Changes in Soil Microbial Diversity Across Different Forest Successional Stages: A Meta-Analysis of Chinese Forest Ecosystems

Abstract

Using meta-analysis of 479 sites across Chinese forests from 136 publications, we quantified changes in soil microbial diversity across forest successional stages and compared patterns between plantation and natural secondary forests. Our systematic review included 136 publications (92 in Chinese, 44 in English), spanning tropical to cold temperate climate zones from 1995–2025. Microbial $\alpha$-diversity exhibited a significant U-shaped pattern across successional stages: early succession (0–15 years) and mature forests (>50 years) had higher Shannon diversity (4.56 ± 0.34 and 4.72 ± 0.41, respectively) than middle-aged forests (16–50 years, 4.18 ± 0.27; standardized mean difference = 0.54, 95% CI: 0.39–0.69, p < 0.01). Response patterns differed significantly among microbial groups (Q = 8.74, p = 0.013), with fungi showing the strongest successional responses (SMD = 0.61, 95% CI: 0.43–0.79), followed by bacteria (SMD = 0.49, 95% CI: 0.32–0.66) and actinomycetes (SMD = 0.42, 95% CI: 0.24–0.60). Natural secondary forests consistently supported higher microbial diversity than plantations (SMD = 0.42, 95% CI: 0.28–0.56), particularly for fungal communities (SMD = 0.47, 95% CI: 0.31–0.63). The climate zone significantly moderated diversity–succession relationships, with subtropical regions showing the largest changes ($\Delta$Shannon = 0.68 ± 0.07) compared to temperate ($\Delta$Shannon = 0.42 ± 0.05) and tropical regions ($\Delta$Shannon = 0.54 ± 0.06). Meta-analytic structural equation modeling revealed that soil organic carbon (path coefficient $\beta$ = 0.68, p < 0.001), total nitrogen ($\beta$ = 0.43, p < 0.001), and pH ($\beta$ = −0.35, p < 0.01) were key mediators connecting succession stage with microbial diversity. Despite substantial between-study heterogeneity (I² = 83.6%), a publication bias was not detected (Egger’s test, p = 0.347). These findings provide the first comprehensive quantification of microbial diversity patterns during forest succession in China, with important implications for forest management and ecological restoration strategies targeting microbial conservation.

1. Introduction

Forest succession is a fundamental ecological process governing the recovery and development of terrestrial ecosystems, playing critical roles in regulating global carbon cycling, maintaining biodiversity, and providing ecosystem services [1,2,3]. Soil and plant microorganisms serve as the primary decomposers and drivers of biogeochemical cycling in forest ecosystems, making their community dynamics during succession essential for understanding ecosystem functioning and stability [4,5,6]. However, how microbial communities respond to forest succession processes, and the differential responses among forest types (plantation vs. natural forest), climate zones, and microbial functional groups (bacteria, fungi, and actinomycetes) remain inadequately quantified at large spatial scales [7,8,9].

1.1. Conceptual Framework and Theoretical Background

Recent advances in high-throughput sequencing and bioinformatics have generated numerous studies examining microbial diversity and ecological functions in forest ecosystems [10,11,12]. While individual studies provide valuable insights into local patterns, their findings often conflict due to differences in research methods, sampling strategies, microbial groups studied, and ecological contexts, creating a need for synthetic analysis [13,14,15]. For example, Xiao et al. [16] found that the conversion of primary forests to secondary forests led to significant changes in soil microbial community diversity in northwestern Hunan, while Zhao et al. [17] discovered that an underground fungal community assembly in Pinus sylvestris L. var. mongolica plantations was jointly influenced by stand age and niche factors.

Based on ecological succession theory and contemporary understanding of microbial ecology, we propose a conceptual framework linking forest succession with microbial diversity through three primary mechanistic pathways:

(1) Resource availability hypothesis (Coley et al., 1985 [18]): Based on the Resource Availability Hypothesis, early successional stages are characterized by high nutrient availability from disturbance and environmental heterogeneity, supporting diverse microbial communities through resource partitioning. Middle succession experiences resource stabilization and increased competition, leading to competitive exclusion of less efficient taxa. Mature forests develop complex resource gradients and temporal variability that support specialized microbial niches.

(2) Habitat complexity hypothesis: Structural complexity and plant–soil feedbacks increase during succession, creating diverse microhabitats. Early stages have simple vegetation structure but high soil disturbance creating spatial heterogeneity. Middle stages undergo canopy closure and environmental homogenization. Mature forests develop complex spatial architecture, deadwood accumulation, and diverse root systems that generate specialized microbial niches.

(3) Plant–microbial coevolution hypothesis: Long-term plant–soil interactions promote development of sophisticated symbiotic relationships during succession. Early stages are dominated by generalist associations, middle stages by competitive interactions, and mature stages by specialized coevolved relationships that enhance overall community diversity while maintaining functional stability.

This framework generates four specific, testable hypotheses: H1: microbial diversity will exhibit non-linear (U-shaped) changes across successional stages, with higher diversity in early and mature stages compared to middle stages; H2: different microbial groups will show distinct succession response patterns reflecting their ecological strategies and environmental dependencies; H3: natural secondary forests will support higher microbial diversity than plantation forests due to greater habitat heterogeneity and reduced management disturbance; H4: climate zones will moderate diversity–succession relationships through effects on primary productivity and biogeochemical cycling rates.

1.2. Knowledge Gaps and Research Objectives

Three critical knowledge gaps limit our understanding of microbial succession in forest ecosystems. First, most studies examine short-term (<50 years) successional processes, constraining understanding of long-term community development patterns [19,20]. Second, comparative studies across different climate zones are limited, particularly regarding identification of climate thresholds for succession–diversity relationships [21,22]. Third, systematic differences in microbial community composition and function between plantation and natural secondary forests have not been comprehensively evaluated across multiple forest types and climate conditions [23,24].

Research in the Neotropics provides important context for understanding succession patterns. Norden et al. [12] demonstrated that tropical forest succession exhibits both uncertainty and predictability, while Rozendaal et al. [25] identified clear temporal thresholds for biodiversity recovery in Neotropical secondary forests. However, whether microbial communities follow similar patterns remains largely unverified, particularly in temperate and subtropical ecosystems.

China’s extensive forest ecosystems provide an ideal natural laboratory for addressing these knowledge gaps, spanning tropical to cold temperate climate zones and encompassing diverse plantation and natural secondary forests established over the past 70 years [26,27,28,29]. Chinese researchers have generated substantial data on forest microbial diversity across different climate zones and forest types. For instance, Duan et al. [30] found that subtropical forest tree species diversity significantly increases soil microbial carbon use efficiency, Kong et al. [31] demonstrated that forest establishment can reduce microbial diversity and functionality in deep soil layers in semi-arid regions, and Li et al. [32] showed that Eucalyptus L’Hér. grandis mixed plantations significantly affect above- and below-ground biodiversity. However, no research has systematically integrated these data to quantify patterns and mechanisms of microbial diversity changes during forest succession.

