Context-Dependent Contributions of Shrub Species Richness to Ecosystem Stability Across Tree Diversity Gradients: Insights from Microbial Mediators

As climate change intensifies, forest ecosystems face increasing destabilization and biodiversity loss. While the role of tree species richness in maintaining ecosystem stability is well-established, the influence of shrub diversity—particularly under varying tree species richness—remains insufficiently understood. Here, we investigate the intricate interplay between shrub species richness, soil microbial communities, and environmental factors in driving ecosystem stability, defined as the temporal stability of tree community productivity. Using a large-scale, controlled experimental platform (BEF-China), we examined these dynamics across a gradient of tree species richness: monocultures, two-species mixtures, and four-species mixtures. Our findings reveal that shrub species richness exerts context-dependent effects on ecosystem stability, strengthening stability at higher levels of tree species richness. These effects are mediated primarily by shifts in microbial diversity and community-aggregated genomic traits. In monocultures, stability was predominantly governed by abiotic factors such as soil pH and slope, with minimal contributions from shrub species richness. In contrast, in two-species mixtures, increased shrub species richness significantly enhanced stability by promoting bacterial diversity and restructuring fungal communities. In four-species mixtures, bacterial and fungal genomic traits differentially modulated stability, highlighting the pivotal yet distinct roles of microbial communities in mediating biodiversity-stability relationships. Additionally, we identified key microbial taxa whose contributions to stability varied with tree species richness, further emphasizing the complexity of these interactions. Together, our results underscore the dynamic and context-specific roles of shrub diversity and microbial mediators in shaping ecosystem stability, providing novel insights into the mechanisms underpinning forest resilience in an era of rapid environmental change.

Ecosystem stability - the capacity to maintain structure and function amid disturbance - is a cornerstone of ecological resilience and sustainability (Tilman et al., 2014). As anthropogenic pressures such as climate change and land-use transformation intensify, elucidating the mechanisms that underpin ecosystem stability has become a critical focus in ecolody (Loreau et al., 2001; Hughes et al., 2003; Oliver et al., 2015; Isbell et al., 2017; Hong et al., 2022). Accumulating evidence from grassland and forests highlights biodiversity as a key stabilizing force (Hautier et al., 2014; Chen et al., 2021; Wu et al., 2023), with tree species richness particularly critical in forests (Schnabel et al., 2021). Tree diversity can stabilize productivity via mechanisms such as resource partitioning, functional complementarity, and enhanced resistance to stress (Jucker et al., 2014; Liang et al., 2016; Del Río et al., 2017; Schnabel et al., 2019; Urgoiti et al., 2022; Meng et al., 2023; Qiao et al., 2023; Yan et al., 2024; Chen et al., 2025a). In contrast, the contributions of forest understory components - especially shrubs -remain underexplored. Shrubs regulate key ecosystem processes by contributing chemically distinct litter, improving soil fertility, enhancing water retention, and stabilizing soils (Cornwell & Ackerly, 2010; Wardle et al., 2012). Some, like nitrogen-fixing Fabaceae, further enrich soil nutrient pools (Temperton et al., 2007; Fornara & Tilman, 2008). Shrubs also modify microclimatic conditions, thereby shaping understorey dynamics (Deng et al., 2023). Despite this, how shrub diversity interacts with tree richness to influence temporal ecosystem stability is poorly resolved. Insights from the BEF-China experiment suggest shrubs can modulate tree diversity effects on forest function (Huang et al., 2018; Tao et al., 2023). On one hand, shrubs may compete with trees for light, water, and nutrients, potentially suppressing productivity (Balandier et al., 2022; Lecomte et al., 2022). On the other hand, increasing shrub richness may promote niche complementarity and reduce interspecific competition, enhancing community-level stability (Loreau et al., 2001; Feng et al., 2022; Chen et al., 2025b). However, while such dynamics have been examined primarily in terms of productivity, their implications for temporal stability—a crucial yet underexplored dimension of ecosystem resilience—remain poorly understood. Moreover, soil microbial communities—particularly bacteria and fungi—play pivotal roles in mediating plant–ecosystem feedbacks through nutrient cycling, decomposition, and disease suppression (Van der Heijden et al., 2008; Wagg et al., 2014; Banerjee et al., 2023). Plant diversity influences microbial structure via changes in organic matter, pH, and nutrient availability (Wardle et al., 2004; Bardgett et al., 2014; Delgado-Baquerizo et al., 2016; Crowther et al., 2019; Tedersoo et al., 2020). Microbial diversity and functional composition stabilize ecosystems by buffering against environmental fluctuations and promoting resilience (Eisenhauer et al., 2012; Rillig et al., 2015; Hong et al., 2022; Baldrian et al., 2023; Wu et al., 2023). Recent work has shown that microbial diversity and genomic traits (e.g., community-weighted mean GC content) can act as integrative indicators of microbial stability and function (Torsvik et al., 2002; Dick et al., 2009). Here, using data from the BEF-China platform, we examine how shrub species richness affects temporal stability of tree productivity across gradients of tree diversity. We hypothesize that shrub richness will enhance stability more strongly in higher-diversity tree communities, where microbial community traits are expected to mediate this relationship. At low tree richness, we expect environmental factors such as pH and slope to dominate. Our findings aim to inform biodiversity-based forest management strategies that enhance resilience in a changing world.

