the national standard “The Degradation Classification Criteria for Grazing Classification Standard of Natural Grassland Degradation, Desertification, and Salt Stains” (GB19377-2003). Vegetation coverage of wetlands was the primary consideration80,81,82. We selected four sampling sites that varied in their level of degradation. These include a non-degraded wetland (WM) with high vegetation coverage between 90 and 100%, a lightly degraded wetland (LD) with medium to high vegetation coverage between 60 to 89%, a moderately degraded wetland (MD) with medium vegetation coverage between 30 and 59%, and a severely degraded wetland (SD) with low vegetation coverage below 30%83,84,85 (Fig. 8, Supplementary Table 3). The selection criteria were as follows: 1) all sampling sites had similar environmental conditions before degradation occurred; 2) the non-degraded site (i.e., WM) has remained undisturbed by human activities for at least the past 20 years. Based on the county annuals, the WM site was unaffected by grazing as it remained distant from the local herdsmen’s habitation areas; 3) the sampling sites for degradation (i.e., LD, MD, and SD) exhibited a long history of grazing. The degradation intensity varied primarily because certain sites were impacted by grazing more heavily than others; 4) the sampling sites exhibit similarities in elevation, terrain, and microclimate conditions, while avoiding both the driest and the most humid locations, as well as areas with dung residue. At each site, three plots were randomly selected. Each plot comprised three quadrats measuring 0.5 m × 0.5 m. The measurements taken for each quadrat included the number of species, species coverage, and total coverage. After the vegetation survey, the aboveground parts of the plants in the sampling area were cut, placed into envelopes, and brought to the laboratory for determination of AB. Soil samples were collected using a soil drill, with three samples taken along the diagonal in each quadrat and divided into three layers (0–10 cm, 10–30 cm, and 30–50 cm). Samples within the same layer were thoroughly mixed until homogenous. The collected soil samples were separated into two parts, with one part designated for the examination of soil physical and chemical characteristics in a laboratory setting, and the other preserved at a temperature of −80 °C for high-throughput microbial sequencing.

Measurements of soil properties

The content of TOC and TN was determined using an elemental analyzer (EA1108, Carlo Erba, Turin, Italy). The content of soil TP was determined using alkali fusion molybdenum-antimony resistance spectrophotometry. Soil AP content was determined using the molybdenum antimony resistance colorimetric method via 0.5 ml·L−1 NaHCO3 extraction. The NO3− concentration, NH4+ concentration, and DOC concentration were determined using a continuous flow analyzer (Skalar San + +, Skalar Co., The Netherlands). The pretreatment for soil DOC, NH4+, and NO3− measurement mainly followed the method outlined by Bolan et al.86. The SWC was determined using the drying method. The MBC and MBN of soil were determined using chloroform fumigation and the steam extraction method87.

DNA extraction, PCR amplification, and sequencing

The FastDNA® SPIN Kit for Soil (MOBIO Laboratories, USA) was used to extract microbial DNA from 0.5 g of frozen (−80 °C) soil samples. An ultraviolet spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) was used to measure the concentration and purity of the extracted DNA samples. All purified samples were stored at −80 °C until further analysis. The V4-V5 fragments of bacterial 16 S rRNA were amplified with the primers 515 F (5ʹ-GTGCCAGCMGCCGCGGTAA-3ʹ) and 907 R (5ʹ-CCGTCAATTCCTTTGAGTTT-3ʹ). The ITS-1 fragment of the fungus was amplified using the primers ITS1F (5ʹ-CTTGGTCATTTAGAGGAAGTAA-3ʹ) and ITS2R (5ʹ- GCTGCGTTCTTCATCGATGC-3ʹ). The thermal cycling conditions were as follows: initial denaturation at 95 °C for 5 min; 40 cycles of denaturation at 98 °C for 10 s, annealing at 50 °C for 30 s, and extension at 72 °C for 30 s; and a final extension at 72 °C for 5 min. The PCR products were preserved at 4 °C. DNA extraction quality was evaluated using 2% agarose gel electrophoresis and purified using the AxyPrep DNA gel extraction kit (Oxygen Biosciences, Union City, CA, USA). Finally, sequencing was performed on the Illumina MiSeq PE300 platform (Illumina, San Diego, USA).

Bioinformatics analysis

QIIME 2 software was used to conduct bioinformatics analysis88. Priming, mass filtering, noise reduction, splicing, and chimeric sequence removal were performed using the DADA2 method89. After all libraries were denoised, the ASV/OTU sequences and OTU tables were merged, and singleton OTUs were removed. QIIME2 online analytical tools were used to analyze α-diversity and β-diversity based on Illumina sequencing data and identify microbial species.

Statistical analyses

All data were log-transformed to meet the assumptions of parametric tests prior to analysis. A one-way analysis of variance (ANOVA) was used to assess variation in AB, edaphic factors and bacterial and fungal α-diversity under different wetland degradation levels. Subsequently, multiple comparisons were conducted using the least significant difference method. Spearman correlation analyses were used to investigate the relationships of soil and plant characteristics with bacterial and fungal richness and diversity. RDA was used to explore the relationships between edaphic factors, plant characteristics, and bacterial and fungal communities. The classification of the identified phyla into oligotrophic and copiotrophic taxa is based on published researches (Supplementary Table 4). The relationships of environmental variables with the composition of oligotrophic and copiotrophic bacteria were examined using Mantel tests. The univariate regression analyses were performed to investigate the relationship between the relative abundance of oligotrophs and pH and C/P as well as the relationship between O/C ratio and Chao1 index of bacteria (Chao1_B). The desirable regression curve was selected based on a smaller Akaike’s information criterion (AIC)90 and a relatively larger R2 simultaneously. The lower the AIC and the higher the R2, the better the model. The Bray–Curtis dissimilarities were estimated based on the Bray–Curtis distance between bacterial communities, which was computed using OTU tables. The environmental distance between each site was assessed by calculating the Euclidean distance using the matrices of standardized environmental variables91. The “vegan” package was used to compute Bray–Curtis distances and Euclidean distances92. All analyses were conducted using R version 3.6.293.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.