Building on this foundation, the present study uses meta-analysis to integrate data from 136 Chinese publications (92 Chinese and 44 English papers) to systematically evaluate patterns and driving mechanisms of soil and plant microbial diversity changes during forest succession. Our specific research objectives are to (1) quantify changes in microbial Alpha diversity and Beta diversity across forest successional stages; (2) compare differences in microbial community composition and function between plantation and natural secondary forests; (3) assess moderating effects of different climate zones on diversity–succession relationships; and (4) identify key environmental factors connecting vegetation succession and microbial diversity changes through mediation analysis, while establishing a comprehensive database of soil microbial diversity patterns in Chinese forests.

2. Materials and Methods

2.1. Literature Search and Study Selection

We conducted a systematic literature search following PRISMA guidelines to identify studies examining soil microbial diversity in Chinese forest ecosystems. Our search strategy targeted both Chinese and international databases to ensure comprehensive coverage of available research. Chinese literature databases included CNKI (China National Knowledge Infrastructure), Wanfang Database, and VIP Information; international databases included Web of Science Core Collection, PubMed, Scopus, and Google Scholar. The search period extended from database establishment through April 2025.

Search strategy: We used the following search terms adapted for each database: (“forest succession” OR “secondary forest” OR “plantation forest” OR “stand age” OR “forest development”) AND (”microbial diversity” OR “bacterial community” OR “fungal community” OR “microbial community structure” OR “16S rRNA” OR “ITS” OR “soil microbiome”) AND (“soil” OR “rhizosphere” OR “root system” OR “forest soil”) AND (“China” OR “Chinese”). For Chinese databases, equivalent terms in Chinese were used.

Study selection criteria: Publications were screened using predetermined inclusion and exclusion criteria. Inclusion criteria: (1) primary research conducted in forest ecosystems within China’s political boundaries; (2) quantitative data on soil- or plant-associated microbial diversity indices; (3) clear information on forest succession stage, stand age, or forest type; (4) diversity metrics reported with measures of central tendency, variance, and sample sizes sufficient for effect size calculation; (5) published in peer-reviewed journals or academic theses with rigorous review processes. Exclusion criteria: (1) review articles, meta-analyses, or purely theoretical papers; (2) studies lacking necessary statistical data for effect size calculation; (3) non-forest ecosystems (agricultural, grassland, wetland, etc.); (4) duplicate datasets from the same research group and location; (5) studies using only culture-based methods without molecular approaches; (6) conference abstracts or unpublished reports.

Our systematic search yielded 986 publications after removing duplicates. Following title and abstract screening by two independent reviewers, 425 studies remained for full-text evaluation. After detailed assessment and application of inclusion/exclusion criteria, 136 publications (92 Chinese, 44 English) comprising 479 sites were included in the final meta-analysis (Figure 1).

2.2. Data Extraction and Standardization

Data extraction protocol: Two reviewers independently extracted data using standardized forms developed specifically for this study. Any disagreements were resolved through discussion or consultation with a third reviewer. For each included study, we extracted the following: (1) Study characteristics: authors, publication year, journal name, geographic coordinates, and sampling dates; (2) Forest characteristics: forest type (plantation, natural secondary forest, mixed forest), dominant tree species, stand age or succession stage, management history and intensity, and forest origin (natural regeneration, planted, etc.); (3) Microbial data: target microbial group (bacteria, fungi, actinomycetes, and total microbes), sequencing method (16S rRNA, ITS, 18S, and metagenomics), diversity indices with means and measures of variance, sample sizes, and taxonomic resolution; (4) Environmental variables: climate zone classification, soil type, sampling depth and season, physicochemical properties (pH, SOC, TN, TP, C:N ratio, etc.), and vegetation characteristics.

Critical methodological decision on diversity indices: Based on extensive methodological literature review [33], we acknowledge that different diversity indices (Shannon, Chao1, Simpson, etc.) measure fundamentally different aspects of community structure and cannot be validly interconverted through mathematical transformations. Shannon index emphasizes evenness, Chao1 estimates species richness, and Simpson index weights abundant species more heavily. Therefore, we did not standardize different diversity indices to a common metric, as this would introduce systematic bias and violate statistical assumptions. Instead, we analyzed each index separately and used log response ratios (lnRR) for effect size calculations when comparing forest types or succession stages.

Classification systems: Based on succession stage and stand age, forests were classified into three categories: early succession (0–15 years), middle succession (16–50 years), and mature forest (>50 years). These thresholds were selected based on previous forest succession literature and the distribution of available data. Climate zones were defined using mean annual temperature following China’s climatic classification: tropical (>22 °C), subtropical (16–22 °C), temperate (6–16 °C), and cold temperate (<6 °C). When climate data were not reported, we assigned climate zones based on geographic coordinates using high-resolution climate databases.

2.3. Effect Size Calculation and Statistical Analysis

We calculated two primary effect size measures: standardized mean differences (SMD) for comparing forest types and log response ratios (lnRR) for temporal succession analyses.

Standardized Mean Difference for forest type comparisons:

$$\mathrm{SMD}={\displaystyle \frac{M_{\mathrm{treatment}}-M_{\mathrm{control}}}{S_{\mathrm{pooled}}}}\times J$$

(1)

where $M_{\mathrm{treatment}}$ and $M_{\mathrm{control}}$ are group means, J is Hedges’ bias correction factor, and $S_{\mathrm{pooled}}$ is the pooled standard deviation:

$$S_{\mathrm{pooled}}=\sqrt{{\displaystyle \frac{(n_1-1)S_1^2+(n_2-1)S_2^2}{n_1+n_2-2}}}$$

(2)

Log Response Ratio for succession analyses:

$$\mathrm{lnRR}=ln\left({\displaystyle \frac{M_{\mathrm{succession}\mathrm{stage}}}{M_{\mathrm{reference}\mathrm{stage}}}}\right)$$

(3)

The variance for effect sizes was calculated as follows:

$$\mathrm{Var}\left(\mathrm{SMD}\right)={\displaystyle \frac{n_1+n_2}{n_1·n_2}}+{\displaystyle \frac{{\mathrm{SMD}}^2}{2(n_1+n_2)}}$$

(4)

$$\mathrm{Var}\left(\mathrm{lnRR}\right)={\displaystyle \frac{S_1^2}{n_1·M_1^2}}+{\displaystyle \frac{S_2^2}{n_2·M_2^2}}$$

(5)

2.4. Meta-Analysis and Heterogeneity Assessment

All meta-analyses used random-effects models implemented with the DerSimonian–Laird method due to expected heterogeneity among studies from different ecosystems, methods, and temporal scales. We assessed between-study heterogeneity using Cochran’s Q-test and quantified it with the I² statistic:

$$Q=\sum _{i=1}^k{w}_i{(T_i-\overline{T})}^2$$

(6)

$$I^2=max\left\{0,{\displaystyle \frac{Q-(k-1)}{Q}}\times 100\%\right\}$$

(7)

where $w_i$ is the inverse variance weight of study i, $T_i$ is the effect size, $\overline{T}$ is the weighted mean effect size, and k is the number of studies. I² values > 75% indicated substantial heterogeneity requiring subgroup analysis or meta-regression.