and methods. Experimental design and soil sampling. This study utilized data from the BEF-China platform (www.bef-china.com) in Xingangshan, Jiangxi Province, China (29°08′N–29°11′N, 117°90′E–117°93′E). The BEF-China experiment, established at two sites (A in 2009 and B in 2010), explores the effect of tree species richness on ecosystem functions (Bruelheide et al., 2014; Huang et al., 2018). For this study, 16, 8, and 4 plots were selected with tree species richness levels of 1, 2, and 4, respectively, from both sites A and B. These plots were nested within four different shrub species richness treatments (SR0, SR2, SR4, SR8). Four individual trees per species were randomly selected within each plot, resulting in 4, 8, and 16 sampled individuals for tree species richness levels of 1, 2, and 4, respectively (Figure 1a, Table S1). Soil samples were collected in October 2018. For each tree, four soil cores (0–10 cm depth) were taken from the upper half of the canopy projection area in cardinal directions, combined into a composite sample. A total of 192 samples were collected per site, with each plot contributing 64 composite samples. Samples were sieved (2-mm mesh) and stored for analysis of various soil properties. Environmental factors and microbial biomass A Digital Elevation Model was used to estimate the mean plot slope, altitude, aspect, and the component of north-south (NS) and east-west (EW) slope aspect, as described in the BEF-China data portal (Bruelheide et al., 2013a, b). Soil moisture content (SM) was determined by drying the soil samples at 105 °C for 48 hours. Soil pH was measured in a 1:2.5 soil-to-water suspension (w/v) using a Delta 320 pH meter (Mettler-Toledo Instruments Co., Shanghai, China). Soil organic carbon (C) and nitrogen (N) content were analyzed by direct combustion of air-dried soils using a CHNOS Elemental Analyzer (vario EL III; Elementar Analysensysteme GmbH, Germany). Soil available phosphorus (AP) was quantified using the molybdenum blue method, and a UV-visible spectrophotometer (UV-2550; Shimadzu, Kyoto, Japan). Nitrate (NO₃⁻-N) and ammonium (NH₄⁺-N) were extracted with 2 M KCl and quantified with a continuous flow analyzer (SAN++, Skalar, Breda, Netherlands). Soil microbial biomass carbon (MBC) and microbial biomass nitrogen (MBN) were measured using the chloroform fumigation-extraction method (Brookes et al., 1985). MBC and MBN were calculated using standard calibration coefficients (Vance et al., 1987; Brookes et al., 1985). Tree stand volume and community stability Tree stand volume data were obtained from a previous study (Bongers et al., 2021), with individual tree volume calculated using the formula: \(\text{Proxy}\ \text{Volume}=H\ \)× π (BR) 2 where H is the tree height and BR is the basal radius at ground level. Stand-level tree volume was computed by summing the volumes of surviving trees in the central 36 planting positions of each plot (Huang et al. 2018). Community stability was quantified as the inverse of the coefficient of variation (CV) of tree stand volume (Isbell et al., 2015; Craven et al., 2018) over the observation period (2015-2019). \papertype Original Article Amplicon sequencing and Bioinformatics DNA was extracted from 0.25 g soil using the PowerSoil® DNA Isolation Kit (MO BIO Laboratories, USA). Bacterial 16S rRNA genes (V3-V4) were amplified with primers 338F and 806R (Caporaso et al., 2011), and fungal ITS2 regions were amplified with primers ITS3F and ITS4R (Toju et al., 2012). PCR products were sequenced on an Illumina MiSeq platform (PE300) in Shanghai, China. Raw sequences were processed in QIIME2 (Bolyen et al., 2019), version 2023-05. Quality control, including trimming, denoising, merging, and chimera detection, was performed using the qiime dada2 denoise-paired plugin, as implemented in DADA2 (Callahan et al., 2016). Alpha and Beta diversity analyses were conducted on the rarefied ASV abundance tables using the core-metrics-phylogenetic pipeline in the q2-diversity plugin. Taxonomic classification of bacterial ASVs was performed using the qiime feature-classifier classify-sklearn plugin, utilizing the pre-trained Naïve Bayes Greengenes classifier for the V3-V4 region of the 16S rRNA gene. Fungal ASVs were classified using the UNITE database (https://unite.ut.ee). ASVs that were absent from at least five samples or had a total read count below ten across all samples were excluded to reduce noise in the dataset. Relationship between environmental factors and microbial communities Stability was modeled as a function of soil pH, moisture, NO₃⁻-N, NH₄⁺-N, microbial biomass C and N (MBC, MBN), MBC:MBN ratio, total C, available P, C:N, N:P, altitude, slope, and slope aspect (EW and NS components). Linear models were implemented with the lm() function in R (R Core Team, 2021), and results visualized via forest plots showing coefficients, 95% confidence intervals, and p-values. Mantel tests (via mantel_test() in the linkET package; Huang, 2021) assessed correlations between environmental variables and microbial diversity indices across tree species richness levels. Alpha diversity included Richness, Shannon, Simpson, Evenness, and Abundance; beta diversity was based on unweighted UniFrac distances. Genomic traits and their relationship with microbial community stability GC content and base count were calculated for each ASV using its representative DNA sequences. Community-weighted means for GC content (CWM-GC) and base count (CWM-BC) were then computed by weighting each ASV’s trait value by its relative abundance: \begin{equation} CWM-Trait=\sum_{i=1}^{n}\left(\text{Relative\ Abundance}_{i}\ \times\ \text{Trait\ Value}_{i}\right)\nonumber \\ \end{equation} where i denotes individual ASVs, and trait values refer to GC content or base count. Linear regressions were used to examine the relationships between CWM-GC, CWM-BC, and microbial community stability. All analyses were performed in R using the lm() (R Core Team, 2021) and cor() (Makowski et al., 2020) functions. Phylogenetic analysis of hub taxa Hub taxa were identified via co-occurrence networks constructed at different tree species richness levels (TR1, TR2, TR4) using the cor_Big_micro2 function from the ggClusterNet package (Wen et al., 2022). The top 500 most abundant ASVs were included, and hub scores were used to rank and select the top 40 hub taxa. To evaluate their contribution to community stability, Random Forest (RF) models were built for each tree richness level using the randomForest package (Breiman, 2001), with ASVs as predictors and stability as the response. Variable importance was measured by %IncMSE, and hub taxa with non-zero values were visualized using ggplot2 (Wickham, 2016). Shared hub taxa across richness levels were displayed with an UpSet plot (UpSetR; Conway et al., 2017). Phylogenetic relationships were explored by aligning hub ASV sequences via the MUSCLE algorithm (EMBL-EBI). Pairwise distances were calculated using the dist.dna function (raw model), and a neighbor-joining tree was constructed with the njfunction in the ape package (Paradis & Schliep, 2019). The tree was rooted using an archaeal ASV (Feature ID: a3118a46bbe33a2436f7549fa6f63b74) from Tao et al. (2023) as an outgroup. Relative abundance of hub taxa across tree richness levels was visualized with log-transformed heatmaps [log10(Abundance+1)], grouped by tree and shrub richness. Structural equation modeling Structural Equation Modeling (SEM) was used to assess the relationships among shrub species richness, environmental variables, microbial traits, and ecosystem stability across three tree species richness levels (TR1, TR2, TR4). At each level, linear regression models were first developed to examine the effects of shrub richness and key environmental factors (e.g., pH, slope, altitude) on microbial diversity and genomic traits (e.g., bacterial and fungal CWM-GC). These models were then integrated into a unified SEM framework using the psem function from the piecewiseSEMpackage (Lefcheck, 2016). This approach allowed for the evaluation of both direct and indirect effects. Model fit was assessed through path coefficients, p-values, and overall model summaries, with significant paths (p < 0.05) indicating meaningful associations. Relationships among community stability, shrub species richness, soil physicochemical, and environmental factors We investigated the impact of shrub species richness on community stability across three levels of tree species richness (TR1, TR2, and TR4) (Figure 1a; Table S2). Overall, community stability increased with shrub richness (SR0, SR2, SR4), but declined at the highest level (S8) (Figure 1b). The positive relationship between shrub richness and stability strengthened with tree species richness. At TR1, no significant correlation was found between shrub richness and stability (R² = 0.012, p = 0.22), although stability increased with SR2 to SR4 (Figure 1b). At TR2, the correlation became stronger (R² = 0.041, p = 0.021), with stability increasing across a broader range of shrub richness levels (SR0–SR4 and SR2–SR8) (Figure 1b). At TR4, the relationship was even more