Tree diversity and soil chemical properties drive the linkages between soil microbial community and ecosystem functioning

The study addresses how tree diversity and soil chemical properties shape soil microbial community facets (biomass, taxonomic and functional profiles) and how these in turn control microbial functions, particularly microbial respiration that is central to soil carbon balance and ecosystem functioning. Prior research shows vegetation type and diversity affect microbial communities, and abiotic conditions (soil C, N, P, pH, moisture) also influence microbial assembly and activity. Plant diversity can increase carbon inputs via litter and rhizodeposition, enhancing microbial biomass and activity, but the pathways linking plant diversity to microbial respiration via distinct microbial facets remain unclear. The authors consider different microbial community facets: biomass (e.g., PLFA, SIR), taxonomic profiles (16S/ITS), functional profiles (functional genes), and realized functions (physiological potential via MicroResp and microbial respiration). They hypothesize: (H1) tree diversity increases microbial community facets and functions; (H2) microbial biomass, taxonomic and functional profiles are correlated and together drive microbial functions; (H3) microbial physiological potential links microbial community facets to microbial respiration; and (H4) environmental conditions (tree diversity and soil chemistry) co-determine microbial respiration by modulating microbial community facets. The goal is to integrate these facets and drivers within a unified framework to improve understanding and modeling of soil carbon dynamics.

Study site and design: The study was conducted at the BEF-China tree diversity experiment in Jiangxi Province, southeast China (29.08–29.11°N, 117.90–117.93°E). Plots span tree species richness levels of 1, 2, 4, 8, and 16/24 species. In August–September 2018 (pre-litterfall), 150 composite soil samples (from four 5 cm diameter, 10 cm depth cores taken 5 and 20 cm from the midpoint between tree pairs) were collected from 52 plots across the richness gradient. Soils were homogenized via 2 mm sieving. Randomized daily sampling minimized spatio-temporal autocorrelation. Soil chemistry: Soil moisture measured by drying 25 g at 40°C for two days. Soil pH measured in 1:2.5 soil-water suspensions. For TOC, TN, TP, 200 g soil was ground and sieved (0.25 mm). TOC measured by Liqui TOC II analyzer; TN by Kjeldahl method on SEAL auto-analyzer; TP by wet digestion (H2SO4 and HClO4) and UV-VIS spectrophotometry (UV2700). Ratios TOC:TN and TOC:TP (C:N, C:P) computed. Microbial biomass: PLFA analysis on 5 g frozen soil (Frostegård et al.) with biomarkers assigned to functional groups (Ruess et al.). Total microbial biomass equals the sum of all PLFA group biomasses; bacteria:fungi (B:F) ratio computed as bacterial over fungal biomarkers. Active microbial biomass measured using substrate-induced respiration (SIR) method on 6 g soil (Scheu 1992). Taxonomic profile: DNA extracted (PowerSoil kit). DNA quality quantified (NanoDrop), adjusted to 10–15 ng/µl. Amplicon libraries for bacteria and fungi prepared per Schöps et al. and Nawaz et al. Reads processed with QIIME 2 (v2020.2): demultiplexing, primer trimming, chimera removal (cutadapt), denoising/ASV inference and non-target removal (DADA2). ASV tables imported to R (phyloseq). Rarefaction to 28,897 (bacteria) and 16,542 (fungi) reads per sample. Diversity metrics (OTU richness, Shannon diversity, Pielou evenness, Gini dominance) computed (microbiome R package); Shannon diversity used for analyses. Functional profile: DNA extracted (FastDNA Spin Kit for Soil). Concentrations measured by NanoDrop and QuantiFluor dsDNA (SpectraMax M5). DNA diluted to 50 ng/µl and stored at −20°C. Functional genes involved in carbon catabolism quantified using high-throughput qPCR-based SmartChip (QMEC; Zheng et al.). Gene abundances z-scaled across samples and summed for carbon catabolism gene set (Cata). Functional gene Pielou evenness (FG evenness) computed (vegan::diversity). Physiological potential: MicroResp method applied with 14 substrates spanning saccharides, amino acids, and carboxylic acids covering gradients of molecular weights (89–221 g mol−1) and carbon oxidation states. CO2 production used to calculate substrate-induced respiration efficiency (SIR efficiency = Pielou evenness of CO2 across substrates) and SIR range (difference in CO2 between oxalic acid and alanine). Sensitivity analyses showed robustness to substrate selection. Microbial respiration: Basal microbial respiration measured on 6 g fresh soil using O2-microcompensation method (Scheu 1992) without substrate or water addition (reflecting in situ respiration). Active microbial biomass and respiration measured on the same instrument/sample; analyses repeated with/without active biomass to test robustness. Statistical analyses: All analyses in R 4.0.3. Variables standardized (centered and divided by two SDs) using arm::rescale. Tree diversity effects tested via linear models (lm), evaluating linear vs non-linear fits (AIC-based model selection). Model assumptions checked (performance::model_check). Correlations among microbial facets assessed via Pearson correlation. Multivariate linear models with AIC-based stepwise selection (stats::step) assessed effects of microbial biomass (total, active), taxonomic profile (B:F, bacterial and fungal Shannon), and functional profile (Cata, FG evenness) on SIR efficiency, SIR range, and microbial respiration. Variance partitioning among facets with vegan::varpart. Structural equation modeling (SEM) with lavaan::sem modeled causal paths among facets, physiological potential, and respiration; model fit evaluated (RMSEA < 0.10, CFI > 0.9, SRMR < 0.08). Extended SEM included exogenous drivers (tree species richness, TOC, TN, TP, pH, relative humidity) with total effects computed as sums of absolute standardized path coefficients. Full outputs in Supplementary materials.