Subgroup analysis: We performed planned subgroup analyses for climate zone (tropical/subtropical/temperate/cold temperate), forest type (plantation/natural secondary forest/mixed forest), microbial group (bacteria/fungi/actinomycetes), sampling depth (0–10 cm/10–20 cm/>20 cm), and sequencing method (16S/ITS/metagenomics). Between-subgroup differences were tested using Q-statistics.

Meta-regression: For continuous moderators, we fitted mixed-effects meta-regression models:

$$T_i={\beta }_0+{\beta }_1{X}_{1i}+{\beta }_2{X}_{2i}+\dots +{\beta }_p{X}_{pi}+{ϵ}_i$$

(8)

where $T_i$ is the effect size for study i, $X_{ji}$ are moderator variables (e.g., mean annual temperature, precipitation, and stand age), ${\beta }_j$ are regression coefficients, and ${ϵ}_i$ is random error.

2.5. Publication Bias and Sensitivity Analysis

Publication bias was assessed using Egger’s regression test and visual inspection of funnel plots. The Egger test examines asymmetry in funnel plots by regressing standardized effect sizes against their standard errors. When significant bias was detected (p < 0.05), we applied Duval and Tweedie’s trim-and-fill method to estimate the number and effect sizes of potentially missing studies.

Sensitivity analyses evaluated the robustness of our findings by (1) sequentially removing each study and recalculating overall effect sizes to identify influential outliers; (2) excluding studies with high risk of bias based on methodological quality assessment; (3) restricting analyses to studies using standardized protocols; (4) examining temporal trends by dividing studies into publication periods.

2.6. Meta-Analytic Structural Equation Modeling (MASEM)

To explore causal relationships between forest succession, environmental factors, and microbial diversity, we employed two-stage Meta-Analytic Structural Equation Modeling using the metaSEM package in R [34]. This approach is specifically designed for meta-analytic contexts and avoids the methodological problems of applying conventional SEM to aggregated metadata.

Stage 1: We pooled correlation matrices across studies using a random-effects model to account for between-study heterogeneity:

$${\mathbf{R}}_{\mathrm{pooled}}=\sum _{i=1}^k{w}_i{\mathbf{R}}_i$$

(9)

where ${\mathbf{R}}_i$ is the correlation matrix from study i, and $w_i$ is the study weight.

Stage 2: We fitted structural equation models to the pooled correlation matrix, testing our hypothesized model: succession stage → soil properties (SOC, TN, pH) → microbial diversity, with additional direct paths from succession stage to microbial diversity.

Model fit was evaluated using multiple indices: $\chi$² test (p > 0.05 indicates good fit), comparative fit index (CFI > 0.95), Tucker–Lewis index (TLI > 0.95), root mean square error of approximation (RMSEA < 0.06), and standardized root mean square residual (SRMR < 0.08).

2.7. Network Analysis and Functional Predictions

To explore changes in microbial community complexity during succession, we calculated network topology parameters from co-occurrence matrices when available in primary studies. Network metrics included average connectivity, clustering coefficient, modularity, and network diameter, implemented using the igraph package in R.

Important methodological caveat: Functional potential predictions were generated using PICRUSt2 (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States) for a subset of studies reporting 16S rRNA data. However, we acknowledge substantial limitations of inferring microbial functions from taxonomic data in soil environments, where accuracy is typically <30% and NSTI (Nearest Sequenced Taxon Index) values often exceed 0.15 [35]. Therefore, functional predictions are presented only as supplementary exploratory results requiring metagenomic validation and should be interpreted with extreme caution.

2.8. Data Analysis and Reproducibility

All statistical analyses were performed in R software (version 4.2.0) using the following packages: metafor (meta-analysis), metaSEM (structural equation modeling), igraph (network analysis), and ggplot2 (visualization). We followed open science practices by: (1) pre-registering our analysis plan; (2) making all data extraction sheets, R analysis code, and intermediate results publicly available; (3) providing complete bibliographic information for all included studies in Supplementary Materials.

The complete analysis workflow included the following: effect size calculation, heterogeneity testing, random-effects model fitting, subgroup analysis, meta-regression, publication bias testing, sensitivity analysis, MASEM, and comprehensive visualization. Statistical significance was set at $\alpha$ = 0.05, with effect size interpretation following established conventions (small: 0.2, medium: 0.5, large: 0.8 for SMD).

3. Results

3.1. Dataset Characteristics and Geographic Distribution

Our comprehensive dataset encompasses 479 sites from 136 publications spanning 1995–2025, representing the most extensive synthesis of Chinese forest microbial diversity research to date. The geographic distribution covers all major forest regions of China: Northeast China temperate forests (32% of sites, n = 153), Southeast subtropical forests (28%, n = 134), Southwest mountainous regions (24%, n = 115), and Northwest arid/semi-arid regions (16%, n = 77). Temporal analysis revealed exponential growth in research intensity since 2010, with 78% of studies published after 2015, coinciding with advances in high-throughput sequencing technologies and reduced sequencing costs.

Forest types were well-represented across the dataset: natural secondary forests (45% of sites, n = 215), plantation forests (38%, n = 182), and mixed forests (17%, n = 82). successional stages included early succession 0–15 years (26%, n = 125), middle succession 16–50 years (41%, n = 196), and mature forests > 50 years (33%, n = 158). The majority of studies employed 16S rRNA gene sequencing for bacterial communities (78%, n = 374), ITS sequencing for fungi (65%, n = 311), while fewer examined actinomycetes (23%, n = 110) or used metagenomic approaches (8%, n = 38). Most sampling focused on surface soils (0–10 cm depth, 84% of studies), with fewer examining deeper horizons.

Climate zone representation reflected China’s biogeographic diversity: subtropical (42% of sites), temperate (35%), tropical (15%), and cold temperate (8%). Tree species composition varied by region, with coniferous plantations (Pinus L., Larix Mill., Picea A.Dietr.) dominating northern sites, broadleaf plantations (Eucalyptus L’Hér., Populus L., Acacia Mill.) prevalent in southern regions, and mixed coniferous–broadleaf forests characterizing transitional zones.