pronounced (R² = 0.15, p = 8.5 × 10⁻⁶), with significant increases in stability across most shrub richness combinations (SR0–SR2, SR0–SR4, SR0–SR8, SR2–SR4, and SR2–SR8) (Figure 1b). To investigate the role of soil physicochemical properties and topographic factors in community stability, we applied linear regression models (Figure 1c; Table S2). At TR1, TC, EW, and altitude negatively affected stability (p < 0.05), while slope and the C:N ratio had positive associations (p < 0.05). At TR2, NO₃⁻-N positively correlated with stability (p < 0.05), while TC and altitude negatively affected it (p < 0.05). At TR4, soil pH and NO₃⁻-N were negatively correlated with stability (p < 0.05), whereas EW and NS had positive relationships (p < 0.05). Trilateral relationships among stability, microbial community diversity and composition, and environmental factors We explored the relationships between microbial community diversity, composition, and ecosystem stability across different tree and shrub species richness levels (Figure 2a–d; Figures S1-S3). At TR1, bacterial alpha diversity (Simpson index) showed a significant positive correlation with stability, particularly at shrub richness levels SR2 and SR4 (Figure 2a, 2c). In contrast, at TR2, the correlation between bacterial diversity and stability became negative, though not statistically significant. At TR4, a significant negative correlation between bacterial diversity and stability emerged, especially at SR0 and SR8 (Figure 2c). For fungi, fungal beta diversity at TR1 was negatively correlated with stability, primarily driven by SR4 (Figure 2b, 2d). We further explored microbial diversity-environment interactions using Mantel correlation analyses (Figure 2e). At TR1, bacterial diversity correlated positively with soil pH, while fungal diversity showed positive correlations with AP, slope, and EW. Fungal composition was significantly influenced by pH, NH₄⁺-N, MBC, NS, altitude, and the C:N ratio (Table S3). At TR2, fungal alpha diversity correlated positively with TN, C:P, and N:P ratios, while bacterial diversity was positively linked to slope, EW, and NS (Table S4). At TR4, fungal diversity was positively correlated with EW and NS, while bacterial composition correlated with the C:P and N:P ratios, and MBC (Table S5). Relationships between microbial community stability and genomic traits Our analysis revealed distinct roles of bacterial and fungal genomic traits (CWM-GC and CWM-BC) in influencing microbial community diversity and ecosystem stability under different tree (TR) and shrub (SR) species richness. We analyzed 13,549 bacterial ASVs and 17,991 fungal ASVs, finding key differences in genomic characteristics: bacteria had a higher GC content, peaking at 57%, compared to fungi, which peaked at 47%. Fungi also exhibited consistently lower base counts than bacteria (Tables S6–S7, Figure S4). As tree species richness increased, bacterial CWM-GC decreased significantly, while fungal CWM-GC increased (Figure 3a). Fungal CWM-BC also increased with tree species richness, whereas bacterial CWM-BC remained unchanged (Figure 3a). Shrub richness did not significantly affect bacterial genomic traits, but at TR2, fungal CWM-GC decreased with increasing shrub richness (Figure S5a). Bacterial CWM-GC and CWM-BC were strongly negatively correlated (p < 0.01), but no significant correlation was found for fungi (Figure 3b). At TR1 and TR2, no significant relationship was observed between ecosystem stability and CWM-GC or CWM-BC for either bacteria or fungi. However, at TR4, both bacterial and fungal CWM-GC showed significant negative correlations with stability (Figure 3c). At SR4, bacterial CWM-GC was negatively correlated with stability, while at SR0, fungal CWM-GC exhibited the same pattern (Figure 3d). Additionally, fungal CWM-BC negatively affected stability at TR1, with SR levels modulating these effects (Figure S5b). Microbial diversity metrics also showed distinct relationships with genomic traits. For bacteria, the Shannon index was positively correlated with CWM-GC at TR1, while beta dissimilarities were negatively correlated with CWM-BC at TR2 and TR4 (Figure 3e). For fungi, the Shannon index was negatively correlated with CWM-BC across all tree species richness levels (TR1, TR2, TR4). Fungal beta dissimilarities were negatively correlated with CWM-GC at TR2 and TR4, but positively correlated with CWM-BC at TR1 (Figure 3f). Contribution of network hub taxa to ecosystem stability We identified the top 10 hub taxa in fungal and bacterial communities at each tree species richness level (TR1, TR2, TR4) based on network association hub scores (Table S8). Using Random Forest analysis, we assessed their contributions to ecosystem stability, ranking taxa by variable importance. The majority of hub taxa contributed positively to stability across all tree richness levels (TR1: 80%; TR2: 80%; TR4: 75%) (Figure 4a). At TR1, the five most influential hub taxa were predominantly bacterial: Pedosphaerales (ASV218, ASV15), Rhizobiales (ASV55), and Acidobacteriales (ASV123), along with the fungal taxon Dothideomycetes (ASV679). At TR2, the top contributors included four bacterial taxa (Rhodospirillales ASV464, Acidobacteriales ASV235, Thermogemmatisporales ASV498, Solibacterales ASV2819, and one fungal taxon Ascomycota_unidentified ASV720. At TR4, the hub taxa consisted of three bacterial taxa (Gemmatimonadetes ASV865, Pedosphaerales ASV15, Solibacterales ASV10), and two fungal taxa (Melanconiella ASV305 and Venturiaceae ASV951) (Tables S9-S10). To explore overlap and specificity of hub taxa across tree species richness levels, we used an upset plot (Figure 4b). Most ASVs were specific to individual tree richness levels, with minimal overlap. Six ASVs were shared between TR1 and TR4, all of which were bacterial taxa (Solibacterales ASV10, not assigned ASVs 6 and ASV 53, Chlorobi ASV2, RhizobialesASV55, and Pedosphaerales ASV15). These shared ASVs exhibited particularly high abundance, highlighting their potential significance in stabilizing ecosystems across these tree richness levels. Phylogenetic trees constructed for each tree species richness level (Figure 4c, Table S11) revealed that many hub taxa lacked precise taxonomic classification. Notably, fungal ASVs 1250 and 821 (unidentified) and bacterial ASV 3917 (Koribacteraceae) were exclusive to TR2. Fungal ASV 729 (unidentified), abundant only in TR4, positively contributed to ecosystem stability. These findings underscore the ecological importance of hub taxa, particularly those with unresolved classifications but strong links to tree species richness. Their roles in stability and potential functional contributions warrant further investigation. Interactive effects of shrub species richness and environmental factors on ecosystem stability and microbial diversity Structural Equation Modeling (SEM) revealed that the effects of shrub species richness (SR) on ecosystem stability, microbial diversity, and genomic traits varied across tree species richness levels (TR1, TR2, and TR4), with environmental factors playing crucial roles in modulating these relationships (Figure 5). At TR1, shrub richness did not significantly affect ecosystem stability, bacterial diversity, or fungal composition, which were primarily driven by environmental factors like pH and slope. Both pH and slope enhanced ecosystem stability, with pH influencing fungal composition directly and indirectly through its effect on total nitrogen (TN), which in turn shaped fungal composition. Fungal composition also regulated NH₄⁺-N levels, contributing to stability. While pH negatively affected bacterial diversity, this diversity positively contributed to ecosystem stability. At TR2, shrub species richness significantly and positively regulated ecosystem stability. This effect was mediated through multiple pathways, including the enhancement of bacterial diversity, which positively contributed to stability. Shrub richness also boosted soil available phosphorus (AP) through environmental factors like slope and altitude, which enhanced fungal composition. Fungal composition either directly promoted stability or indirectly regulated soil C:N ratios, further contributing to stability. However, total carbon (TC), positively influenced by shrub richness, was negatively affected by environmental factors, leading to an increase in the C:N ratio, which ultimately supported stability. At TR4, shrub species richness positively regulated ecosystem stability. In addition to directly enhancing stability, shrub richness facilitated bacterial CWM-GC (community-weighted mean GC content), which reduced microbial biomass nitrogen (MBN). This reduction in MBN was positively associated with microbial biomass carbon (MBC). Furthermore, increased shrub richness led to higher fungal CWM-GC, which reduced MBC and thereby decreased stability. Environmental factors like soil C:N ratio and water availability (EW) also played significant roles, with both bacterial and fungal CWM-GC positively influenced by these factors. The C:N ratio, positively affected by shrub richness, was negatively impacted by EW.