- Tree species richness significantly increased total microbial biomass (estimate ± SE = 0.020 ± 0.007, p = 0.003), bacterial Shannon diversity (0.017 ± 0.007, p = 0.011), and SIR efficiency (0.022 ± 0.007, p = 0.001); microbial respiration showed a positive trend (0.013 ± 0.007, p = 0.064). Across monocultures to 24-species plots, microbial biomass increased by ~25%, bacterial diversity by ~12%, and physiological potential by ~12%. - Microbial community facets were intercorrelated: total and active biomass (r = 0.45, p < 0.001); functional profile variables (Cata and FG evenness) (r = 0.57, p < 0.001). B:F ratio negatively correlated with microbial biomass and fungal diversity; fungal diversity positively correlated with active biomass (r = 0.20, p = 0.014). - Microbial facets explained up to 50% of microbial respiration variance, but only 19% (SIR efficiency) and 4% (SIR range). Microbial biomass was the dominant driver (explaining up to 43% of respiration, 14% of SIR efficiency, 2% of SIR range), whereas taxonomic and functional profiles explained little (respiration: 6% and <1% respectively; SIR efficiency: 1% and 2%; SIR range: ~1%). - SEM including physiological potential improved explained variance of respiration from 50% to 57%. Significant positive effects on microbial respiration: active biomass (0.590 ± 0.060, p < 0.001), fungal diversity (0.128 ± 0.058, p = 0.027), SIR efficiency (0.176 ± 0.062, p = 0.005), and SIR range (0.213 ± 0.057, p < 0.001). SIR efficiency increased with total biomass (0.209 ± 0.083, p = 0.012) and active biomass (0.258 ± 0.082, p = 0.002), and decreased with functional gene evenness (−0.179 ± 0.089, p = 0.045). - Total effect sizes on respiration: microbial biomass 0.672 (direct 0.590, indirect 0.082); taxonomic profile 0.128 (direct); functional profile 0.031 (indirect); physiological potential 0.389 (direct). - Adding tree diversity and soil chemistry to SEM increased respiration R² to 68% and biomass R² to 46%. Soil chemical properties strongly affected microbial biomass (total effect 1.474), physiological potential (0.799), respiration (0.312), and taxonomic profile (0.199); no net effect detected on functional profile. TOC was the dominant soil driver (total effect 1.383). Specific paths included: TOC → biomass (0.65, p < 0.001) and active biomass (0.41, p < 0.001); RH → respiration (0.33, p < 0.001) but RH → biomass (−0.23, p < 0.001); pH → fungal diversity (0.20, p < 0.05) and pH → SIR efficiency (−0.21, p < 0.05); C:P → SIR efficiency (0.26, p < 0.01); TOC → SIR efficiency (0.27, p < 0.001) and TOC → SIR range (−0.23, p < 0.01); B:F → respiration (−0.14, p < 0.01); bacterial diversity → SIR range (0.21, p < 0.01). - Tree species richness increased total biomass (0.173 ± 0.063, p = 0.006), bacterial diversity (0.164 ± 0.082, p = 0.045), and SIR efficiency (0.152 ± 0.073, p = 0.038), indirectly increasing respiration (indirect effect ≈ 0.014). Overall, microbial biomass, not diversity, primarily mediates the tree diversity effect on respiration.

The findings demonstrate that tree species richness enhances multiple microbial community facets and functions, supporting H1. By integrating biomass, taxonomic, and functional profiles, the study shows these facets are interrelated (H2) and jointly influence microbial physiological potential and respiration. Microbial biomass and physiological potential emerged as the principal determinants of respiration (supporting H3), whereas taxonomic and functional diversity exerted comparatively minor effects. Physiological potential mediates part of the biomass effect and is influenced by biomass and functional gene evenness, highlighting the importance of microbial physiology in linking community composition to realized functions. Soil chemical properties—particularly TOC, with additional roles of pH and moisture—strongly structure microbial community facets and their linkages to respiration (supporting H4). TOC boosts biomass and physiological potential, ultimately enhancing respiration. Tree diversity indirectly increases respiration via increases in biomass and physiological potential, indicating that diversity-driven increases in plant carbon inputs and substrate heterogeneity elevate microbial capacity and activity. These insights suggest that models of soil carbon dynamics should prioritize microbial biomass, physiological potential, and key soil chemical drivers (TOC, pH, moisture) to predict microbial respiration and carbon cycling outcomes.

Tree diversity and soil carbon content are key drivers of soil microbial respiration through their effects on microbial biomass and physiological potential. Across a comprehensive framework linking microbial community facets with environmental drivers, microbial biomass consistently outperformed diversity metrics in predicting respiration. The study suggests a positive feedback whereby tree diversity enhances soil carbon content via increased microbial biomass and functioning, with implications for soil carbon storage and climate mitigation. Future research and models of soil carbon dynamics should explicitly incorporate microbial biomass, physiological potential, soil carbon pools, pH, and moisture, and consider reforestation strategies that enhance diversity to optimize belowground carbon processes.

- Physiological potential measured by MicroResp reflects responses under substrate-rich conditions and may not fully represent in situ respiration dynamics where substrate availability and pathway activation differ. - The study relies on cross-sectional sampling; temporal dynamics and causality beyond SEM inference are not directly tested. Dynamic and mechanistic modeling calibrated with temporal data are needed. - Functional profiling focused on carbon-catabolism-related genes via QMEC; other functional pathways and enzyme activities were not directly measured and may influence respiration. - Soil carbon chemical pool characterization was coarse (bulk TOC); more detailed fractionation could refine mechanisms linking TOC quality to microbial processes. - Context dependency (e.g., climate variability, microclimate, unmeasured edaphic factors) may limit generalizability beyond the subtropical forest experiment.