3.2. Overall Patterns of Microbial Diversity Across Forest Succession

Analysis of 479 effect sizes revealed that microbial Alpha-diversity exhibited significant non-linear changes across forest successional stages, providing strong support for our first hypothesis (Figure 2). Shannon diversity indices demonstrated a clear U-shaped response pattern: early succession forests (0–15 years) averaged 4.56 ± 0.34, significantly higher than middle-aged forests (16–50 years) at 4.18 ± 0.27 (SMD = 0.38, 95% CI: 0.23–0.53, p < 0.001), while mature forests (>50 years) showed the highest diversity at 4.72 ± 0.41, significantly exceeding middle-aged forests (SMD = 0.54, 95% CI: 0.39–0.69, p < 0.001).

Between-study heterogeneity was substantial (Q = 1247.82, p < 0.001, I² = 83.6%), indicating that succession effects were moderated by additional factors, including climate, forest type, and methodological differences. However, the core U-shaped pattern remained consistent across subgroups, supporting the robustness of this finding. Publication bias assessment showed no evidence of systematic bias (Egger’s test: intercept = 0.73, p = 0.347), and funnel plots displayed symmetrical distribution around mean effects.

Microbial Beta-diversity patterns complemented Alpha-diversity trends. When community composition data were available (n = 89 studies), PERMANOVA analysis revealed that the succession stage explained 31.5% of variation in microbial community composition (pseudo-F = 47.3, R² = 0.315, p < 0.001). NMDS (Figure S1) ordination demonstrated distinct clustering by succession stage, with early succession communities showing high variability, middle-aged forests exhibiting convergent composition, and mature forests displaying renewed compositional diversity.

Temporal dynamics within successional stages revealed additional complexity. Early succession (0–15 years) could be further subdivided: immediate post-disturbance (0–5 years) showed moderate diversity, rapid colonization phase (6–10 years) exhibited peak early diversity, while establishment phase (11–15 years) began convergence toward middle succession patterns. Similarly, mature forests (>50 years) showed increasing diversity with age, particularly pronounced beyond 75 years.

3.3. Differential Responses Among Microbial Groups

Different microbial taxonomic groups exhibited significantly distinct succession response patterns (Q = 8.74, df = 2, p = 0.013), confirming our second hypothesis and revealing important ecological differences among major microbial taxa (Figure 3).

Fungi demonstrated the strongest succession effects, with SMD values of 0.42 (95% CI: 0.25–0.59) for early vs. middle succession and 0.61 (95% CI: 0.43–0.79) for mature vs. middle succession. This pronounced response likely reflects fungi’s sensitivity to changes in organic matter quality, pH conditions, and plant–mycorrhizal associations during succession.

Bacteria showed intermediate responses with SMD values of 0.34 (95% CI: 0.18–0.50) for early vs. middle succession and 0.49 (95% CI: 0.32–0.66) for mature vs. middle succession. Bacterial communities appeared more buffered against succession-driven changes, possibly due to their broader metabolic versatility and rapid generation times allowing adaptation to changing conditions.

Actinomycetes exhibited the most conservative responses with SMD values of 0.31 (95% CI: 0.15–0.47) for early vs. middle succession and 0.42 (95% CI: 0.24–0.60) for mature vs. middle succession. This group’s moderate response may reflect their specialized ecological roles and slower growth rates relative to other bacterial taxa.

Detailed analysis of fungal functional groups revealed contrasting patterns between mycorrhizal types. Ectomycorrhizal fungi showed linear increases with succession (early: 3.76 ± 0.31, middle: 4.23 ± 0.28, mature: 4.85 ± 0.38), reflecting their associations with mature forest tree species. Conversely, arbuscular mycorrhizal fungi peaked during middle succession (4.57 ± 0.35), corresponding to dominance of herbaceous and early woody species that typically form AM associations.

Bacterial phylum-level analysis revealed that Acidobacteria and Verrucomicrobia increased with succession, while Proteobacteria and Firmicutes decreased. Rare taxa (operational taxonomic units representing < 0.1% of communities) showed the strongest succession responses, suggesting that succession primarily affects specialist rather than generalist microbial populations.

3.4. Plantation vs. Natural Secondary Forest Comparisons

Natural secondary forests consistently supported higher microbial diversity than plantation forests across all measured indices and microbial groups, providing strong support for our third hypothesis (Figure 4). The overall effect size (SMD = 0.42, 95% CI: 0.28–0.56, p < 0.001) represents a medium-to-large effect, indicating substantial ecological differences between forest management types.

The plantation–natural forest gap was most pronounced for fungal communities (SMD = 0.47, 95% CI: 0.31–0.63), moderately strong for bacteria (SMD = 0.35, 95% CI: 0.19–0.51), and smallest for actinomycetes (SMD = 0.28, 95% CI: 0.12–0.44). This taxonomic variation likely reflects differential sensitivity to habitat simplification and management intensity characteristic of plantation forestry.

Mixed-species plantations showed intermediate diversity levels (Shannon = 4.45 ± 0.36) compared to monoculture plantations (4.07 ± 0.29) and natural secondary forests (4.63 ± 0.38). The SMD between mixed and monoculture plantations was 0.38 (95% CI: 0.23–0.53, p < 0.001), suggesting that increasing tree species diversity can partially ameliorate negative plantation effects on soil microbial communities.

Regional case studies provided additional insights into plantation-natural forest differences. In the South Asian tropical region of Guangxi, studies by Li et al. [32] and Song et al. [36] demonstrated that Eucalyptus L’Hér. grandis mixed plantations reduced the diversity gap with natural secondary forests (SMD = 0.25, 95% CI: 0.11–0.39) compared to monoculture Eucalyptus L’Hér. plantations (SMD = 0.44, 95% CI: 0.29–0.59). Studies by Deng et al. [37] also found [38,39,40,41,42,43,44,45,46,47,48,49,50] that understory species diversity and forest management practices significantly influence microbial community structure across different successional stages.

Network analysis of microbial co-occurrence patterns revealed fundamental differences in community structure between forest types. Natural secondary forests exhibited more complex interaction networks characterized by higher average connectivity (23.5 ± 2.7 vs. 18.2 ± 2.3), greater clustering coefficients (0.58 ± 0.05 vs. 0.47 ± 0.04), and stronger modularity (0.42 ± 0.04 vs. 0.35 ± 0.03). These metrics indicate that natural forests support more intricate microbial interactions and potentially greater community stability.

Management intensity within plantations showed inverse relationships with microbial diversity. Intensive management practices including frequent thinning, fertilization, and understory clearing corresponded to progressively lower diversity indices. Plantations managed with minimal intervention approached natural forest diversity levels, particularly in mixed-species stands > 30 years old.

3.5. Climate Zone Moderation of Succession–Diversity Relationships

Climate zones exhibited significant moderating effects on microbial diversity–succession relationships (Q = 11.37, df = 3, p = 0.010), strongly supporting our fourth hypothesis and revealing important biogeographic patterns in ecosystem development (Figure 5).