From competition to facilitation: how tree species richness alters the effect of shrub diversity on ecosystem stability We found that shrub species richness generally promoted microbial community stability, consistent with BEF theory, which attributes stability gains to mechanisms like species complementarity, resource partitioning, and functional redundancy (Loreau & Hector, 2001; Loreau & de Mazancourt, 2013; Isbell et al., 2015; Schnabel et al., 2021). Importantly, this positive effect was modulated by tree species richness, revealing an interaction between vertical vegetation layers. At low tree richness (TR1), shrub diversity had a weak effect on stability, likely due to limited niche opportunities and resource heterogeneity (Isbell et al., 2015; Craven et al., 2018). However, a noticeable increase in stability at intermediate shrub richness (SR2–SR4) suggests that complementarity among shrubs becomes more effective after a diversity threshold is reached (Hautier et al., 2014). In contrast, under higher tree richness (TR2 and TR4), shrub richness had stronger and more consistent stabilizing effects, implying facilitative interactions and increased functional redundancy between trees and shrubs (Wagg et al., 2014; Feng et al., 2022). Interestingly, stability declined at the highest shrub richness level (SR8), indicating possible over-competition for limited resources—an outcome consistent with theoretical predictions of diminishing returns at very high diversity (Hautier et al., 2015, Fichtner et al., 2018). The influence of environmental variables varied across tree richness levels, highlighting the role of abiotic filters in shaping biodiversity–stability relationships. For example, NO₃⁻-N significantly promoted stability at TR2, likely by enhancing productivity and reducing variability (Xu et al., 2018), while at TR4, elevated pH and NO₃⁻-N were negatively associated with stability, possibly due to shifts in microbial dynamics under high-diversity conditions (Wagg et al., 2014; Delgado-Baquerizo et al., 2017). By integrating shrub and tree richness with soil and topographic variables, our study moves beyond traditional BEF approaches, offering a multifactorial perspective on ecosystem stability. These results emphasize that optimal biodiversity configurations may depend on both biotic interactions and environmental context. \papertype Original Article Multitrophic interactions and environmental modulation of microbial diversity-stability relationships across plant biodiversity gradients Our results reveal complex, context-dependent interactions among microbial diversity, community stability, and environmental factors across varying biodiversity scenarios. At low tree species richness (TR1), bacterial alpha diversity (Simpson index) was positively correlated with stability, particularly at intermediate shrub richness (SR2 and SR4), suggesting that moderate plant diversity promotes microbial functional redundancy and resilience (Wagg et al., 2014; Delgado-Baquerizo et al., 2017). These findings support the idea that plant diversity enhances microbial diversity via increased resource heterogeneity but also highlight the influence of community composition (Fierer et al., 2009). In contrast, at higher tree richness levels (TR2 and TR4), the bacterial diversity–stability relationship changed. A significant negative correlation at TR4—most evident at extreme shrub richness levels (SR0 and SR8)—suggests that excessive diversity across multiple trophic levels may intensify competition or disrupt cooperative dynamics (Hautier et al., 2015; Chen et al., 2020). This contrasts with the typically positive biodiversity-stability relationships observed in earlier studies (Isbell et al., 2015) and highlights the need to consider multitrophic and nonlinear interactions in microbial ecology. The patterns observed in fungal communities further underscore the intricate ways in which microbes influence stability. At TR1, fungal beta diversity negatively correlated with stability, especially at SR4, indicating that high community dissimilarity may reflect destabilizing functional divergence (Van der Heijden et al., 2008). While fungal diversity is often linked to nutrient cycling and productivity (Treseder et al., 2012), our findings suggest that under certain conditions, competition or niche differentiation may reduce stability. Mantel tests revealed that environmental factors modulate these diversity–stability relationships. At TR1, bacterial diversity positively correlated with pH and total nitrogen, while fungal diversity was associated with available phosphorus, slope, and slope aspect—consistent with prior studies on microhabitat heterogeneity and nutrient availability (Lauber et al., 2009, Peay et al., 2016). In contrast, at higher tree richness levels, microbial diversity correlated more strongly with stoichiometric ratios (e.g., C:P, N:P), reflecting shifts in nutrient dynamics and microbial responses under diverse plant communities. Overall, our findings go beyond single-dimension analyses of microbial diversity by demonstrating how both taxonomic richness and community composition contribute to stability. By integrating microbial dynamics within a multitrophic and environmental framework, we provide a more mechanistic understanding of biodiversity–stability relationships and highlight the interplay between biotic complexity and abiotic modulation in shaping ecosystem functioning. Genomic traits as mediators of microbial contributions to ecosystem stability Understanding how microbial genomic traits influence ecosystem stability involves addressing several challenges, particularly the limitations of using full-genome GC content, which is hindered by incomplete genomes and the underrepresentation of uncultured species (Johnson et al., 2019; Straub et al., 2020). Additionally, the uneven representation of microbial species in genomic databases complicates the accurate assessment of ecosystem functions, such as community stability (Wang et al., 2023; Su et al., 2023). To overcome these issues, our study focused on the GC content of amplicon regions (16S rRNA and ITS), which are reliable proxies for microbial functional diversity. This approach is particularly effective since GC content correlates with metabolic pathways, environmental adaptability, and microbial interactions (Yamane et al., 2011; Peay et al., 2016). Our findings reveal complex relationships between community-weighted mean GC content (CWM-GC), community-weighted mean base count (CWM-BC), and microbial communities across varying tree (TR) and shrub (SR) species richness levels, highlighting the nuanced role these genomic traits play in ecosystem stability. At lower tree species richness levels (TR1 and TR2), no significant correlation was found between CWM-GC and stability. However, at the highest tree richness level (TR4), a significant negative correlation emerged between both bacterial and fungal CWM-GC content and stability, particularly at higher shrub richness levels (SR4). This suggests that, under resource-rich conditions, increased microbial genetic diversity may result in heightened competition for limited resources, ultimately reducing stability (Hautier et al., 2015). The GC content of amplicon regions also reflects microbial resilience to extreme environmental conditions, such as in thermophilic communities (Hu et al., 2022; Yamane et al., 2011). Our results align with earlier studies suggesting that microbial genomic traits, including GC content, are indicative of microbial community functions and their interactions with environmental factors (Chuckran et al., 2021). Specifically, at TR1, bacterial Shannon diversity was positively correlated with GC content, indicating that in low-diversity environments, higher genetic diversity within bacterial communities may contribute to functional stability. This suggests that community stability in such environments may depend more on ecological mechanisms like resource partitioning, rather than solely on genomic traits (Loreau & Hector, 2001; Hautier et al., 2014). Overall, our study emphasizes the usefulness of amplicon-based GC content as an informative and practical measure of microbial genomic traits, especially when whole-genome data is limited. This approach offers valuable insights into how microbial diversity and functional stability are influenced by genomic traits, shedding light on the mechanisms through which these traits mediate microbial contributions to ecosystem stability. Ecological significance and functional insights of microbial hub taxa Our study highlights the critical role of microbial hub taxa, both bacterial and fungal, in stabilizing ecosystems, with their contributions varying across different biodiversity contexts. Hub taxa, characterized by high connectivity and disproportionate effects on community function, are key drivers of ecosystem stability (Banerjee et al., 2018). Most identified hub taxa (TR1: 80%; TR2: 80%; TR4: 75%) positively contributed to stability, illustrating the dynamic nature of microbial contributions across biodiversity gradients. Notably, TR2 showed the highest number of unique hub amplicon sequence variants (ASVs, 20), with no overlap with other tree richness levels, emphasizing the complexity of microbial interactions at intermediate biodiversity levels. This supports the ”diversity-interaction hypothesis” (Loreau et al., 2001), which suggests that intermediate biodiversity maximizes functional complementarity and resource utilization. In contrast, six shared bacterial ASVs between TR1 and TR4, which were abundant and contributed stably to ecosystem stability, may act as keystone taxa in microbial networks (Power et al., 1996). A significant portion of the identified hub taxa lacked clear taxonomic classification, in line with findings from other studies that many environmentally significant microorganisms remain unidentified (Lennon & Locey, 2016). Unique fungal ASVs (e.g., ASV729 and ASV821) and bacterial ASVs (e.g., ASV3917), strongly associated with specific tree richness levels, underscore the need for further exploration of these taxa using metagenomic and transcriptomic approaches to understand their functional roles in ecosystem stability (Jansson & Hofmockel, 2020). Context-dependent effects of shrub richness, environmental factors, and microbial traits on ecosystem stability Using Structural Equation Modeling (SEM), we identified context-dependent mechanisms through which shrub richness, environmental factors, and microbial traits shape ecosystem stability. Consistent with previous studies (Tilman et al., 2014; Craven et al., 2018), our findings show that shrub species richness generally promotes ecosystem stability, probably by enhancing functional complementarity and resource use efficiency. However, the stabilizing effect of shrub richness was contingent on tree species richness. At TR1, no significant effect of shrub richness on stability was observed, emphasizing the importance of biodiversity context in shaping biodiversity-stability relationships (Loreau & Hector, 2001). At TR2, shrub richness positively influenced bacterial diversity and increased AP, both contributing to enhanced stability, aligning with studies highlighting the role of plant diversity in microbial diversity and nutrient cycling (Chen et al., 2020). In contrast, at TR4, shrub richness positively affected bacterial CWM-GC and MBC, suggesting that microbial genomic traits are critical for stabilizing ecosystems in high tree species richness environments. Environmental factors, such as pH, slope, and east-west (EW) aspect, significantly modulated the relationships between shrub richness, microbial traits, and stability. At TR1, soil pH was the primary driver, influencing fungal composition and indirectly regulating stability through nitrogen and ammonium levels, consistent with prior research on the impact of pH on microbial communities (Rousk et al., 2010). At TR2 and TR4, environmental gradients like slope, altitude, and EW interacted with shrub richness to shape microbial diversity and ecosystem functions, supporting findings that topographic and climatic factors influence microbial traits and stability (Wu et al., 2021). This highlights the need to account for environmental heterogeneity when examining biodiversity-stability relationships. A novel aspect of our study was the integration of community-weighted mean GC content (CWM-GC) into assessments of ecosystem stability. Unlike studies focused solely on microbial diversity, we show that functional genomic traits provide deeper insights into the mechanisms underlying stability. At TR4, bacterial and fungal CWM-GC influenced MBC and stability differently, emphasizing the distinct roles of microbial taxa in nutrient cycling and resource allocation (Fierer et al., 2021). Additionally, shrub richness promoted AP, total carbon (TC), and soil C:N ratios, reinforcing the idea that plant diversity enhances nutrient availability and soil organic matter dynamics (Delgado-Baquerizo et al., 2016). However, the negative effects of fungal CWM-GC on MBC and stability suggest the need for further exploration of competitive interactions or resource trade-offs among microbial communities.

We would like to express our sincere gratitude to Keping Ma, Xiaojuan Liu, and the staff of the BEF-China platform for their invaluable support and assistance throughout this study. Their contributions were essential in facilitating our research and ensuring its success. Conflict of Interest Statement The authors declare no competing interests. Funding This study was supported by the National Key Research and Development Project of China (grant number 2022YFF1303201) and the National Natural Science Foundation of China (grant number 32071644).

1. Tilman D, Isbell F, Cowles JM. (2014). Biodiversity and ecosystem functioning. Annu Rev Ecol Evol Syst, 45: 471–93. 2. Hughes TP, Baird AH, Bellwood DR, et al. (2003). Climate change, human impacts, and the resilience of coral reefs. Science, 301: 929–33. 3. Oliver TH, et al. (2015). Biodiversity and resilience of ecosystem functions. Trends Ecol Evol, 30: 513–23. 4. Isbell F, Gonzalez A, Loreau M, et al. (2017). Linking the influence and dependence of people on biodiversity across scales. Nature, 546: 65–72. 5. Hong P, Schmid B, De Laender F, et al. (2022). Biodiversity promotes ecosystem functioning despite environmental change. Ecol Lett, 25(2): 555–69. 6. Schnabel F, Liu X, Kunz M, et al. (2021). Species richness stabilizes productivity via asynchrony and drought-tolerance diversity in a large-scale tree biodiversity experiment. Sci Adv, 7: eabk1643. 7. Meng Y, Li S, Wang S, et al. (2023). Scale-dependent changes in ecosystem temporal stability over six decades of succession. Sci Adv, 9: eadi1279. 8. Qiao X, Lamy T, Wang S, et al. (2023). Latitudinal patterns of forest ecosystem stability across spatial scales as affected by biodiversity and environmental heterogeneity. Glob Change Biol, 29: 2242–55. 9. Yan Y, Piao S, Hammond WM, et al. (2024). Climate-induced tree-mortality pulses are obscured by broad-scale and long-term greening. Nat Ecol Evol, 8: 912–23. 