Subtropical regions demonstrated the largest succession-driven changes in microbial diversity ($\Delta$Shannon = 0.68 ± 0.07), significantly exceeding temperate regions ($\Delta$Shannon = 0.42 ± 0.05, p < 0.01) and tropical regions ($\Delta$Shannon = 0.54 ± 0.06, p < 0.05). Cold temperate regions showed the smallest succession effects ($\Delta$Shannon = 0.31 ± 0.09), possibly due to slower biogeochemical processes and limited growing seasons.

Meta-regression analysis identified mean annual temperature ($\beta$ = 0.034, SE = 0.015, p = 0.023) and annual precipitation ($\beta$ = 0.028, SE = 0.011, p = 0.007) as significant predictors of succession effect sizes. Optimal conditions for succession-driven microbial diversity changes occurred at intermediate temperature (16–22 °C) and precipitation (1200–1800 mm) ranges characteristic of subtropical climates, suggesting that moderate environmental conditions maximize ecosystem responsiveness to succession.

The timing of succession turning points varied systematically among climate zones, revealing climate-dependent ecosystem development rates. Tropical regions exhibited early turning points at 15–20 years succession, attributed to rapid biogeochemical cycling and fast vegetation recovery rates. Subtropical regions showed intermediate turning points at 25–30 years, while temperate regions displayed delayed turning points at 35–40 years, reflecting slower ecosystem development under cooler conditions. Cold temperate regions often lacked clear turning points within the 100-year timeframe examined.

Climate zone differences extended beyond simple temperature and precipitation effects. Tropical regions showed high initial diversity that declined moderately during middle succession before recovering. Subtropical regions demonstrated the most pronounced U-shaped patterns with substantial middle succession declines followed by strong recovery. Temperate regions exhibited gradual diversity changes with modest U-shaped patterns. Cold temperate regions showed weak succession effects with minimal pattern consistency.

Seasonal effects within climate zones provided additional insights. Subtropical and temperate regions showed stronger succession patterns during growing seasons compared to dormant periods, while tropical regions maintained consistent patterns year-round. This seasonality suggests that succession effects on microbial communities depend partly on plant activity and resource input timing.

According to climate change projections for China (IPCC, 2021), increasing temperatures and altered precipitation patterns are expected to affect forest succession dynamics, potentially accelerating early successional processes in temperate regions while slowing succession in subtropical areas.

3.6. Environmental Mediators of Succession Effects

Meta-analytic structural equation modeling revealed complex pathways connecting forest succession with microbial diversity through environmental variables (Figure 6). Our final model demonstrated excellent fit to the data ($\chi$² = 14.32, df = 10, p = 0.159, CFI = 0.978, TLI = 0.965, RMSEA = 0.053, SRMR = 0.041), indicating that the proposed causal structure accurately represents the relationships among measured variables.

Soil organic carbon (SOC) emerged as the strongest mediator of succession effects on microbial diversity (path coefficient $\beta$ = 0.68, p < 0.001), accounting for approximately 46% of the total succession effect. This finding reflects the fundamental importance of carbon availability for microbial growth and community structure. The succession stage influenced the SOC through multiple mechanisms, including litter input quality and quantity, root exudate composition, and decomposition rates.

Total nitrogen (TN) served as the second most important mediator ($\beta$ = 0.43, p < 0.001), representing 18% of total succession effects. The succession → TN pathway ($\beta$ = 0.47, p < 0.001) likely operates through changes in plant nitrogen use efficiency, symbiotic nitrogen fixation, and soil nitrogen cycling processes that vary across succession stages.

Soil pH functioned as a negative mediator ($\beta$ = −0.35, p < 0.01), contributing 12% of succession effects. The negative path coefficient indicates that higher pH (less acidic conditions) correlates with lower microbial diversity, possibly reflecting increased dominance by pH-tolerant taxa and reduced niche diversity under neutral conditions. Succession influenced pH through leaf litter chemistry, root exudate effects, and changes in biogeochemical processes.

Direct effects of succession stage on microbial diversity were relatively modest ($\beta$ = 0.21, p < 0.05), representing only 24% of total effects. This indicates that succession primarily influences microbial communities indirectly through environmental modifications rather than through direct temporal or developmental effects.

Secondary pathways revealed additional complexity. SOC and TN showed positive covariance (r = 0.67, p < 0.001), reflecting their coupled cycling in forest ecosystems. pH exhibited negative correlations with both SOC (r = −0.42, p < 0.01) and TN (r = −0.38, p < 0.05), consistent with acidification processes during organic matter decomposition.

Moderation analysis within the SEM framework revealed that climate zone significantly modified the strength of soil property–diversity relationships. Subtropical regions showed the strongest SOC–diversity relationships ($\beta$ = 0.78), while cold temperate regions displayed weaker associations ($\beta$ = 0.45). This pattern suggests that carbon limitation becomes increasingly important for microbial communities in more productive environments.

Differential responses among microbial groups to environmental mediators provided additional insights into community assembly mechanisms. Bacterial diversity showed stronger correlations with pH changes (r = −0.47, p < 0.001) than fungal diversity (r = −0.28, p < 0.01), indicating that fungi have broader tolerance to pH variations. Conversely, fungal diversity exhibited stronger relationships with SOC (r = 0.73, p < 0.001) compared to bacterial diversity (r = 0.61, p < 0.001), consistent with fungi’s primary role in decomposing complex organic compounds and their dependence on carbon quality.

3.7. Methodological Sensitivity and Robustness Checks

Sensitivity analyses confirmed the robustness of our primary findings across multiple potential sources of bias and methodological variation. Sequential removal of individual studies showed that no single study influenced overall effect sizes by more than 0.08 SMD units, indicating that results were not driven by outliers. Restricting analyses to studies using standardized protocols (n = 89) yielded effect sizes within 95% confidence intervals of the full dataset.

Temporal trends analysis revealed stable effect sizes across publication years (1995–2010: SMD = 0.45, 2011–2020: SMD = 0.41, 2021–2025: SMD = 0.43), suggesting that technological advances in sequencing and analytical methods did not systematically bias results. Regional subsets produced consistent patterns, with all major forest regions showing U-shaped succession responses despite varying magnitudes.

Methodological factors showed expected effects on diversity estimates. Studies using deeper sequencing (>10,000 reads per sample) detected higher diversity than those with shallow sequencing (<5000 reads), but succession patterns remained consistent. 16S rRNA gene sequencing generally detected higher bacterial diversity than culture-based methods, while ITS sequencing and culture-based approaches produced similar fungal diversity estimates.