10. Liang J, Crowther TW, Picard N, et al. (2016). Positive biodiversity-productivity relationship predominant in global forests. Science, 354: aaf8957. 11. Fichtner A, Härdtle W, Bruelheide H, et al. (2018). Neighbourhood interactions drive overyielding in mixed-species tree communities. Nat Commun, 9: 1144. 12. Huang Y, Chen Y, Castro-Izaguirre N, et al. (2018). Impacts of species richness on productivity in a large-scale subtropical forest experiment. Science, 362: 80–3. 13. Chen L, Jiang L, Jing X, et al. (2021). Above‐ and belowground biodiversity jointly drive ecosystem stability in natural alpine grasslands on the Tibetan Plateau. Glob Ecol Biogeogr, 30(7): 1418–29. 14. Wu L, Chen H, Chen D, et al. (2023). Soil biota diversity and plant diversity both contributed to ecosystem stability in grasslands. Ecol Lett, 26(6): 858–68. 15. Chen C, Xiao W, Chen HYH. (2025a). Meta-analysis reveals global variations in plant diversity effects on productivity. Nature ; https://doi.org/10.1038/s41586-024-08407-8. 16. Jucker T, Bouriaud O, Avacaritei D, et al. (2014). Stabilizing effects of diversity on aboveground wood production in forest ecosystems: linking patterns and processes. Ecol Lett, 17: 1560–9. 17. Del Río M, Pretzsch H, Ruíz‐Peinado R, et al. (2017). Species interactions increase the temporal stability of community productivity in Pinus sylvestris–Fagus sylvatica mixtures across Europe. J Ecol, 105: 1032–43. 18. Schnabel F, Schwarz JA, Dănescu A, et al. (2019). Drivers of productivity and its temporal stability in a tropical tree diversity experiment. Glob Change Biol, 25: 4257–72. 19. Feng Y, et al. (2022). Multispecies forest plantations outyield monocultures across a broad range of conditions. Science, 376: 865–8. 20. Urgoiti J, et al. (2022). No complementarity no gain—net diversity effects on tree productivity occur once complementarity emerges during early stand development. Ecol Lett, 25: 851–62. 21. Cornwell WK, Ackerly DD. (2010). A link between plant traits and abundance: evidence from coastal California woody plants. J Ecol, 98: 814–21. 22. Wardle DA, Jonsson M, Bansal S, et al. (2012). Linking vegetation change, carbon sequestration and biodiversity: insights from island ecosystems in a long‐term natural experiment. J Ecol, 100: 16–30. 23. Temperton VM, Mwangi PN, Scherer-Lorenzen M, et al. (2007). Positive interactions between nitrogen-fixing legumes and four different neighbouring species in a biodiversity experiment. Oecologia, 151: 190–205. 24. Fornara DA, Tilman D. (2008). Plant functional composition influences rates of soil carbon and nitrogen accumulation. J Ecol, 96: 314–22. 25. Deng J, Fang S, Fang X, et al. (2023). Forest understory vegetation study: current status and future trends. For Res, 3: 1. 26. Tao S, Veen GF, Zhang N, et al. (2023). Tree and shrub richness modifies subtropical tree productivity by regulating the diversity and community composition of soil bacteria and archaea. Microbiome, 11: 261. 27. Balandier P, Mårell A, Prévosto B, et al. (2022). Tamm review: Forest understorey and overstorey interactions: So much more than just light interception by trees. For Ecol Manag, 526: 120584. 28. Lecomte X, Paulo JA, Tomé M, et al. (2022). Shrub understorey clearing and drought affects water status and growth of juvenile Quercus suber trees. For Ecol Manag, 503: 119760. 29. Loreau M, Hector A. (2001). Partitioning selection and complementarity in biodiversity experiments. Nature, 412: 72–6. 30. Loreau M, Naeem S, Inchausti P, et al. (2001). Biodiversity and ecosystem functioning: current knowledge and future challenges. Science, 294(5543): 804–8. 31. Isbell F, Craven D, Connolly J, et al. (2015). Biodiversity increases the resistance of ecosystem productivity to climate extremes. Nature, 526: 574–7. 32. Chen C, Yan GY, Schmid B, et al. (2025b). Understory shrub diversity: equally vital as overstory tree diversity to promote forest productivity. Natl Sci Rev, nwaf093. 33. Van Der Heijden MGA, Bardgett RD, Van Straalen NM. (2008). The unseen majority: soil microbes as drivers of plant diversity and productivity in terrestrial ecosystems. Ecol Lett, 11: 296–310. 34. Wagg C, Bender SF, Widmer F, et al. (2014). Soil biodiversity and soil community composition determine ecosystem multifunctionality. Proc Natl Acad Sci USA, 111: 5266–70. 35. Banerjee S, Van Der Heijden MGA. (2023). Soil microbiomes and one health. Nat Rev Microbiol, 21: 6–20. 36. Wardle DA, Bardgett RD, Klironomos JN, et al. (2004). Ecological linkages between aboveground and belowground biota. Science, 304: 1629–33. 37. Bardgett RD, Van Der Putten WH. (2014). Belowground biodiversity and ecosystem functioning. Nature, 515: 505–11. 38. Delgado-Baquerizo M, Maestre FT, Reich PB, et al. (2016). Microbial diversity drives multifunctionality in terrestrial ecosystems. Nat Commun, 7: 10541. 39. Crowther TW, Van den Hoogen J, Wan J, et al. (2019). The global soil community and its influence on biogeochemistry. Science, 365: eaav0550. 40. Tedersoo L, Bahram M, Zobel M. (2020). How mycorrhizal associations drive plant population and community biology. Science, 367: eaba1223. 41. Eisenhauer N, Scheu S, Jousset A. (2012). Bacterial diversity stabilizes community productivity. PLoS One, 7: e34517. 42. Rillig MC, Aguilar-Trigueros CA, Bergmann J, et al. (2015). Plant root and mycorrhizal fungal traits for understanding soil aggregation. New Phytol, 205: 1385–8. 43. Hong P, Schmid B, De Laender F, et al. (2022). Biodiversity promotes ecosystem functioning despite environmental change. Ecol Lett, 25(2): 555–69. 44. Baldrian P, López-Mondéjar R, Kohout P. (2023). Forest microbiome and global change. Nat Rev Microbiol, 21(8): 487–501. 45. Torsvik V, Øvreås L. (2002). Microbial diversity and function in soil: from genes to ecosystems. Curr Opin Microbiol, 5(3): 240–5. 46. Dick GJ, Andersson AF, Baker BJ, et al. (2009). Community-wide analysis of microbial genome sequence signatures. Genome Biol, 10: 1–16. 47. Lauber CL, Hamady M, Knight R, et al. (2009). Pyrosequencing-based assessment of soil pH as a predictor of soil bacterial community structure at the continental scale. Appl Environ Microbiol, 75(15): 5111–20. 48. Jansson JK, Hofmockel KS. (2020). Soil microbiomes and climate change. Nat Rev Microbiol, 18(1): 35–46. 49. Beugnon R, Du J, Cesarz S, et al. (2021). Tree diversity and soil chemical properties drive the linkages between soil microbial community and ecosystem functioning. ISME Commun, 1(1): 41. 50. Bruelheide H, Schmidt K, Seidler G, Nadrowski K. (2013a). Site A plots: diversity treatments, coordinates, topography. BEF-China data portal, http://china.befdata.biow.uni-leipzig.de/datasets/71. 51. Bruelheide H, Schmidt K, Seidler G, Nadrowski K. (2013b). Site B plots: diversity treatments, coordinates, topography. BEF-China data portal, http://china.befdata.biow.uni-leipzig.de/datasets/46. 52. Bruelheide H, Nadrowski K, Assmann T, et al. (2014). Designing forest biodiversity experiments: general considerations illustrated by a new large experiment in subtropical China. Methods in Ecology and Evolution, 5(1): 74-89. 53. Brookes PC, Landman A, Pruden G, et al. (1985). Chloroform fumigation and the release of soil nitrogen: a rapid direct extraction method to measure microbial biomass nitrogen in soil. Soil Biol Biochem, 17: 837–42. 54. Vance ED, Brookes PC, Jenkinson DS. (1987). An extraction method for measuring soil microbial biomass C. Soil Biol Biochem, 19: 703–7. 55. Bongers FJ, Schmid B, Bruelheide H, et al. (2021). Functional diversity effects on productivity increase with age in a forest biodiversity experiment. Nat Ecol Evol, 5: 1–10. 