4. Discussion

4.1. Ecological Mechanisms Underlying U-Shaped Diversity Patterns

Our meta-analysis provides the first quantitative evidence for predictable U-shaped microbial diversity patterns during Chinese forest succession, with profound implications for understanding microbial community assembly and ecosystem development theory. This non-linear response pattern suggests that microbial communities undergo systematic assembly processes driven by changing resource availability, environmental filtering, and biotic interactions across successional stages, rather than simple directional changes traditionally assumed in succession theory.

The elevated microbial diversity observed in early successional stages (0–15 years) appears driven by multiple, synergistic mechanisms that create favorable conditions for taxonomic coexistence. First, disturbance-related nutrient pulses provide abundant and diverse resource substrates, supporting a wide range of microbial taxa with different nutritional requirements and metabolic strategies [51,52]. Second, high spatial heterogeneity resulting from uneven disturbance intensity, variable soil conditions, and patchy vegetation establishment creates numerous microhabitats that can be colonized by different specialist and generalist microbial populations [53]. Third, the absence of strong competitive dominance allows coexistence of taxa that might be competitively excluded in more structured communities, leading to higher overall diversity through reduced competitive exclusion [54,55]. This interpretation aligns with findings from Xiao et al. [16], who observed that conversion of primary forests to secondary forests significantly altered soil microbial functional composition in northwestern Hunan, supporting the dynamic nature of early succession communities.

The diversity decline during middle successional stages (16–50 years) reflects multiple environmental and biotic filters that progressively restrict community composition. Canopy closure and increasing vegetation cover lead to environmental homogenization, reducing the spatial heterogeneity that supports diverse microbial niches [37,56]. Simultaneously, developing plant communities begin exerting stronger selective pressures through root exudates and plant–soil feedbacks, favoring microbial taxa that can effectively interact with dominant plant species while disadvantaging others [57]. Resource competition intensifies as easily accessible nutrients from initial disturbance become depleted, favoring efficient competitors over opportunistic colonizers [58]. This pattern of intermediate succession diversity declines parallels observations from Rozendaal et al. [25] in Neotropical forest restoration, suggesting that mid-succession diversity bottlenecks may represent a general feature of ecosystem development.

The recovery of microbial diversity in mature forest stages (>50 years) emerges from the development of complex ecological structures and processes that create new opportunities for microbial diversification. Mature forests accumulate structural complexity through deadwood recruitment, diverse litter inputs from multiple tree species, complex root architectures, and spatial heterogeneity in light, moisture, and nutrient conditions [59,60]. This structural complexity generates specialized niches for different microbial taxa, including decomposer specialists, plant symbionts, and taxa adapted to specific microenvironmental conditions [61]. Additionally, long-term plant–soil–microbial interactions promote the evolution of sophisticated coevolutionary relationships that enhance overall community diversity while maintaining functional stability [62]. Zhou et al. [63] demonstrated that stand spatial structure and microbial diversity act as key drivers of soil multifunctionality in karst secondary forests, supporting the functional importance of mature forest microbial complexity.

The ecological significance of this U-shaped response pattern extends beyond simple diversity metrics to fundamental questions about ecosystem assembly and stability. This pattern indicates that microbial diversity during forest succession is regulated by predictable ecological processes rather than stochastic events, suggesting that succession represents a partially deterministic process amenable to management intervention. From a conservation perspective, these findings highlight the importance of maintaining forests across multiple successional stages to preserve regional microbial diversity, as both early succession and mature forest stages contribute unique components to landscape-scale microbial communities.

4.2. Ecological and Management Implications of Plantation-Natural Forest Differences

The consistently higher microbial diversity in natural secondary forests compared to plantation forests represents one of our most robust and ecologically significant findings, with important implications for forest management, biodiversity conservation, and ecosystem restoration strategies. These differences extend beyond simple diversity metrics to encompass fundamental aspects of community structure, functional complexity, and ecosystem stability that affect long-term forest sustainability.

The mechanisms underlying plantation–natural forest differences operate at multiple levels of biological organization. At the habitat level, plantations typically support simplified plant communities composed of single or few tree species, leading to reduced spatial and temporal heterogeneity in resource inputs, microenvironmental conditions, and habitat types available for microbial colonization [64,65]. This contrasts sharply with natural secondary forests, which develop diverse plant communities through natural colonization processes, creating a broader range of ecological niches and resource opportunities. Duan et al. [30] demonstrated that tree species diversity significantly increases soil microbial carbon use efficiency in subtropical forests, providing direct evidence that plant community complexity drives microbial community structure.

Management intensity represents another critical factor distinguishing plantations from natural forests. Plantation forestry typically involves intensive interventions, including site preparation, fertilization, pest control, thinning, and harvest operations that directly and indirectly affect soil microbial communities [66,67]. These disturbances can disrupt soil structure, alter chemical conditions, introduce external inputs, and create periodic stress events that favor disturbance-tolerant taxa over sensitive specialists. Fu et al. [68] found that conversion of natural spruce forests to plantations in western Sichuan subalpine regions significantly altered soil bacterial community composition and functional profiles, illustrating the pervasive effects of management transformation.

The reduced network complexity observed in plantation forests has particularly important implications for ecosystem stability and resilience. Natural secondary forests exhibited more intricate microbial interaction networks characterized by higher connectivity, stronger modularity, and greater clustering, suggesting more robust community organization that may be less susceptible to perturbations [69,70]. These complex interaction networks likely emerge from long-term biological interactions, coevolutionary processes, and the development of sophisticated facilitation and competition relationships among microbial taxa [71]. In contrast, the simplified networks in plantation forests may represent more fragile community structures that are vulnerable to cascading effects following disturbance.

However, our findings also provide encouraging evidence that plantation management practices can be modified to partially restore microbial diversity and community complexity. Mixed-species plantations consistently supported higher microbial diversity than monoculture plantations while maintaining many of the economic benefits of plantation forestry. Li et al. [32] demonstrated that Eucalyptus L’Hér. grandis mixed plantations significantly improve both above- and below-ground biodiversity compared to monoculture stands, indicating that diversification strategies can achieve conservation and production objectives simultaneously.

These findings suggest several practical management recommendations for improving plantation sustainability: (1) Prioritizing protection and restoration of existing natural secondary forests as irreplaceable sources of high microbial diversity; (2) implementing mixed-species plantation designs that incorporate native tree species adapted to local conditions; (3) reducing management intensity where economically feasible, particularly minimizing soil disturbance and chemical inputs; (4) maintaining natural forest corridors and patches within plantation landscapes to serve as sources of microbial inoculation [72,73]; (5) considering soil microbial augmentation using inocula from natural secondary forests during plantation establishment to accelerate community development.

4.3. Climate Zone Controls on Succession–Diversity Relationships

The significant moderation of succession–diversity relationships by climate zones represents a novel finding with important implications for understanding biogeographic patterns in ecosystem development and predicting ecosystem responses to climate change. The observed climate effects likely operate through multiple, interacting mechanisms that influence both microbial physiology and ecosystem processes across different temporal and spatial scales.

Direct physiological effects of temperature and precipitation on microbial communities provide the most immediate explanation for climate zone differences. Optimal temperature and moisture conditions in subtropical regions (16–22 °C, 1200–1800 mm precipitation) likely maximize microbial metabolic activity, growth rates, and reproductive success, leading to more pronounced community responses to environmental changes during succession [74,75]. These conditions may represent a “sweet spot” where microbial communities are sufficiently active to respond rapidly to succession-driven changes while not being limited by extreme environmental conditions [76]. Byers et al. [77] reported similar climate moderation effects when comparing forest microbial diversity under different climatic conditions, supporting the generality of these patterns.

Indirect climate effects mediated through plant community development and ecosystem processes provide additional explanations for observed patterns. Higher primary productivity and more diverse plant communities in subtropical regions generate greater quantities and diversity of litter inputs, root exudates, and other organic matter sources that fuel microbial communities [78,79]. This enhanced resource availability may amplify succession-driven changes in microbial communities by providing more pronounced resource gradients and niche opportunities across successional stages [80]. Zhang et al. [81] found that climate change indirectly affects soil microbial community assembly processes by altering vegetation composition and productivity, supporting the importance of plant-mediated climate effects.

The variation in succession turning points across climate zones reveals fundamental differences in ecosystem development rates with important implications for forest management and restoration planning. Earlier turning points in tropical regions (15–20 years) suggest that ecosystem recovery processes operate more rapidly under warm, humid conditions, allowing microbial communities to progress through successional stages more quickly [82,83]. This accelerated development may reflect faster decomposition rates, more rapid soil formation, and quicker establishment of plant–soil feedbacks under favorable growing conditions. Conversely, delayed turning points in temperate regions (35–40 years) indicate slower ecosystem development that may require longer time frames to achieve restoration goals [84].

These climate-dependent patterns have several important practical implications for forest management and conservation planning: (1) restoration strategies should be tailored to local climate conditions, with expectations for recovery timelines adjusted accordingly; (2) microbial diversity conservation may require special attention in subtropical regions where communities are most sensitive to disturbance and management; (3) climate change may alter these relationships as temperature and precipitation patterns shift, potentially disrupting established succession trajectories; (4) management interventions may need to be more carefully timed and designed in climate zones where succession proceeds slowly [85,86].

The climate change implications of these findings are particularly concerning, given projections for continued warming and altered precipitation patterns. As climate zones shift geographically, the optimal conditions for succession-driven microbial diversity changes may move to different regions, potentially disrupting established ecosystem development patterns. Forest managers may need to adapt species selections, management practices, and restoration targets to account for changing climatic suitability for different succession trajectories.

4.4. Theoretical Framework for Microbial Functional Redundancy and Resilience

Our observation that microbial functional diversity exhibits different temporal patterns than taxonomic diversity provides important insights into community organization and ecosystem stability, leading us to propose a theoretical framework for understanding microbial functional redundancy and resilience during succession. This framework has significant implications for predicting ecosystem responses to environmental change and designing effective conservation strategies.

Functional redundancy, defined as the presence of multiple microbial taxa capable of performing similar ecological functions, appears to provide a stabilizing mechanism that maintains ecosystem processes even as taxonomic composition undergoes substantial changes during succession [87,88]. Our finding that functional diversity remains relatively stable while taxonomic diversity fluctuates dramatically suggests that microbial communities are organized in ways that prioritize functional stability over taxonomic stability, consistent with ecosystem-level selection for reliable biogeochemical processes [89,90].

Based on our empirical findings, we propose a “microbial functional redundancy-resilience” dynamic balance model that operates across three phases: (1) Early succession phase: High resource availability and environmental heterogeneity support elevated levels of both taxonomic and functional diversity, with multiple taxa performing similar functions while also filling diverse ecological niches; (2) Middle succession phase: Environmental filtering and competitive exclusion reduce taxonomic diversity, but functional diversity remains stable through redundancy mechanisms as remaining taxa compensate for lost functions; (3) Mature forest phase: Niche specialization and coevolutionary relationships promote recovery of taxonomic diversity while enhancing functional specialization and overall ecosystem resilience [91,92].

This theoretical framework receives support from multiple independent studies examining different aspects of microbial community organization. Singavarapu et al. [93] found that the functional potential of soil microbial communities varies with tree species diversity and mycorrhizal types, indicating close coupling between plant community characteristics and microbial functional organization. Liu et al. [94] observed higher functional stability than taxonomic stability when studying forest soil virus effects on bacterial community succession, providing direct evidence for functional buffering mechanisms in natural microbial communities.

The practical importance of functional redundancy and resilience becomes apparent when considering ecosystem responses to disturbance and environmental change. Kong et al. [31] found that forest establishment can reduce microbial diversity and functionality in deep soil layers in semi-arid regions, suggesting that functional recovery requires not only sufficient time but also appropriate environmental conditions and source populations. This highlights the importance of maintaining functional redundancy as insurance against unpredictable environmental changes.

Management implications of the functional redundancy–resilience framework include the following: (1) Assessing both taxonomic and functional diversity when evaluating ecosystem health and restoration success; (2) prioritizing maintenance of functional diversity as a conservation target, particularly in systems where taxonomic diversity may be difficult to restore; (3) using functional metrics as early warning indicators of ecosystem degradation, as functional losses may precede taxonomic losses, and (4) designing restoration strategies that focus on reestablishing key functional groups rather than specific taxonomic assemblages [95,96].

Future research should focus on testing specific predictions of this framework, particularly regarding the relative importance of functional vs. taxonomic diversity for ecosystem stability, the environmental conditions that promote functional redundancy, and the timescales required for functional recovery following disturbance.

4.5. Study Limitations and Future Research Directions

While our meta-analysis provides the most comprehensive assessment of forest succession and microbial diversity patterns to date, several important limitations constrain our interpretations and highlight priorities for future research. Acknowledging these limitations is essential for appropriate interpretation of our findings and for guiding future research efforts.

Temporal resolution and succession dynamics: Our study relies primarily on space-for-time substitutions rather than true temporal data, which limits our ability to track individual microbial communities through succession and may overlook site-specific development trajectories [97,98]. Additionally, our dataset includes limited representation of very long-term succession sequences (>100 years), constraining understanding of late-succession dynamics and potential alternative stable states. Future research should prioritize establishment of permanent monitoring plots with regular microbial community sampling to validate space-for-time inferences and capture temporal dynamics that may be missed in cross-sectional studies.

Methodological heterogeneity and technical standardization: Despite our efforts to account for methodological differences, substantial variation in sequencing technologies, bioinformatics pipelines, and sampling protocols across studies may have introduced systematic biases [99,100]. Earlier studies using less sensitive methods may have underestimated diversity, while more recent high-throughput approaches may detect rare taxa that were previously unobservable. Standardization of sampling and analytical protocols would greatly improve future meta-analyses and enable more precise comparisons across studies and regions.