56. Craven D, Eisenhauer N, Pearse WD, et al. (2018). Multiple facets of biodiversity drive the diversity–stability relationship. Nat Ecol Evol, 2(10): 1579–87. 57. Wickham H. (2011). ggplot2. Wiley Interdiscip Rev Comput Stat, 3(2): 180–5. 58. Caporaso JG, Lauber CL, Walters WA, et al. (2011). Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proc Natl Acad Sci USA, 108: 4516–20. 59. Toju H, Tanabe AS, Yamamoto S, et al (2012). High-coverage ITS primers for the DNA-based identification of ascomycetes and basidiomycetes in environmental samples. PLoS One, 7(7): e40863. 60. Xiong H, Zhang N, Takayoshi K, et al (2023). Tree mycorrhizal types and the interaction between tree and shrub species richness shape soil fungal communities in a subtropical forest in China. Plant Soil, 493(1): 61–77. 61. Bolyen E, Rideout JR, Dillon MR, Bokulich NA, Abnet CC, Al-Ghalith GA, et al (2019). Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol, 37: 852–857. 62. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP (2016). DADA2: high-resolution sample inference from Illumina amplicon data. Nat Methods, 13: 581–583. 63. R Core Team (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available at: https://www.R-project.org/. 64. Huang H (2021). linkET: Everything is Linkable. R package version 0.0.3. 65. Põlme S, Abarenkov K, Nilsson RH, et al (2020). FungalTraits: a user-friendly traits database of fungi and fungus-like stramenopiles. Fungal Divers, 105: 1–16. 66. Wickham H, Çetinkaya-Rundel M, Grolemund G (2023). R for Data Science. O’Reilly Media, Inc. 67. Douglas GM, Maffei VJ, Zaneveld JR, et al (2020). PICRUSt2 for prediction of metagenome functions. Nat Biotechnol, 38(6): 685–688. 68. Makowski D, Ben-Shachar M, Patil I, Lüdecke D (2020). Methods and algorithms for correlation analysis in R. J Open Source Software, 5(51): 2306. 69. Wen T, Xie P, Yang S, et al (2022). ggClusterNet: An R package for microbiome network analysis and modularity-based multiple network layouts. Imeta, 1(3): e32. 70. Breiman L (2001). Random forests. Mach Learn, 45: 5–32. 71. Conway JR, Lex A, Gehlenborg N (2017). UpSetR: an R package for the visualization of intersecting sets and their properties. Bioinformatics, 33(18): 2938–2940. 72. Paradis E, Schliep K (2019). ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics, 35(3): 526–528. 73. Lefcheck JS (2016). piecewiseSEM: Piecewise structural equation modelling in R for ecology, evolution, and systematics. Methods Ecol Evol, 7(5): 573–579. 74. Loreau M, De Mazancourt C (2013). Biodiversity and ecosystem stability: a synthesis of underlying mechanisms. Ecol Lett, 16: 106–115. 75. Hautier Y, Seabloom EW, Borer ET, et al (2014). Eutrophication weakens stabilizing effects of diversity in natural grasslands. Nature, 508(7497): 521–525. 76. Hautier Y, Tilman D, Isbell F, et al (2015). Anthropogenic environmental changes affect ecosystem stability via biodiversity. Science, 348(6232): 336–340. 77. Xu Z, Li MH, Zimmermann NE, et al (2018). Plant functional diversity modulates global environmental change effects on grassland productivity. J Ecol, 106(5): 1941–1951. 78. Delgado-Baquerizo M, Reich PB, Khachane AN, et al (2017). It is elemental: soil nutrient stoichiometry drives bacterial diversity. Environ Microbiol, 19(3): 1176–1188. 79. Fierer N, Strickland MS, Liptzin D, et al (2009). Global patterns in belowground communities. Ecol Lett, 12(11): 1238–1249. 80. Chen C, Chen HYH, Chen X (2020). Functional diversity enhances, but exploitative traits reduce tree mixture effects on microbial biomass. Funct Ecol, 34(2): 276–286. 81. Treseder KK, Balser TC, Bradford MA, et al (2012). Integrating microbial ecology into ecosystem models: challenges and priorities. Biogeochemistry, 109: 7–18. 82. Peay KG, Kennedy PG, Talbot JM (2016). Dimensions of biodiversity in the Earth mycobiome. Nat Rev Microbiol, 14(7): 434–447. 83. Johnson JS, Spakowicz DJ, Hong BY, et al (2019). Evaluation of 16S rRNA gene sequencing for species and strain-level microbiome analysis. Nat Commun, 10(1): 5029. 84. Straub D, Blackwell N, Langarica-Fuentes A, et al (2020). Interpretations of environmental microbial community studies are biased by the selected 16S rRNA (gene) amplicon sequencing pipeline. Front Microbiol, 11: 550420. 85. Wang C, Yu QY, Ji NN, Zheng Y, Taylor JW, Guo LD, Gao C (2023). Bacterial genome size and gene functional diversity negatively correlate with taxonomic diversity along a pH gradient. Nat Commun, 14(1): 7437. https://doi.org/10.1038/s41467-023-43297-w. 86. Su L, Teske AP, MacGregor BJ, et al (2023). Thermal selection of microbial communities and preservation of microbial function in Guaymas Basin hydrothermal sediments. Appl Environ Microbiol, 89(3): e00018–23. 87. Yamane K, Hattori Y, Ohtagaki H, Fujiwara K (2011). Microbial diversity with dominance of 16S rRNA gene sequences with high GC contents at 74 and 98 °C subsurface crude oil deposits in Japan. FEMS Microbiol Ecol, 76(2): 220–235. 88. Hu EZ, Lan XR, Liu ZL, et al (2022). A positive correlation between GC content and growth temperature in prokaryotes. BMC Genomics, 23: 110. 89. Chuckran PF, Hungate BA, Schwartz E, Dijkstra P (2021). Variation in genomic traits of microbial communities among ecosystems. FEMS Microbes, 2(1): xtab020. 90. Wagg C, Schlaeppi K, Banerjee S, et al (2019). Fungal-bacterial diversity and microbiome complexity predict ecosystem functioning. Nat Commun, 10(1): 4841. 91. Martiny JBH, Jones SE, Lennon JT, et al (2015). Microbiomes in light of traits: a phylogenetic perspective. Science, 350(6261): aac9323. 92. Fierer N, Wood SA, de Mesquita CPB (2021). How microbes can, and cannot, be used to assess soil health. Soil Biol Biochem, 153: 108111. 93. Smith SE, Read DJ (2008). Mycorrhizal Symbiosis . London, UK: Academic Press and Elsevier. 94. Banerjee S, Schlaeppi K, van der Heijden MGA (2018). Keystone taxa as drivers of microbiome structure and functioning. Nat Rev Microbiol, 16(9): 567–576. 95. Power ME, Tilman D, Estes JA, et al (1996). Challenges in the quest for keystones: identifying keystone species is difficult—but essential to understanding how loss of species will affect ecosystems. BioScience, 46(8): 609–620. 96. Lennon JT, Locey KJ (2016). The underestimation of global microbial diversity. mBio, 7(5): 10-1128. 97. Rousk J, Bååth E, Brookes PC, et al (2010). Soil bacterial and fungal communities across a pH gradient in an arable soil. ISME J, 4(10): 1340–1351. 98. Wu L, Chen H, Chen D, et al (2023). Soil biota diversity and plant diversity both contributed to ecosystem stability in grasslands. Ecology Letters, 26(6): 858–868. LEGENDS Figure 1. Effects of Shrub Species Richness, Soil Physicochemical, and Environmental Factors on Tree Community Stability Across Three Tree Species Richness Levels. (a) Experimental design illustrating how the interaction among varying levels of tree species richness (TR1, TR2, and TR4) and shrub species richness (SR0, SR2, SR4, SR8) collectively affects community stability. The diagram further highlights the pivotal role of microbial communities in driving these processes, while also illustrating how microbe-driven processes are influenced by soil properties and topographic factors. (b) Relationship between shrub species richness and community stability at three tree species richness levels: monoculture (TR1), two-species mixture (TR2), and four-species mixture (TR4). Regression lines indicate trends, with R² values and p-values representing significance levels. (c) Forest plots summarizing the outcomes of linear regression for community stability, incorporating soil physicochemical factors, terrain attributes, and directional gradients as predictors. Effect