Functional prediction limitations and validation needs: Our functional diversity analyses based on PICRUSt2 predictions have substantial limitations that must be acknowledged. Functional inference from taxonomic data in soil environments typically achieves <30% accuracy, with NSTI values often exceeding 0.15, indicating poor representation in reference databases [101,102]. These predictions provide only preliminary insights into functional patterns and require validation through direct metagenomic sequencing, metatranscriptomic analysis, or biochemical assays. Future studies should prioritize direct functional measurements over inference-based approaches to develop robust understanding of functional succession patterns.

Geographic scope and global generalizability: Our analysis focuses exclusively on Chinese forest ecosystems, which limits the global generalizability of our findings despite covering diverse climate zones and forest types within China. Different biogeographic regions may exhibit distinct succession patterns due to evolutionary history, species pools, disturbance regimes, and environmental constraints that are not represented in our dataset. Expanding similar analyses to other continents and biogeographic regions would test the universality of U-shaped diversity patterns and climate moderation effects.

Mechanistic understanding and experimental validation: While our correlational analyses identify important patterns and relationships, they cannot establish definitive causal mechanisms underlying observed succession patterns. Controlled experiments manipulating specific factors (resource availability, plant diversity, disturbance intensity) would provide stronger evidence for proposed mechanisms and enable more precise management recommendations.

Scale dependencies and landscape effects: Our study focuses primarily on local-scale (plot-level) diversity patterns, potentially overlooking landscape-scale processes, such as dispersal limitation, source-sink dynamics, and spatial heterogeneity effects, that may influence community assembly during succession. Future research should examine how landscape context, connectivity, and spatial scale affect succession–diversity relationships.

Future research priorities emerging from our findings include the following: (1) Long-term monitoring networks: Establishing permanent plots across succession gradients with standardized sampling protocols to track temporal dynamics directly; (2) Multi-omics integration: Combining metagenomics, metatranscriptomics, and metaproteomics to understand functional succession patterns more comprehensively; (3) Experimental manipulations: Conducting controlled experiments to test specific mechanisms underlying succession patterns; (4) Global comparative analyses: Expanding meta-analyses to other biogeographic regions to test generality of patterns; (5) Predictive modeling: Developing models that integrate microbial diversity with ecosystem functions and environmental change projections; (6) Management applications: Translating research findings into practical management tools and restoration guidelines [103,104]; (7) Database development: Creating comprehensive, standardized databases of forest microbial diversity to support future synthetic research and management applications [105,106].

5. Conclusions

This comprehensive meta-analysis of 479 sites across Chinese forests provides the first systematic quantification of soil microbial diversity patterns during forest succession, revealing several key findings with broad ecological and practical implications for forest management and biodiversity conservation.

Primary findings: (1) Microbial $\alpha$-diversity exhibits a robust U-shaped pattern across forest successional stages, with early succession (0–15 years) and mature forests (>50 years) supporting significantly higher diversity than middle-aged forests (16–50 years). This pattern was consistent across multiple diversity indices and geographic regions, suggesting fundamental ecological processes governing community assembly during ecosystem development. (2) Different microbial taxonomic groups show distinct succession responses, with fungi exhibiting the strongest effects (SMD = 0.61), followed by bacteria (SMD = 0.49) and actinomycetes (SMD = 0.42). These differences likely reflect varying life history strategies, environmental dependencies, and ecological roles among major microbial taxa. (3) Natural secondary forests consistently support higher microbial diversity than plantation forests (SMD = 0.42), with the largest differences observed for fungal communities. Mixed-species plantations show intermediate diversity levels, suggesting that diversification strategies can partially mitigate negative plantation effects. (4) Climate zones significantly moderate succession–diversity relationships, with subtropical regions showing the largest succession-driven changes ($\Delta$Shannon = 0.68), indicating optimal conditions for microbial community development. (5) Soil organic carbon ($\beta$ = 0.68), total nitrogen ($\beta$ = 0.43), and pH ($\beta$ = −0.35) serve as key environmental mediators connecting successional stages with microbial diversity changes, providing mechanistic insights into succession processes.

Theoretical contributions: Our findings advance ecological theory by demonstrating that microbial succession follows predictable, non-linear patterns analogous to plant succession while revealing unique features of microbial community assembly. The U-shaped diversity pattern suggests that both early successional and mature forest stages are critical for maintaining landscape-scale microbial diversity, challenging simple assumptions about monotonic diversity increases during succession. The proposed functional redundancy–resilience framework provides a theoretical foundation for understanding how microbial communities maintain ecosystem functions despite taxonomic turnover during succession.

Management implications: These results have important practical applications for forest management and restoration strategies. The consistently higher diversity in natural secondary forests emphasizes the critical importance of protecting existing natural forest remnants while working to improve plantation management through increased species diversity and reduced management intensity. The climate-dependent nature of succession–diversity relationships indicates that restoration strategies should be tailored to local climatic conditions, with appropriate expectations for recovery timelines and management interventions.

Conservation significance: Our findings support landscape-level conservation approaches that maintain forests across multiple successional stages rather than focusing exclusively on old-growth preservation. Both early successional habitats and mature forests contribute unique components to regional microbial diversity, suggesting that diverse forest age structures enhance overall biodiversity conservation. The pronounced climate zone effects highlight the particular importance of subtropical forest conservation given the high sensitivity of microbial communities in these regions.

Global change context: The climate-dependent nature of succession–diversity relationships has important implications for ecosystem responses to global climate change. As climate zones shift geographically with continued warming, established succession trajectories may be disrupted, requiring adaptive management strategies that account for changing environmental conditions. Forest managers may need to adjust species selections, management practices, and restoration targets to maintain desired microbial community outcomes under future climate scenarios.

Future perspectives: While this study provides a comprehensive foundation for understanding microbial succession patterns, several important research needs remain. Long-term monitoring studies are essential for validating space-for-time inferences and understanding temporal dynamics. Direct functional measurements through metagenomic approaches are needed to move beyond taxonomic diversity patterns to functional succession dynamics. Experimental studies can test specific mechanisms underlying observed patterns and improve management recommendations. Expansion of similar analyses to other biogeographic regions will test the global generalizability of these findings.

As China continues its massive reforestation and forest restoration efforts, incorporating microbial diversity considerations into forest planning could substantially enhance both biodiversity conservation and ecosystem service provision. The predictable patterns identified in this study provide a scientific foundation for evidence-based forest management that considers the complex relationships between succession, climate, management practices, and microbial community development. Moving forward, integrating microbial diversity metrics into forest management planning and restoration monitoring will be essential for achieving sustainable forest management objectives in an era of global environmental change.