sizes (regression coefficients), 95% confidence intervals, and significance levels (p-values) are shown for each predictor at TR1, TR2, and TR4. Positive and negative effects are indicated in blue and red, respectively, with significance levels denoted by *p < 0.05, **p < 0.01, and ***p < 0.001. Abbreviations: SM, soil moisture; NO₃⁻-N, nitrate; NH₄⁺-N, ammonium; MBC, microbial biomass carbon; MBN, microbial biomass nitrogen; TC, total carbon; AP, available phosphorus; C:N ratio, carbon-to-nitrogen ratio; N:P ratio, nitrogen-to-phosphorus ratio; EW, east-west component of slope aspect; NS, north-south component of slope aspect. Figure 2. Relationships Between Microbial Community Diversity, Ecosystem Stability, and Environmental Factors Across Tree Species Richness Levels. (a, b) Relationships between alpha diversity (Simpson index) and beta diversity (unweighted UniFrac distance dissimilarity) of both bacteria and fungi with ecosystem stability at three tree species richness levels (TR1, TR2, and TR4). (c, d) Effects of shrub species richness (SR0, SR2, SR4, and SR8) on the relationships between microbial diversity and ecosystem stability across tree species richness levels. Significant relationships for specific shrub species richness levels are indicated by solid lines and p-values (*p < 0.05). (e) Mantel correlation analysis illustrating the relationships between microbial diversity (alpha and beta) and environmental factors (e.g., soil physicochemical properties, topographic features, and nutrient elements) across three levels of tree species richness. Correlation plots display Spearman’s rank correlation coefficients, using a color gradient to indicate the strength of correlations. Size and color coding indicate Mantel’s p-values and correlation coefficients, respectively. Statistical significance is indicated by a color scheme and size legend, while insignificant correlations are denoted by a grey color. Figure 3. Effects of Tree Species Richness on Community Weighted Mean Guanine-Cytosine Content (CWM-GC) and Base Count (CWM-BC) and Their Relationship with Community Stability and Microbial Diversity in Fungal and Bacterial Communities. (a)Violin plots illustrating the changes in bacterial and fungal CWM-GC and CWM-BC across tree species richness (TR) levels. (b) Correlation between bacterial and fungal CWM-GC and CWM-BC. (c) Relationship between CWM-GC and community stability across tree species richness (TR) levels. (d) Modulation of CWM-GC-stability relationships by shrub species richness (SR) at TR4. (e-f) Relationship between bacterial and fungal diversity metrics (Shannon index and UniFrac dissimilarities) and CWM-GC and CWM-BC. Significance levels are indicated by *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Figure 4. Hub taxa and their contribution to ecosystem stability across tree species richness levels. (a) Random forest analysis of the top 10 hub taxa in fungal and bacterial communities across three tree species richness levels (TR1, TR2, and TR4). Taxa are ranked according to their variable importance in relation to ecosystem stability. (b) The overlap and specificity of hub taxa across the three tree species richness levels. (c) A phylogenetic tree delineating the hub taxa identified at each tree species richness level, with color coding corresponding to their taxonomic classification. Each ASV is also represented by a heatmap showing its abundance across the four levels of shrub species richness at TR1, TR2, and TR4. Figure 5. Structural Equation Modeling (SEM) Elucidating the Interconnections Among Shrub Species Richness, Environmental Factors, and Ecosystem Stability Across Three Tree Species Richness Levels (TR1, TR2, and TR4). The SEM analysis illustrates how shrub species richness (SR), in conjunction with environmental factors influences soil physicochemical properties, microbial community diversity, composition, and genomic traits across varying levels of tree species richness (TR1, TR2, and TR4). These microbial factors subsequently modulate soil nutrients and, ultimately, ecosystem stability. Arrows indicate direct effects, with solid lines representing significant relationships and dashed lines representing non-significant effects. Statistical significance is denoted by *p < 0.05, **p < 0.01, and ***p < 0.001. Abbreviations: EF, environmental factors; NH₄⁺-N, ammonium; MBC, microbial biomass carbon; MBN, microbial biomass nitrogen; TC, total carbon; TN, total nitrogen; AP, available phosphorus; C: N, carbon-to-nitrogen ratio; EW, east-west component of slope aspect. SUPPLEMENTARY FILE Figure S1. Relationships between bacterial community alpha diversity (Shannon diversity, evenness, abundance, and richness) and ecosystem stability across three tree species richness levels (TR1, TR2, and TR4). Statistical significance is denoted by *p < 0.05, **p < 0.01, and ***p < 0.001 Figure S2. Relationships between fungal community alpha diversity (Shannon diversity, evenness, abundance, and richness) and ecosystem stability across three tree species richness levels (TR1, TR2, and TR4). Statistical significance is denoted by *p < 0.05, **p < 0.01, and ***p < 0.001 Figure S3. Relationships between bacterial and fungal community alpha diversity (Shannon diversity, evenness, abundance, and richness) and ecosystem stability across four shrub species richness levels (SR0, SR2, SR4, and SR8) within three tree species richness levels (TR1, TR2, and TR4). Statistical significance is denoted by *p < 0.05, **p < 0.01, ***p < 0.001, and ns: not significant. Figure S4. Distribution and Density of Guanine-Cytosine (GC) Content and Base Count in Bacterial and Fungal Amplicon Sequencing Variants. Figure S5. Effects of Shrub Species Richness on Community Weighted Mean Guanine-Cytosine Content (CWM-GC) and Base Count (CWM-BC) and Their Relationship with Community Stability Across Three Tree Species Richness Levels. Statistical significance is denoted by *p < 0.05, **p < 0.01, ***p < 0.001, and ns: not significant. Table S1. Overview of sample information across the experimental plots. Table S2. Environmental factors and microbial biomass measurements for each sample. Tables S3-S5. Mantel correlations between fungal and bacterial community diversity indices and environmental factors across three tree species richness levels (TR1, TR2, and TR4). Table S6. Guanine-cytosine (GC) content and base count of each bacterial amplicon sequence variant, along with their relative abundance in each sample. Table S7. Guanine-cytosine (GC) content and base count of each fungal amplicon sequence variant, along with their relative abundance in each sample. Table S8. Hub taxa in bacterial and fungal communities across three tree species richness levels (TR1, TR2, and TR4). Table S9. Taxonomic assignment of each bacterial amplicon sequence variant. Table S10. Taxonomic assignment of each fungal amplicon sequence variant. Table S11. FASTA sequences of hub taxa in bacterial and fungal communities used for phylogenetic tree construction. Information & Authors Information Version history Copyright This work is licensed under a Non Exclusive No Reuse License.

Authors Metrics & Citations Metrics Article Usage 285views 80downloads Citations Download citation Siqi Tao, Laiye Qu, Paul Kardol, et al. Context-Dependent Contributions of Shrub Species Richness to Ecosystem Stability Across Tree Diversity Gradients: Insights from Microbial Mediators. Authorea. 28 April 2025. DOI: https://doi.org/10.22541/au.174584633.30760731/v1 DOI: https://doi.org/10.22541/au.174584633.30760731/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu.