Functional diversity of soil microbial communities increases with ecosystem development

Sveen, Tord Ranheim; Viketoft, Maria; Bengtsson, Jan; Strengbom, Joachim; Lejoly, Justine; Buegger, Franz; Pritsch, Karin; Fritscher, Joachim; Hildebrand, Falk; Osburn, Ernest; Hannula, S. Emilia; Bahram, Mo

doi:10.1038/s41467-025-66544-8

Download PDF

Article
Open access
Published: 22 November 2025

Functional diversity of soil microbial communities increases with ecosystem development

Nature Communications volume 16, Article number: 10408 (2025) Cite this article

8456 Accesses
3 Citations
14 Altmetric
Metrics details

Subjects

Abstract

Land abandonment is the single largest process of land-use change in the Global North driving succession and afforestation at continental scales, but assessing its impacts on soil microbial communities remains a challenge. Here, we establish a nationwide successional gradient of paired grassland and forest sites to track developments in microbial structure and functioning following land abandonment and gradual land-use change to forests. We show that microbes generally respond through threshold dynamics, leading to increasing functional but decreasing taxonomic diversity. Succession also entailed specialization of microbial nutrient (C-N-P) cycling genetic repertoires while decreasing genetic redundancy. This highlights a putative trade-off between two desirable ecosystem properties: functional diversity and functional redundancy. Fungal functional diversity underpins higher microbial C-cycling capacity, underscoring the link between functional traits and ecosystem processes. Changing litter quality similarly provides a mechanistic link between plant and microbial communities despite otherwise largely decoupled successional developments above- and belowground. While land abandonment is frequently touted as an opportunity to increase biodiversity and carbon storage, our results show that deeper knowledge about the multifaceted development of soil microbial communities and their links to plant communities during succession may be needed to fully grasp the impacts of global land abandonment processes.

Patterns in soil microbial diversity across Europe

Article Open access 08 June 2023

Root carbon inputs outweigh litter in shaping grassland soil microbiomes and ecosystem multifunctionality

Article Open access 19 December 2024

The soil microbiome as an indicator of ecosystem multifunctionality in European soils

Article Open access 14 December 2025

Introduction

Agricultural land abandonment (hereafter land abandonment) and afforestation are the two largest processes of land-use change in the Global North¹, profoundly altering landscapes, ecosystems, and biodiversity patterns². It is estimated that ~500 Mha of agricultural land has been abandoned globally since the onset of industrialization in the mid-1800s³. After management ceases, abandoned fields are usually colonized by pioneer grasses and herbs, followed by shrubs, bushes and trees, until a stage of full afforestation is reached⁴. These changes occur gradually, through a directional and largely predictable ecosystem development referred to as secondary succession⁵ (hereafter succession). However, the traditional view of a putatively undisturbed “climax” forest (sensu Clements⁶) as the successional endpoint is increasingly disputed^7,8,9, and the vast majority of abandoned grasslands in Europe instead ending up as production forests incorporated into management regimes comprising low diversity, short rotation times, and clear-cutting^10,11,12,13. The joint processes of land abandonment and forestry management pose serious threats to a range of organismal groups aboveground^{14,15,16,17,18}, but few studies have comprehensively examined their effects on soil microbial diversity and functioning.

Soil microbes constitute the largest pool of terrestrial diversity and undergo gradual changes in response to changing resource quality and soil properties during succession. For instance, soil fungal communities typically respond to decreasing litter quality of plants (resource-driven succession) through changes in the leaf-dry matter content (LDMC)¹⁹, whereas bacteria are generally more sensitive to changes in abiotic properties like pH and soil organic matter quality (soil C:N)²⁰. However, in contrast to the largely predictable patterns observed during plant succession, microbes often exhibit unimodal (i.e., hump-shaped) or threshold responses^21,22,23. This could be related to sudden shifts in the metabolic trait distribution of the microbial decomposer communities²⁴, leading to abrupt transitions between microbial generalists and specialists as plant litter becomes increasingly recalcitrant to microbial degradation^25,26. Ecological theory posits that niche differentiation and specialization of the microbial saprotrophic community increase with increasing litter complexity and heterogeneity²⁷, and that this is accompanied by “tighter” and more efficient nutrient cycling processes. Yet, contrasting evidence also show high overlap of microbial genes related to nutrient cycling during succession^22,28, suggesting that metabolic traits remain unaltered during afforestation. This points to a high level of genetic redundancy, where the genes needed to perform nutrient cycling are widely shared across multiple microbial taxa²⁹. Functional specialization and functional redundancy are both important features underpinning ecosystem functioning and stability, but few studies have reconciled these dimensions across ecosystems undergoing land-use change. Existing studies of microbial successional dynamics are moreover often hampered by limitations inherent to space-for-time designs (i.e., chronosequences)³⁰, which often fail to capture the spatial scales of land abandonment¹³.

Here, we establish a nationwide gradient spanning from managed and open grasslands to abandoned and increasingly afforested sites undergoing succession (Fig. 1a, c). Each grassland site was paired with an adjacent forest site representing its land-use change endpoint³¹ (Fig. 1b), typically a coniferous-dominated forest managed for wood production because historical land-use maps show that managed forest sites better represent the outcome of abandoned agricultural land than undisturbed forests^{11,13,32,33,34}. We use amplicon (16S for bacteria, ITS for fungi) and metagenomic sequencing (Fig. 1d) to study how microbial taxonomic diversity and functional genes encoding enzymes involved in C-N-P cycling developed with abandonment and afforestation, including links to measured degradation rates of a range of carbon substrates. We hypothesize that increasing resource complexity would increase the functional diversity and reduce the overlap of functional genes, leading to divergent levels of genetic redundancy and specialization with succession. We further expect genetic specialization and redundancy to be reflected in the degradation of simple and complex substrates, with high genetic redundancy leading to high degradation rates of low-complexity substrates and vice versa for specialization. Lastly, we expect these community changes to occur predominantly as unimodal or threshold-like responses (Fig. 1e).

Results

Threshold dynamics in soil microbial communities during grassland afforestation

Our successional gradient revealed a striking decrease of microbial (taxonomic) diversity and marked shifts in community composition with land-use change between grasslands and forests (Fig. 2). By contrast, grassland microbes were overall structurally and functionally similar, independent of the successional stage (i.e., managed, recently abandoned, and late-stage successional grasslands, Fig. 1c and of the increasing levels of afforestation (Table S6-7). The abrupt decrease of bacterial diversity is consistent with threshold-like tipping points (Fig. 1e) occurring between late-stage successional grasslands and fully afforested sites (Fig. 2b). In contrast, fungal diversity declined gradually, and this evened out during the later stage of succession (Fig. 2a). Community composition differed strongly between grassland and forest sites (Fig. 2c, d), with fungal community composition additionally differed between managed and abandoned grasslands (Permanova: p < 0.05 for all comparisons, see Table S8 for full test details), indicating that a compositional shift occurs after the cessation of grassland management. Effect sizes (partial ω²) were derived from comparisons between paired grassland and forest sites while including pairs and spatial distance between pairs as random factors to account for the paired structure of the dataset. These showed overall larger effect sizes between managed grasslands and forests compared to abandoned and successional grasslands (Fig. 2e), and considerably higher for bacteria than for fungi (partial ω² = 0.46 and 0.19 for bacterial alpha and beta diversity, and partial ω² = 0.07 and 0.05 for fungal alpha diversity and beta diversity, respectively). Variable selection analyses revealed that thresholds in diversity changes coincided with a sharp decline in soil pH, increasing soil C:N ratio, and higher levels of leaf dry matter content (LDMC; Table S3–S4), whereas fungal diversity was significantly related only to the geological parent material of the soils (Fig. S2a, c). Partitioning the contribution of these variables to the observed changes in alpha diversity revealed soil pH as the primary determinant of bacterial diversity decline during afforestation (Fig. S2d), whereas fungal diversity was mainly driven by changes in total soil organic carbon (SOC), parent material, and C:N (Fig. S2b). The inclusion of geographical distance between paired grassland-forest sites to account for spatial autocorrelation and climate factors (temperature, precipitation) did not alter these results (Table S6).

**Fig. 2: Threshold changes to microbial diversity between grasslands and forests.**

Because the paired forest reference sites were managed and heterogeneous in age and structure, we acknowledge that thresholds between late-stage successional sites and forests could be due to management factors rather than effects of afforestation. To test the influence of forest management on the results obtained here, we retrieved stand age and evenness as proxies of forestry management and examined whether the structure of microbial communities differed depending on these characteristics. Results showed that fungal community composition but not alpha diversity varied according to stand evenness (Permanova: Pseudo-F_1,2 = 1.28, p = 0.02), whereas bacterial diversity and community composition did not vary with any of the factors (Table S9–S10).

Functional diversity increases relative to taxonomic diversity during land abandonment

We next examined how the diversity of functional genes encoding enzymes involved in soil biogeochemical cycling (i.e., C, N, P-related genes) changed with succession and land-use change. Also, here, we found no differences between grasslands across differing stages of management and succession (p > 0.05 for all pairwise comparisons; see Table S11 for full test details). Conversely, threshold effects were notable in the response of fungal C-cycling genes, which increased in forests compared to grasslands (Fig. 3a–c). By contrast, the diversity of bacterial C-cycling and all microbial P-cycling genes remained at similar levels across all land uses (Fig. 3d–f). The diversity of bacterial N-cycling genes also decreased in forests compared to grasslands (Fig. 3g, h). As with taxonomic diversity patterns (Fig. 2e, f), effect sizes for functional genes were generally higher between managed grasslands and forests compared to successional (recent, late) grasslands (Fig. 3c, f, h). Forest age and evenness structure did not significantly affect the diversity and composition of any of the functional gene categories (Table S9, S10). Moreover, when juxtaposing the results obtained from the taxonomic and functional profiling (Figs. 2, and 3), a pattern of increasing functional relative to taxonomic diversity emerged, suggesting changes to microbial redundancy and functioning.

**Fig. 3: Contrasting genetic diversity development during succession.**

Grassland abandonment leads to loss of genetic redundancy

The absence of direct relationships between taxonomic and functional diversity is considered a strong indicator of functional redundancy, as taxa can be replaced without losing genetic potential to carry out ecosystem functions^35,36. To examine how the observed patterns in taxonomic and functional diversity related to functional redundancy, we combined all grassland sites into one single grassland category as the differences between successional stages were minimal (Figs. 2, and 3). Next, we used ordinary least-square regression (OLS) to test the strength and direction of the relationship between taxonomic diversity (predictor) and functional diversity (response variable). For fungi, the marked increase of C-cycling genes in the forest sites occurred despite a general loss of taxonomic diversity (Figs. 3a vs 2a), producing a negative relationship (OLS: df = 85, estimate = −0.15, p = 0.011; Fig. 4a) which was not significant in grasslands (OLS: df = 40, estimate = −0.04, p = 0.694). Comparing the slopes of both ecosystems (grasslands, forests) showed that these differed significantly (Anova: F_1,86 = 42.6, p < 0.001), indicating that the coupling of taxonomic diversity and genetic C-cycling potential undergoes substantial alteration as grasslands transition to forests. Similar results were obtained when assessing bacterial C-cycling and N-cycling genetic diversity (Fig. 4b, e). Both these genetic pools were moreover significantly related to taxonomic diversity in forests but not in grassland soils (OLS, p < 0.05 for both comparisons; see Table S13 for full test details). Because the number of functional genes scales with genome size for bacteria³⁷, the increase in functional genes relative to taxonomic diversity could be due to increasing bacterial genomes in forest soils. To account for this, we included average genome size (AGS) as a fixed effect in all models assessing bacteria, but this did not alter the direction or significance of the taxonomic-functional diversity relationships (Table S13). For P-cycling genes, diversity coupling was significant in both ecosystem types for bacteria (Fig. 4d), albeit stronger in forests than in grasslands (Anova, F_1,87 = 34.2, p < 0.001). Interestingly, no corresponding coupling was found for fungi (Fig. 4c), indicating that functional redundancy may be relatively high when it comes to fungal P-cycling independent of ecosystem development. Soil pH and C:N were the main factor related to genetic C-cycling diversity for both fungi and bacteria (Fig. S3a-f), with hierarchical partitioning additionally showing high influence ( ~ 14% explained variation) of bulk density (Fig. S3b). The diversity of bacterial N-cycling genes was similarly driven by C:N but additionally showed high influence stemming from leaf litter quality (Fig. S3i, j). In contrast, the diversity of P-cycling genes was mainly driven by a mix of pH, total N content, parent material, leaf litter quality, and mean annual temperature (Fig. S3c-h).

**Fig. 4: Genetic redundancy decreases with change from grasslands to forests.**

Widespread genetic specialization of the soil microbiome during ecosystem development

To examine whether microbial communities become more functionally specialized after land abandonment, we quantified the average niche overlap of C-N-P-related genes among successional stages (see “Methods”). Results showed decreasing genetic overlap of genes related to C- and P- cycling when going from managed grasslands to forests for both fungi and bacteria (Fig. 5a–d). Second-order polynomial regressions provided the best fits for P-cycling genes and bacterial C-cycling genes (Table S15), whereas fungal C-cycling genes decreased linearly with ecosystem development (Fig. 5a). In contrast, the overlap of N-cycling genes, calculated for bacteria only due to the lack of databases containing fungal N-cycling genes, showed a weakly unimodal pattern peaking in early successional grasslands (Fig. S4). We next partitioned genetic overlaps across differing substrates of increasing carbohydrate complexity (C-cycling; Fig. 5f, g) and pathways (N-P-cycling; Fig. S4, 5). Also here, we found that genetic overlap decreased across the full spectrum of simple and complex substrates. Notably, P-cycling specialization increased gradually for fungi, whereas the overlap between many of the P-cycling pathways decreased sharply between managed and abandoned grasslands for bacteria (Fig. S5).

**Fig. 5: Community-wide genetic specialization of microbial nutrient cycling increases with succession.**

A diversity-redundancy framework for soil bacterial communities during succession

Our results strongly indicate that communities shift from high genetic redundancy to increasing genetic specialization during land-use change from grasslands to forests. We formalized and tested this putative shift using the framework developed in ref. ³⁸, where functional and taxonomic dimensions of a community is combined to yield three interrelated properties describing taxonomic diversity, functional uniqueness (hereafter used synonymously with specialization) and functional redundancy. These interrelated properties can be further visualized along a main axis of redundancy and specialization (Fig. 6a)³⁹. Importantly, as these analyses require traits-by-taxa matrices, we first predicted bacterial metagenomes from amplicon sequences (n = 204) using picrust2⁴⁰ and then inferred C-N-P-cycling genes based on KEGG orthology (see “Methods”). Predicted metagenomes are of limited value for analyzing specific pathways⁴¹, so we focused only on overall changes in the diversity of metabolic pathways between ecosystems. Based on the predicted metagenomes, we assessed redundancy and specialization under the hypothesis that grasslands and forest sites would cluster differentially along the vertical axis separating these diversity dimensions (Fig. 6a). As expected, functional redundancy related to C-cycling decreased sharply between grasslands and forests, whereas functional specialization instead increased (Fig. S6), consistent with a threshold response to afforestation occurring late in the successional gradient. The quantity (Total C) and quality (C:N) of organic carbon were the main variables driving redundancy and specialization (Fig. S7a), differing from the factors affecting bacterial alpha diversity and community composition (Fig. S2). Clustering of grassland and forest sites occurred mainly along the redundancy-specialization axis, in line with our hypothesis (Fig. 6b; Fig. S8), and these results were further corroborated by compositional differences between the two land-use types (Table S19). Notably, N- and P-cycling showed the same redundancy-specialization divergence (Fig. S8), whereas no differences were found between any of the differing grassland successional stages. However, the main factors underlying these results varied, with N-cycling more related to soil pH in addition to soil carbon quality (Fig. S7b) whereas P-cycling was related to the successional stage of the site and plant leaf traits (Fig. S7c).

**Fig. 6: A general framework of redundancy-diversity divergence during succession.**

Functional diversity is of key importance for substrate degradation

Lastly, we tested whether increasing functional diversity and specialization during succession would relate to degradation rates of substrates across a gradient of increasing complexity. For this, we incubated a subset of paired grassland-forest samples (n = 154) and measured substrate-induced respiration (SIR) across six substrates ranging from easily degradable oligosaccharides (glucose) to complex recalcitrant substrates like lignin and chitin. Degradation rates were generally higher in forest than in grassland soils but increased with abandonment and succession (Fig. 7a, and Fig. S9). Interestingly, the least and the most recalcitrant substrates (glucose, chitin) were exceptions to this, with no differences between any of the grassland and forest soils (Table S19). When combining the full range of degradation rates into an aggregate measure of functional capacity (MSIR, see “Methods”) and using this as a response variable with microbial taxonomic and functional diversity as predictors, we found that MSIR was most associated with fungal C-cycling genetic diversity (Fig. 7b). Notably, this relationship did not differ between grasslands and forests (Anova: F_1,70 = 0.015, p = 0.903), suggesting fungal control over substrate degradation across the whole range of grasslands and forests investigated. In contrast, we found no corresponding relationship for bacterial C-cycling genetic diversity (Fig. 7c). MSIR moreover correlated negatively with bacterial taxonomic diversity (df = 143, r = −0.37, p < 0.001) and showed no significant relationship with fungal diversity (df = 142, r = −0.06, p = 0.49).

**Fig. 7: Fungal functional diversity drives substrate degradation in grassland and forest soils.**

Discussion

Elucidating the patterns and processes occurring after land abandonment is a notorious challenge in ecology because the gradients used to infer ecosystem development are spatially confined and fail to match the extent of land abandonment. By establishing a nationwide gradient of paired grassland and forest sites (Fig. 1), we could examine how microbial communities respond to land abandonment and land-use change at an unprecedented scale. Contrary to the notion of the putatively undisturbed “climax” forests (sensu Clements⁶) inherent to much successional theory, we opted for using conventionally managed forest sites across a broad range of stand age and evenness (Fig. S1) as reference endpoints for adjacent grasslands. This choice reflects an assumption that grasslands, once abandoned and if not restored, will predominantly end up as managed production forests, as reflected in historical land-use maps across Sweden and Europe^{11,13,32,33,34}.

We found that microbial community responses during grassland abandonment and afforestation varied depending on organismal group (i.e., bacteria vs fungi) and the diversity aspect examined (i.e., taxonomic vs functional diversity). For bacteria, considerable taxonomic diversity was lost in the transition between late-stage successional grasslands and forests (Fig. 2b), whereas the corresponding decrease of fungal diversity was gradual (Fig. 2a). The overlap of annotated C, N, and P-cycling genes similarly decreased across the land-use gradient (Fig. 4a-d), despite overall few changes (Fig. 3b–d) or threshold dynamics (Fig. 3a, e) of the overall functional gene diversity. These results both confirm and contrast with our hypothesis and with foundational successional theory, where development is thought to occur as a gradual and largely predictable series of changes, and highlight discrepancies across the community properties examined. Reduced overlaps of annotated functional genes were generally better fitted to second-order polynomial than OLS-regressions (Table S14), indicating that changing niche space is non-linear and likely driven by a mix of the change in resource quality with changing plant community composition and by abiotic filtering²². Metabolic specialization of the soil microbiome occurring during grassland succession is therefore likely to be reflected in a non-linear transition between generalist and specialist taxa, consistent with findings within the overall framework of the fundamental microbial metabolic niche⁴² where non-linear taxa overlap and transitions between generalist and specialist have been previously shown across environmental gradients^43,44. Indeed, a recent survey of more than 230 soil bacterial communities across a continental scale showed similar bimodal distributions of generalist and specialist taxa, based on their average niche breadth²⁶.

Here, we observed that threshold responses in taxonomic and functional diversity coincided with a sharp decline in soil pH, increasing soil C:N, and higher levels of leaf dry matter content between grassland and forest sites (Table S3). These are all emblematic changes in soils undergoing succession after land abandonment in boreal biomes¹⁹. Of the key ecosystem properties identified, soil pH, C:N, total soil organic carbon (SOC), and leaf quality (LDMC) were moreover consistently singled out as key properties related to both taxonomic and functional diversity in both variable selection analyses and in hierarchical partitioning of fixed effects (Fig. S2, S3). Notably, LDMC provides a powerful “functional marker” of multiple ecosystem processes during secondary succession⁴⁵, including microbially driven litter decomposition¹⁹, whereas increases in SOC have been linked to increasing microbial taxonomic diversity and enzyme potential during secondary succession previously²². The depletion of base cations through tree growth leading to soil acidification⁴⁶ and a build-up of recalcitrant leaf litter⁴⁵ could be important processes underpinning these abiotically driven changes, whereas the accumulation of plant detritus during afforestation, visible in the increasing levels of SOC in forest sites (Table S3), has also been shown to support a more active microbial community⁴⁷. This effect is however likely modulated also by the quality of the litter²². Thus, together, LDMC, C:N, and SOC reflect changing resource quality and input during grassland afforestation which, together with soil pH, profoundly alter soil microbial community structure and functioning.

Afforestation of grasslands increase soil acidity by several pH units⁴⁸, but to our knowledge, there are no time-series to indicate whether this development tends to be linear or threshold-based. Because our chosen forest sites were generally managed, the observed threshold effects could also reflect a “management gap” between late-stage successional grasslands and the fully afforested endpoints, potentially setting these apart from the trajectory of a naturally developing ecosystem after abandonment. Yet, as we found generally few differences in the taxonomic and functional components of the soil microbiome across forests of differing ages and evenness classes (Table S9-10), we are inclined to think that forest management is not a main driver of the thresholds observed. A similar gap between late-stage successional grasslands and an unmanaged forest reference site was moreover found in ref. ²², indicating that threshold dynamics late in the successional gradient could be an intrinsic feature of microbial community development after land abandonment. Resolving this question likely requires time-series that circumvent the pitfalls of space-for-time substitutions altogether³⁰.

The absence of significant correlations between taxonomic and functional diversity is indicative of high functional redundancy, because removing taxa in redundant systems will thereby not affect the genetic diversity⁴⁹. In other words, functional redundancy is visible in the decoupling between taxonomic and functional diversity⁵⁰. We found differences in this relationship between grasslands and forests, with notably higher functional diversity relative to taxonomic diversity in forest soils (Figs. 2–4). This suggests that fewer taxa contribute to the overall pool of genetic diversity in forest soils and is corroborated by the generally strong associations between taxonomic and functional diversity in these sites but not in grassland (Table S13). These relationships were independent of the communities’ average genome size (AGS; Table S13), suggesting that a shift toward taxa with larger genome size is not the main driver behind the loss of bacterial redundancy in forest soils⁵¹. By contrast, larger fungal genomes in forest soils could explain the negative relationship observed between fungal taxonomic and C-cycling genetic diversity (Fig. 4a), where taxa with smaller genomes containing fewer C-cycling genes are presumably replaced by fewer taxa with larger genomes harboring more genetic diversity. If so, this would be another indication of functional specialization, in line with the decreasing overlap of fungal C-cycling genes across the gradient (Fig. 5a). Fungal genome sizes have been reported to increase in nutrient-poor soils⁵², where fertility was defined as the availability of inorganic N and P. Given the large increase of SOC and C:N in forest compared to grassland soils here, it is plausible that the same elements are stoichiometrically limited and could yield the same pattern of increasing fungal genomes. However, due to the lack of tools to reliably infer AGS for eukaryotes, this could not be tested explicitly here. We also note the generally negligible influence of climate (mean annual temperature and precipitation) on microbial alpha and beta diversity (Figs. S2, S3), indicating consistent microbial successional dynamics across differing climatic zones.

The generally decreasing overlaps in the genetic repertoires of both fungal and bacterial communities with succession moreover indicate that specialization occurrs in all measured aspects of nutrient cycling (Fig. 5a–d; Fig. S4) and across all pathways (Fig. 5e, f, Figs. S4, 5). This is in line with recent evidence showing that bacterial specialization along one axis of environmental variation may explain up to 80% of the variation in all other axes²⁶, meaning that specialization within only one of multiple metabolic dimensions (e.g., P-cycling) is rare. Yet, it is important to acknowledge that the lack of databases for functional annotation of eukaryote N-cycling genes impedes similar analyses of fungal N-cycling gene redundancy and specialization.

Whether the loss of genetic redundancy also translates into a loss of functional redundancy is an open question⁵⁰. Conversely, we may ask whether increased genetic diversity and specialization increase the rates or efficiency of ecological processes such as C-N-P-cycling. We measured the degradation rates of a range of increasingly complex substrates representative of changing litter quality during succession and found that these were generally higher in forest than grassland soils (Fig. 7a). This contrasts with our hypothesis of higher degradation rates in grassland due to genetic redundancy of C-cycling genes and suggests that functional specialization may be an important factor driving degradation of individual litter chemicals during decomposition⁵³, possibly due to higher complementarity in communities with low metabolic overlaps. In forest soils, the litter-derived pool of organic matter is typically dominated by structurally complex substrates enriched in recalcitrant lignin and chitin⁵⁴, and the enzymatic degradation of these into progressively simpler molecules selects for a highly specialized web of microbial cross-feeding⁵⁵. By contrast, grassland soils are characterized by simpler and more easily degradable carbohydrate inputs, in which metabolic generalists with broad enzymatic repertoires dominate the decomposer community⁵⁶. Thus, efficient resource partitioning among a community of specialist taxa provides a framework through which the higher substrate degradation rates in lower-diversity forest soils can be plausibly explained. This is in line also with a recent study finding both empirical and theoretical evidence for increasing dominance of specialist bacterial taxa with the supply of additional glycolytic carbon sources⁵⁷, which is likely to be an accurate reflection of the build-up of multiple sources of increasingly complex substrates in forest soils during succession⁵⁸. However, we stress the observational nature of our study and that that our results are not necessarily causative, underscoring the need for more research to trace carbon flows throughout the decomposer community and pinpoint its mediating metabolic pathways. We also observed significant correlations between fungal but not bacterial diversity of C-cycling genes and aggregated degradation rates (i.e., MSIR) across all substrates in both grassland and forest soils (Fig. 6b), suggesting that saprotrophic fungi are of high functional importance also in these systems.

Functional diversity and functional redundancy are both crucial components supporting ecosystem functioning and stability^59,60,61. Our results indicate that these two diversity components may be inversely related during succession (Fig. 6c), in line with our first hypothesis. Using computationally predicted bacterial metagenomes, we found that grassland and forest sites clustered regularly along the redundancy-specialization axis for all nutrient cycling processes (Fig. 6a; and Fig. S8), and that this clustering was likely driven by changing soil C:N, pH, and LDMC (Fig. S7). This corroborates our hypothesis that changes in resource complexity and soil acidity trigger shifts from taxonomically diverse but functionally redundant (i.e., more similar with respect to functioning) to functionally rich communities with more specialized communities. An intriguing outcome of this presumed trade-off is that ecosystems undergoing afforestation after land abandonment may become more efficient when it comes to carbon- and nutrient cycling due to higher functional diversity and specialization, but less resilient to disturbances due to lower functional redundancy⁶⁰. Again, we emphasize that these results are based on the diversity of metabolic traits, here inferred through functional annotation of genes at community level, for which the relationship with actual process rates is highly context dependent^62,63 and may blur at the level of higher-order functions such as decomposition⁵⁰. Similarly, the divergence pattern identified here relates to predicted metagenomes of soil bacteria only, with the absence of adequate tools to predict fungal metagenomes from marker sequences precluding corresponding analyses for fungi and will need to be verified using communities with known taxonomy and functional profiles. However, while the idea of a redundancy-specialization trade-off has been previously proposed in the context of global bird, mammal, and plant distributions^39,64, it has not yet been evaluated and tested for soil microbes. Our results provide a valuable insight into these dynamics in the context of the well-studied phenomenon of ecological succession, anchored in a nationwide gradient with high generalization. Based on these findings, we predict that other findings of negative relations between taxonomic and functional diversity across diverse environmental gradients^51,65,66 would yield the corresponding redundancy-specialization trade-off observed here. In conclusion, our study reveals that grassland abandonment and afforestation entail a broad range of changes to the soil microbiome when compared to adjacent forest sites, including a loss of taxonomic diversity while the diversity of functional genes encoding for C-N-P-related enzymes remains stable or increases. These dynamics imply threshold responses and a gradual transition from high genetic redundancy to functionally specialized microbiomes, with potential consequences for their capacity to degrade differing sources of carbohydrates. Further studies are needed to investigate whether the transition from functionally redundant to specialized microbiomes affects the long-term stability of ecosystems.

Methods

Study design

The study comprised 102 grassland and 105 forest sites (Table S1) derived from environmental monitoring programs conducted at the Swedish University of Agricultural Sciences (SLU). Grasslands were initially categorized into differing successional stages (Managed, recently abandoned, Late successional) based on site descriptions, and this was corroborated by permutational distance-based multivariate analyses (db-MANOVA, Table S2) for differences in plant community composition coupled with indicator species analysis⁶⁷. As historical records show that abandoned grasslands in Sweden predominantly end up as conventionally managed production forests^11,12,13,68, we selected production forests (i.e., > 1 m³ ha⁻¹ year⁻¹ biomass production) as the reference endpoint for the abandoned grasslands. These forest sites were characterized by considerable variation in stand age and evenness (Fig. S1), indicating differences in management and/or growth cycle. To increase the accuracy of the forest reference, we paired each grassland with an adjacent forest site based on a geographical proximity criterion ( < 6.5 km distance between paired sites, median distance = 3.63 km), under the explicit assumption that grasslands, which, if and when abandoned and fully afforested, would share approximately the same vegetation and soil microbiome characteristics as their adjacent forest sites³⁰. Altogether, 190 (92 %) of the 207 sites were paired. Although geographical pairing ensures that sites are located within the same regions and increase the likelihood that grassland and forest sites share broadly the same environmental characteristics, we acknowledge that local differences in climate, parent material, and topography exist at this spatial scale and could lead to discrepancies in the expected successional trajectory. Climate data comprising mean annual temperature (MAT) and precipitation (MAP) for each site was collected from the Swedish Meteorological and Hydrological Institute’s (SMHI) long-term local weather stations. Information about the parent material for each site was gathered from the Swedish Geological Survey (SGU).

Soil sampling

The sites were sampled between July and September 2020 following the standardized protocol established in ref. ⁶⁹. Briefly, at each site, 15 samples were collected across a circular area of ~ 200 m² using a soil corer (diameter = 3 cm, depth = 10 cm), and pooled into a bulk sample. Surface layers constituting loose litter and fibric material was removed prior to sampling but otherwise no distinction was made between the organic and mineral layers when sampling. By sampling at a standard soil depth, we therefore compare samples with differing levels of organic horizon depths, which strongly shapes resource availability, mineralogy, and physicochemical conditions. In particular, peaty soils with no discernible mineral horizon differ strongly from mineral soils. A total of 12 forest sites (11 % of all forest soils) were peaty soils without any mineral horizon, but this category also comprised a total of 7 grassland soils (6.8 % of all grassland sites). The alternative approach to sample at a consistent 0-10 cm depth in the mineral horizon was discarded as this would have failed to capture most of the resource-driven community changes occurring in the organic soil horizon. A subset of the pooled sample was air dried ( < 40 °C) within 24 h of collection and stored in a zip-lock plastic bag with silica gel to minimize humidity and prevent development of molds during transit and stored frozen (−20 °C) until molecular analyses. All equipment was sterilized with 95% ethanol between sites.

Abiotic drivers of soil diversity

Approximately 15 g of soil was used to analyze pH (1:5 soil:water suspension). Available phosphorus (P-AL) and potassium (K-AL) were extracted using ammonium lactate and acetic acid at pH 3.75⁷⁰ and analyzed using the stannous chloride-molybdate procedure (P-AL) and inductively coupled plasma atomic emission spectroscopy (ICP-AES), respectively. Exchangeable calcium and magnesium concentrations were measured in ammonium acetate extract (pH = 7.0) using ICP-AES at Agrilab Uppsala, Sweden. Total C and N contents was determined on aliquots (1-20 mg) of air-dried soil using an Elemental Analyzer (Euroa EA, Eurovector, Milano, Italy). We used leaf dry matter content (LDMC) as a proxy for leaf quality, as this has been shown to correlate well with litter quality and decomposition in successional systems in our study area¹⁹ LDMC values were derived from the TRY database⁷¹ (Ver. 6.0) and covered 120 out of the total 235 plant species (51.1 %) recorded across all sites, and converted to community-weighted means. While we acknowledge that traits derived from plant databases disregard intraspecific variation that could affect the results, it has been shown that interspecific trait variation is of considerably greater importance than intraspecific variation⁷².

Substrate-induced respiration

We performed catabolic profiling on a subset (n = 156, or n = 77 paired grassland-forest sites, Table S1) of air-dried soils using substrate-induced respiration (SIR) for a range of carbon substrates of differing complexity, following the protocol established in refs. ^73,74. Briefly, 8 mL solutions of glucose, glycine, oxalic acid, autolyzed yeast, lignin, and chitin was added to 4 g dry-weight equivalent of fresh soil (1 analytical replicate per solution). After 1 h of pre-incubation, soils were incubated for 4 h at 20 °C except for lignin and chitin which were incubated for 24 h. After incubation, respiration for each amendment was determined using Gas Chromotography (Trace CG Ultra Gas Chromatograph (Thermo Fisher Scientific, Milan, Italy)). Basal respiration rates, measured using 8 mL of added ddH₂O, were subtracted from each substrate-induced respiration rate to account for the added effect of water flushing (Birch effect)⁷⁵. We used Multiple substrate-induced respiration (MSIR) as a proxy for the communities’ catabolic profile and total functional capacity⁷⁶, which was calculated by subtracting basal respiration rates from each substrate-amended soil, and subsequently summing the individual substrate respiration rates for each sample.

Molecular analyses

DNA extraction and sequencing

DNA was extracted from 200 mg of dried and milled soil samples using the PowerMax Soil DNA Isolation Mini kit (Qiagen GmbH, Hilden, Germany) following the manufacturer’s instructions. The extracted DNA was quality-checked with 260/280 and 260/230 nm wavelength ratios using a NanoDrop™ (Thermo Scientific, Massachusetts, USA) and stored at −20 °C until sequencing. For bacterial amplicons, the universal prokaryote primers 515 F and 926 R were used to amplify the 16S V4 subregion of rRNA gene⁷⁷. DNA samples were amplified using the following conditions in three replicate runs: 95 °C for 15 min, followed by 26 cycles of 95 °C for 30 s, 50 °C for 30 s and 72 °C for 1 min with a final extension step at 72 °C for 10 min. The 25 μl PCR mix consisted of 18 μl sterilized H₂O, 5 μl 5 × HOT FIREPol Blend MasterMix 0.5 μl of each primer (20 μl) and 1 μl template DNA (final concentration of 400 nM). For amplicons of eukayotes/fungal ITS, we used the universal eukaryote PCR primers ITS9mun and ITS4ngsuni⁷⁸. PCR amplification followed the protocol described in ref. ⁷⁹. Briefly, 0.5 µl of each forward and reverse primer (20 mM), 1 µl of DNA extract and 18 µl ddH₂O were used in combination with 5 µl of 5 × HOT FIREPol Blend Master Mix (Solis Biodyne, Tartu, Estonia). Thermal cycling followed an initial denaturation at 95 °C for 15 min; 25–30 cycles of denaturation for 30 s at 95 °C, annealing for 30 s at 57 °C, elongation for 1 min at 72 °C; final elongation at 72 °C for 10 min; and storage at 4 °C. Duplicate PCR products were pooled and quality checked on a 1 % agarose gel. All primers were tagged with a 10-base pair barcode for sample identification. Blanks containing ddH₂O instead of DNA template were used as negative controls in the library preparation. The amplicons from the replicates were pooled, purified using a purification kit containing agarose gel (FavourPrep Gel/PCR Purification mini Kit-300 Preps; Favourgen) and shipped for library preparation in the sequencing service facility of University of Tartu (the Estonian Biocenter). Sequencing of ITS was performed on the PacBio Sequel System at Novogene (UK), and 16S libraries were sequenced on two runs using an Illumina MiSeq platform (2 × 250 bp paired-end chemistry).

Shotgun metagenomic sequencing was performed on pooled equimolar amounts of DNA from a subset (n = 94, or n = 47 pairs) of grassland and forest sites (Table S1). Library preparation and sequencing were performed at the service provider facilities (Novogene Europe) using the Novogene NGS DNA Library Prep Set kit and sequenced on Illumina NovaSeq with 2 × 150 bp paired end reads.

Bioinformatics

Metabarcoding

We used the LotuS2 version 2.22⁸⁰ pipeline to quality-filter, demultiplex, and process the filtered reads into OTUs. Chimera detection and removal was done using Uchime⁸¹ with all singletons and sequences shorter than 100 bp discarded. For fungal sequences, the ITS region was extracted post-clustering, and used to extract and restrict Blast searches using ITSx⁸². Clustering of sequences was done using a de-novo clustering algorithm in UPARSE⁸³ based on a 97% similarity threshold. Taxonomy was assigned against the SILVA (ver. 138.1) and UNITE (ver. 8.1) databases for prokaryotic and fungal sequences respectively. All datasets were manually curated to remove contaminant sequences based on negative controls. OTUs representing archaea, chloroplasts, eukaryotes, and mitochondria were omitted from the bacterial dataset, and OTUs unassigned at class level were omitted from the ITS dataset after OTU clustering. To reduce biases due to low sequence coverage, we discarded samples with < 100 reads (n = 8 samples) and < 3000 reads (n = 3 samples) from the fungal and bacterial dataset, respectively. This resulted in a total of 204 samples comprising 8,532,758 reads (sample mean ± SD = 41,8127 ± 33,514) across 8937 OTUs for bacteria, and 199 samples with a total of 437,006 reads (sample mean ± SD = 2196 ± 2750) covering 2818 OTUs for fungi.

Shotgun metagenomics

Analysis of metagenomic reads was done using MATAFILER pipeline⁸⁴ using a workflow optimized for complex environmental metagenomes⁸⁵. Briefly, reads obtained from the shotgun metagenomic sequencing of soil samples were quality-filtered by removing reads shorter than 70% of the maximum expected read length (150 bp), with an observed accumulated error > 2 or an estimated accumulated error > 2.5 with a probability of ≥0.01, > 1 ambiguous position or if base quality dropped below 20, using sdm (version 1.46)⁸⁰. All 95 samples produced sufficient quantity of reads (average 37 186 296 ± 7 365 270 reads per sample) and were retained for statistical analyses. To estimate the functional composition of each sample, similarity search using DIAMOND (version 2.0.5; options -k 5 -e 1e-4 –sensitive) in blastx mode⁸⁶ was employed. Prior to that, the quality-filtered read pairs were merged using FLASH (version 1.2.10)⁸⁷. The mapping scores of two unmerged query reads that mapped to the same target were combined to avoid double counting. In these cases, the hit scores were combined by averaging the percent identity of both hits. The best hit for a given query was based on the highest bit score and highest percent identity to the subject sequence. Using this method, we calculated the relative abundance of (clusters of) orthologous gene groups (OG) by mapping quality-filtered reads against the eggnog database (version 5)⁸⁸. We used the MicrobiomeCensus pipeline to estimate average genome size (AGS) of bacterial communities⁸⁹.

Inferring pathways related to C-N-P cycling

Quality-filtered reads were blasted against the CAZyme database⁹⁰ to derive OGs related to C-cycling. These were further mapped against CAZymes with known substrate affinities based on a custom database provided in ref. ⁹¹. For N- and P-cycling genes, we annotated genes from the Kyoto Encyclopedia of Genes and Genomes (KEGG) Orthology database⁹² and mapped these against custom databases containing N- and P-cycling genes provided in refs. ^93,94. The total number of metagenomic reads, including reads and OG for each nutrient cycling process are shown in Table S4.

Predicted metagenomes

Testing for trade-offs between functional redundancy and specialization require joint taxa-function matrices to assess functional distances between taxa³⁹. We used Picrust2 (ver. 2.5.2)⁴⁰ to predict functional metagenomes (PFM) from quality filtered 16S rRNA sequences. The weighted nearest sequenced taxon index (NSTI) of the PFM ranged between 0.129 and 0.312, indicating generally good predictions. From the PFMs, we next extracted the functional pathways related to C-N-P cycling based on OG from the KEGG database for subsequent analyses.

Statistical analyses

All statistical analyses were performed using R (ver. 4.4.1) using the following packages effectsize⁹⁵, vegan⁹⁶, adiv⁹⁷, lme4⁹⁸, microniche⁹⁹, rfpermute¹⁰⁰, SRS¹⁰¹, glmm.hp¹⁰². For all analyses involving multiple comparisons, p-values were adjusted with Benjamin-Hochberg corrections.

Diversity analyses

We used Shannon’s H’ to estimate taxonomic (i.e., OTU) and functional (C-N-P-cycling genes) alpha diversity as it considers both richness and evenness. We used non-parametric tests and generalized linear mixed models (LMM) to test for differences in diversity between land-use stages. Specifically, we used Kruskal-Wallis tests to assess differences in diversity between grassland sites, and LMMs to test for differences between paired grassland and forest sites with sample pair and spatial distance between paired sites included as random factors. As forests are the putative successional endpoint of the abandoned grasslands, we set these as the reference level in the LMMs, with subsequent pairwise comparisons in a separate model without a set reference level to examine differences between grassland stages. Two-way ANOVA tests were used to assess whether taxonomic and functional diversity differed depending on differences in forest stand age and evenness, including their interaction. Differences in community composition (beta diversity) were tested using perMANOVA (10^⁴ permutations) on Bray-Curtis distances for taxonomic and functional matrices. To estimate the magnitude of the diversity effects (alpha, beta) between grassland and forest sites, we derived effect sizes (partial ω²) between paired grassland and forest sites for each grassland category. Effects of forest stand age and evenness on microbial taxonomic and functional diversity were assessed using two-way ANOVA models, including an interaction term of the two factors. Corresponding analyses for community and functional gene composition were done using perMANOVA analyses (10^⁴ permutations) based on Bray-Curtis distances. We used random forest models with combined permutation tests to extract the main covariates influencing microbial taxonomic and functional diversity, including community and functional composition. Based on this variable selection process, we next used hierarchical partitioning of fixed effects in the LMMs to partition the effects of the covariates on diversity variables. Prior to all diversity analyses, OTU matrices were normalized using scaling with ranked subsampling (SRS)¹⁰¹, with a minimum sequencing depth (cmin = 200 and 3000) for fungi and bacteria, respectively. Functional matrices were normalized by rarefying to minimum sequencing depth.

Assessing functional redundancy and specialization, and their links to carbon-cycling capacity

We used a set of approaches to examine the genetic redundancy related to nutrient cycling during succession. Firstly, taxonomic diversity (Shannon’s H’ of OTU tables) was used as a predictor of functional diversity (Shannon’s H’ of C-N-P-genes from metagenomes) in ordinary least-square regression (OLS) models. For these analyses, all grassland stages were pooled into a single grassland category to match the sample size of forest sites.

Significant fits between taxonomic and functional diversity was deemed indicative of low functional redundancy⁴⁹. This reasoning follows the assumption that functional redundancy ensues from the accumulation of taxa sharing the same effect (i.e., metabolic) trait¹⁰³. Removing taxa from the community will not immediately decrease the diversity of functional genes in systems with high redundancy. In systems with low redundancy, the opposite is true, and the two variables are consequently significantly related. We next used Levin’s niche overlap¹⁰⁴ (implemented in the Microniche package⁹⁹ to quantify the metabolic overlap of C-N-P genes across land uses, using rarefied metagenome matrices. Only genes above the default limit of quantification (LOQ = 1.65) were kept for subsequent analyses. We assumed that genes with high overlaps are occur widely within a given land use, akin to generalist taxa in classical niche theory, whereas low overlaps indicate specialization¹⁰⁴. The analyses were repeated using specific nutrient cycling pathways. We used OLS and second-order polynomial regressions to assess changes in average genetic overlap across successional stages, with pairwise Wilcoxon tests were used to test for differences in average overlap of substrates (C-cycling) and nutrient cycling pathways (N-P-cycling) between land use-stages. Differences in SIR between paired grassland and forest sites were tested through effect size measurements (Hedge’s d) for each assessed substrate. We then examined the relationship between the communities’ total carbon cycling capacity and diversity by regressing MSIR as response variable against taxonomic and functional diversity as predictor variables in OLS models.

Assessing trade-offs between functional redundancy and diversity

Trade-offs between community-level functional redundancy and diversity, as well as drivers of both measures, were examined with predicted bacterial metagenomes based on the framework developed in refs. ^38,39. We considered KO pathways related to C-N-P-cycling as traits and related these to the relative abundances of bacterial OTUs at each site. Specifically, we first scaled all pathways (range 0-1) by their minimum and maximum values and calculated the functional Euclidean distances between all pairs of OTUs. This step was done for each land-use stage separately to avoid the inclusion of OTUs not found within the meta-community. Functional distances were then scaled by division with their land-use specific maximum values. From this, functional diversity (Rao’s quadratic diversity Q), functional redundancy R, and the Simpson dominance index D were calculated using the adiv package⁹⁷. The position of each site along the axes of these three measures were assessed and visualized using ternary plots and perMANOVA tests (Bray-Curtis distances, 9999 permutations). Note that the limited availability of fungal genomes precludes similar analyses for the fungal dataset.

Reporting summary

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

Data availability

The sequence data generated in this study have been deposited in the NCBI database (amplicon sequences) under accession PRJNA1238768, and in SRA (shotgun sequences) under the project ID PRJEB56463. The 16S and ITS metabarcoding data generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) under the Bio Project accession number PRJNA994701. The carbohydrate active enzymes (CAZy) Database is available under (http://www.cazy.org/). The nitrogen cycling gene (NCyc) Database is available under (https://github.com/qichao1984/NCyc) (https://doi.org/10.1093/bioinformatics/bty741), the phosphorus cycling gene (PCyCDB) Database is available under (https://github.com/ZengJiaxiong/Phosphorus-cycling-database) (https://doi.org/10.1186/s40168-022-01292-1). The Evolutionary genealogy of genes: Non-supervised Orthologous Groups (EggNOG) Database is available under (http://eggnog5.embl.de/#/app/home). The Kyoto Encyclopedia of Genes and Genomes (KEGG) Database is available under (https://www.genome.jp/kegg/kegg1.html). Picrust2 (ver. 2.5.2) was used to predict bacterial metagenomes and is available under (https://huttenhower.sph.harvard.edu/picrust/). Data and code to reproduce the figures are available at: (https://zenodo.org/records/17176048).

Code availability

Data and code to reproduce all analyses and figures after bioinformatic processing are accessible on Github under the following repo: (https://github.com/tranheim/Succession_functional_diversity/tree/main) and (https://zenodo.org/records/17176048).

References

Winkler, K., Fuchs, R., Rounsevell, M. & Herold, M. Global land use changes are four times greater than previously estimated. Nat. Commun. 12, 2501 (2021).
Article PubMed PubMed Central ADS CAS Google Scholar
Newbold, T. et al. Global effects of land use on local terrestrial biodiversity. Nature 520, 45–50 (2015).
Article PubMed ADS CAS Google Scholar
Campbell, J. E., Lobell, D. B., Genova, R. C. & Field, C. B. The global potential of bioenergy on abandoned agriculture lands. Environ. Sci. Technol. 42, 5791–5794 (2008).
Article PubMed ADS CAS Google Scholar
Walker, L. R. & Moral, R. del. Primary Succession and Ecosystem Rehabilitation. (Cambridge University Press, 2003).
Horn, H. S. The ecology of secondary succession. Annu. Rev. Ecol. Syst. 5, 25–37 (1974).
Article Google Scholar
Clements, F. E. Nature and structure of the climax. J. Ecol. 24, 252–284 (1936).
Article Google Scholar
Simberloff, D. A succession of paradigms in ecology: essentialism to materialism and probabilism. Synthese 43, 3–39 (1980).
Article Google Scholar
Vera, F. W. M. et al. Grazing Ecology and Forest History. (CABI, Oxford, 2000).
Johnson, E. A. & Miyanishi, K. Testing the assumptions of chronosequences in succession. Ecol. Lett. 11, 419–431 (2008).
Article PubMed Google Scholar
Arias-Navarro, C., Baritz, R. & Jones, A. The State of Soils in Europe. https://doi.org/10.2760/7007291 (2024).
Eriksson, O., Cousins, S. A. O. & Bruun, H. H. Land-use history and fragmentation of traditionally managed grasslands in Scandinavia. J. Veg. Sci. 13, 743–748 (2002).
Article Google Scholar
Gustavsson, E., Lennartsson, T. & Emanuelsson, M. Land use more than 200 years ago explains current grassland plant diversity in a Swedish agricultural landscape. Biol. Conserv. 138, 47–59 (2007).
Article Google Scholar
Auffret, A. G., Kimberley, A., Plue, J. & Waldén, E. Super-regional land-use change and effects on the grassland specialist flora. Nat. Commun. 9, 3464 (2018).
Article PubMed PubMed Central ADS Google Scholar
Bremer, L. L. & Farley, K. A. Does plantation forestry restore biodiversity or create green deserts? A synthesis of the effects of land-use transitions on plant species richness. Biodivers. Conserv. 19, 3893–3915 (2010).
Article Google Scholar
Prangel, E. et al. Afforestation and abandonment of semi-natural grasslands lead to biodiversity loss and a decline in ecosystem services and functions. J. Appl. Ecol. 60, 825–836 (2023).
Article Google Scholar
Kindermann, E. et al. Resurveying inner-alpine dry grasslands after 70 years calls for integrative conservation efforts. Biol. Conserv. 289, 110393 (2024).
Article Google Scholar
Öckinger, E., Eriksson, A. K. & Smith, H. G. Effects of grassland abandonment, restoration and management on butterflies and vascular plants. Biol. Conserv. 133, 291–300 (2006).
Article Google Scholar
Moyano, J. et al. Unintended consequences of planting native and non-native trees in treeless ecosystems to mitigate climate change. J. Ecol. 112, 2480–2491 (2024).
Article Google Scholar
Quested, H., Eriksson, O., Fortunel, C. & Garnier, E. Plant traits relate to whole-community litter quality and decomposition following land use change. Funct. Ecol. 21, 1016–1026 (2007).
Article Google Scholar
Zhou, Z., Wang, C., Jiang, L. & Luo, Y. Trends in soil microbial communities during secondary succession. Soil Biol. Biochem. 115, 92–99 (2017).
Article CAS Google Scholar
Tian, J. et al. Ecological succession pattern of fungal community in soil along a retreating glacier. Front. Microbiol. 8, 1028 (2017).
Cline, L. C. & Zak, D. R. Soil microbial communities are shaped by plant-driven changes in resource availability during secondary succession. Ecology 96, 3374–3385 (2015).
Article PubMed Google Scholar
Egidi, E., Coleine, C., Delgado-Baquerizo, M. & Singh, B. K. Assessing critical thresholds in terrestrial microbiomes. Nat. Microbiol. 8, 2230–2233 (2023).
Article PubMed CAS Google Scholar
Allison, S. D. A trait-based approach for modelling microbial litter decomposition. Ecol. Lett. 15, 1058–1070 (2012).
Article PubMed CAS Google Scholar
Cortez, J., Garnier, E., Pérez-Harguindeguy, N., Debussche, M. & Gillon, D. Plant traits, litter quality and decomposition in a Mediterranean old-field succession. Plant Soil 296, 19–34 (2007).
Article CAS Google Scholar
Hernandez, D. J., Kiesewetter, K. N., Almeida, B. K., Revillini, D. & Afkhami, M. E. Multidimensional specialization and generalization are pervasive in soil prokaryotes. Nat. Ecol. Evol. 7, 1408–1418 (2023).
Article PubMed Google Scholar
Hättenschwiler, S., Tiunov, A. V. & Scheu, S. Biodiversity and litter decomposition in terrestrial ecosystems. Annu. Rev. Ecol. Evol. Syst. 36, 191–218 (2005).
Article Google Scholar
Hannula, S. E. & van Veen, J. A. Primer sets developed for functional genes reveal shifts in functionality of fungal community in soils. Front. Microbiol. 7, 1897 (2016).
Louca, S. et al. Function and functional redundancy in microbial systems. Nat. Ecol. Evol. 2, 936–943 (2018).
Article PubMed Google Scholar
Walker, L. R., Wardle, D. A., Bardgett, R. D. & Clarkson, B. D. The use of chronosequences in studies of ecological succession and soil development. J. Ecol. 98, 725–736 (2010).
Article Google Scholar
Cousins, S. A. O. & Eriksson, O. After the hotspots are gone: Land use history and grassland plant species diversity in a strongly transformed agricultural landscape. Appl. Veg. Sci. 11, 365–374 (2008).
Article Google Scholar
Mason, W. L. Changes in the management of British forests between 1945 and 2000 and possible future trends. Ibis 149, 41–52 (2007).
Article Google Scholar
Hooftman, D. A. P. & Bullock, J. M. Mapping to inform conservation: a case study of changes in semi-natural habitats and their connectivity over 70 years. Biol. Conserv. 145, 30–38 (2012).
Article Google Scholar
Mietkiewicz, N., Kulakowski, D., Rogan, J. & Bebi, P. Long-term change in sub-alpine forest cover, tree line and species composition in the Swiss Alps. J. Veg. Sci. 28, 951–964 (2017).
Article Google Scholar
da Silva Camilo, G., de Freitas Terra, B. & Araújo, F. G. Using the relationship between taxonomic and functional diversity to assess functional redundancy in streams of an altered tropical watershed. Environ. Biol. Fishes 101, 1395–1405 (2018).
Article Google Scholar
Chen, H. et al. Functional redundancy in soil microbial community based on metagenomics across the globe. Front. Microbiol. 13, 878978 (2022).
Bobay, L.-M. & Ochman, H. The evolution of bacterial genome architecture. Front. Genet. 8, 72 (2017).
Ricotta, C. et al. Measuring the functional redundancy of biological communities: a quantitative guide. Methods Ecol. Evol. 7, 1386–1395 (2016).
Article Google Scholar
Ricotta, C. et al. The ternary diagram of functional diversity. Methods Ecol. Evol. 14, 1168–1174 (2023).
Article Google Scholar
Douglas, G. M. et al. PICRUSt2 for prediction of metagenome functions. Nat. Biotechnol. 38, 685–688 (2020).
Article PubMed PubMed Central CAS Google Scholar
Sun, S., Jones, R. B. & Fodor, A. A. Inference-based accuracy of metagenome prediction tools varies across sample types and functional categories. Microbiome 8, 46 (2020).
Article PubMed PubMed Central CAS Google Scholar
Malard, L. A. & Guisan, A. Into the microbial niche. Trends Ecol. Evol. 38, 936–945 (2023).
Article PubMed CAS Google Scholar
Gubry-Rangin, C., Aigle, A., Herrera-Alsina, L., Lancaster, L. T. & Prosser, J. I. Niche breadth specialization impacts ecological and evolutionary adaptation following environmental change. ISME J. 18, wrae183 (2024).
Article PubMed PubMed Central CAS Google Scholar
Fang, J. et al. Threshold effects of soil pH regulate the biogeography of bacterial communities in inland wetlands across eastern China. Soil Ecol. Lett. 7, 240274 (2024).
Article Google Scholar
Garnier, E. et al. Plant functional markers capture ecosystem properties during secondary succession. Ecology 85, 2630–2637 (2004).
Article Google Scholar
Berthrong, S. T., Jobbágy, E. G. & Jackson, R. B. A global meta-analysis of soil exchangeable cations, pH, carbon, and nitrogen with afforestation. Ecol. Appl. 19, 2228–2241 (2009).
Article PubMed Google Scholar
Tscherko, D., Hammesfahr, U., Marx, M.-C. & Kandeler, E. Shifts in rhizosphere microbial communities and enzyme activity of Poa alpina across an alpine chronosequence. Soil Biol. Biochem. 36, 1685–1698 (2004).
Article CAS Google Scholar
Jobbágy, E. G. & Jackson, R. B. Patterns and mechanisms of soil acidification in the conversion of grasslands to forests. Biogeochemistry 64, 205–229 (2003).
Article Google Scholar
Fierer, N. et al. Reconstructing the microbial diversity and function of pre-agricultural tallgrass prairie soils in the United States. Science 342, 621–624 (2013).
Article PubMed ADS CAS Google Scholar
Inkpen, S. A. et al. The coupling of taxonomy and function in microbiomes. Biol. Philos. 32, 1225–1243 (2017).
Article Google Scholar
Wang, C. et al. Bacterial genome size and gene functional diversity negatively correlate with taxonomic diversity along a pH gradient. Nat. Commun. 14, 7437 (2023).
Article PubMed PubMed Central ADS Google Scholar
Zhang, H.-Y., Bissett, A., Aguilar-Trigueros, C. A., Liu, H.-W. & Powell, J. R. Fungal genome size and composition reflect ecological strategies along soil fertility gradients. Ecol. Lett. 26, 1108–1118 (2023).
Article PubMed Google Scholar
Bhatnagar, J. M., Peay, K. G. & Treseder, K. K. Litter chemistry influences decomposition through activity of specific microbial functional guilds. Ecol. Monogr. 88, 429–444 (2018).
Article Google Scholar
Berg, B. Litter decomposition and organic matter turnover in northern forest soils. Ecol. Manag. 133, 13–22 (2000).
Article Google Scholar
Silverstein, M. R., Bhatnagar, J. M. & Segrè, D. Metabolic complexity drives divergence in microbial communities. Nat. Ecol. Evol. 8, 1493–1504 (2024).
Article PubMed Google Scholar
Hernández, D. L. & Hobbie, S. E. The effects of substrate composition, quantity, and diversity on microbial activity. Plant Soil 335, 397–411 (2010).
Article Google Scholar
Dal Bello, M., Lee, H., Goyal, A. & Gore, J. Resource–diversity relationships in bacterial communities reflect the network structure of microbial metabolism. Nat. Ecol. Evol. 5, 1424–1434 (2021).
Article PubMed Google Scholar
Foote, R. L. & Grogan, P. Soil carbon accumulation during temperate forest succession on abandoned low productivity agricultural lands. Ecosystems 13, 795–812 (2010).
Article CAS Google Scholar
Bongers, F. J. et al. Functional diversity effects on productivity increase with age in a forest biodiversity experiment. Nat. Ecol. Evol. 5, 1594–1603 (2021).
Article PubMed Google Scholar
Biggs, C. R. et al. Does functional redundancy affect ecological stability and resilience? a review and meta-analysis. Ecosphere 11, e03184 (2020).
Article Google Scholar
Auber, A. et al. A functional vulnerability framework for biodiversity conservation. Nat. Commun. 13, 4774 (2022).
Article PubMed PubMed Central ADS CAS Google Scholar
Rocca, J. D. et al. Relationships between protein-encoding gene abundance and corresponding process are commonly assumed yet rarely observed. ISME J. 9, 1693–1699 (2015).
Article PubMed Google Scholar
Chen, J. & Sinsabaugh, R. L. Linking microbial functional gene abundance and soil extracellular enzyme activity: Implications for soil carbon dynamics. Glob. Change Biol. 27, 1322–1325 (2021).
Article ADS CAS Google Scholar
Cooke, R. S. C., Bates, A. E. & Eigenbrod, F. Global trade-offs of functional redundancy and functional dispersion for birds and mammals. Glob. Ecol. Biogeogr. 28, 484–495 (2019).
Article Google Scholar
Kerfahi, D. et al. Elevation trend in bacterial functional gene diversity decouples from taxonomic diversity. CATENA 199, 105099 (2021).
Article CAS Google Scholar
Bahram, M. et al. Structure and function of the global topsoil microbiome. Nature 560, 233–237 (2018).
Article PubMed ADS CAS Google Scholar
Ricotta, C., Pavoine, S., Cerabolini, B. E. L. & Pillar, V. D. A new method for indicator species analysis in the framework of multivariate analysis of variance. J. Veg. Sci. 32, e13013 (2021).
Article Google Scholar
Cousins, S. A. O. Analysis of land-cover transitions based on 17th and 18th century cadastral maps and aerial photographs. Landsc. Ecol. 16, 41–54 (2001).
Article Google Scholar
Tedersoo, L. et al. Global diversity and geography of soil fungi. Science 346, 1256688 (2014).
Article PubMed Google Scholar
Egnér, H., Riehm, H. & Domingo, W. R. Untersuchungen uber die chemische Bodenanalyse als Grundlage fur die Beurteilung des Nährstoffzustandes der Böden. II. Chemische Extraktionsmethoden zur Phosphor- und Kaliumbestimmung. K. Lantbrukshögskolans Ann. 26, 199–215 (1960).
Google Scholar
Kattge, J. et al. TRY – a global database of plant traits. Glob. Change Biol. 17, 2905–2935 (2011).
Article ADS Google Scholar
Roscher, C. et al. Interspecific trait differences rather than intraspecific trait variation increase the extent and filling of community trait space with increasing plant diversity in experimental grasslands. Perspect. Plant Ecol. Evol. Syst. 33, 42–50 (2018).
Article Google Scholar
Manrubia, M. et al. Soil functional responses to drought under range-expanding and native plant communities. Funct. Ecol. 33, 2402–2416 (2019).
Article PubMed PubMed Central Google Scholar
Bradford, M. A. et al. Thermal adaptation of soil microbial respiration to elevated temperature. Ecol. Lett. 11, 1316–1327 (2008).
Article PubMed Google Scholar
Birch, H. F. The effect of soil drying on humus decomposition and nitrogen availability. Plant Soil 10, 9–31 (1958).
Article CAS Google Scholar
Bongiorno, G. et al. Soil management intensity shifts microbial catabolic profiles across a range of European long-term field experiments. Appl. Soil Ecol. 154, 103596 (2020).
Article Google Scholar
Walters, W. et al. Improved bacterial 16S rRNA gene (V4 and V4-5) and fungal internal transcribed spacer marker gene primers for microbial community surveys. mSystems https://doi.org/10.1128/msystems.00009-15 (2015).
Tedersoo, L. & Lindahl, B. Fungal identification biases in microbiome projects. Environ. Microbiol. Rep. 8, 774–779 (2016).
Article PubMed Google Scholar
Tedersoo, L. et al. The Global Soil Mycobiome consortium dataset for boosting fungal diversity research. Fungal Divers. 111, 573–588 (2021).
Article Google Scholar
Özkurt, E. et al. LotuS2: an ultrafast and highly accurate tool for amplicon sequencing analysis. Microbiome 10, 176 (2022).
Article PubMed PubMed Central Google Scholar
Edgar, R. C., Haas, B. J., Clemente, J. C., Quince, C. & Knight, R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27, 2194–2200 (2011).
Article PubMed PubMed Central CAS Google Scholar
Bengtsson-Palme, J. et al. Improved software detection and extraction of ITS1 and ITS2 from ribosomal ITS sequences of fungi and other eukaryotes for analysis of environmental sequencing data. Methods Ecol. Evol. 4, 914–919 (2013).
Article Google Scholar
Edgar, R. C. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat. Methods 10, 996–998 (2013).
Article PubMed CAS Google Scholar
Hildebrand, F. et al. Dispersal strategies shape persistence and evolution of human gut bacteria. Cell Host Microbe 29, 1167–1176.e9 (2021).
Article PubMed PubMed Central CAS Google Scholar
Bahram, M. et al. Metagenomic assessment of the global diversity and distribution of bacteria and fungi. Environ. Microbiol. 23, 316–326 (2021).
Article PubMed CAS Google Scholar
Buchfink, B., Xie, C. & Huson, D. H. Fast and sensitive protein alignment using DIAMOND. Nat. Methods 12, 59–60 (2015).
Article PubMed CAS Google Scholar
Magoč, T. & Salzberg, S. L. FLASH: fast length adjustment of short reads to improve genome assemblies. Bioinformatics 27, 2957–2963 (2011).
Article PubMed PubMed Central Google Scholar
Huerta-Cepas, J. et al. eggNOG 5.0: a hierarchical, functionally and phylogenetically annotated orthology resource based on 5090 organisms and 2502 viruses. Nucleic Acids Res. 47, D309–D314 (2019).
Article PubMed CAS Google Scholar
Nayfach, S. & Pollard, K. S. Average genome size estimation improves comparative metagenomics and sheds light on the functional ecology of the human microbiome. Genome Biol. 16, 51 (2015).
Article PubMed PubMed Central Google Scholar
Levasseur, A., Drula, E., Lombard, V., Coutinho, P. M. & Henrissat, B. Expansion of the enzymatic repertoire of the CAZy database to integrate auxiliary redox enzymes. Biotechnol. Biofuels 6, 41 (2013).
Article PubMed PubMed Central CAS Google Scholar
Piton, G. et al. Life history strategies of soil bacterial communities across global terrestrial biomes. Nat. Microbiol. 8, 2093–2102 (2023).
Article PubMed CAS Google Scholar
Kanehisa, M., Furumichi, M., Tanabe, M., Sato, Y. & Morishima, K. KEGG: new perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res. 45, D353–D361 (2017).
Article PubMed CAS Google Scholar
Zeng, J. et al. PCycDB: a comprehensive and accurate database for fast analysis of phosphorus cycling genes. Microbiome 10, 101 (2022).
Article PubMed PubMed Central CAS Google Scholar
Tu, Q., Lin, L., Cheng, L., Deng, Y. & He, Z. NCycDB: a curated integrative database for fast and accurate metagenomic profiling of nitrogen cycling genes. Bioinformatics 35, 1040–1048 (2019).
Article PubMed CAS Google Scholar
Ben-Shachar, M., Lüdecke, D. & Makowski, D. effectsize: estimation of effect size indices and standardized parameters. J. Open Source Softw. 5, 2815 (2020).
Article ADS Google Scholar
Oksanen, J. et al. vegan: Community Ecology Package. Package version: 2.7-2. https://doi.org/10.32614/CRAN.package.vegan (2025).
Pavoine, S. adiv: An r package to analyse biodiversity in ecology. Methods Ecol. Evol. 11, 1106–1112 (2020).
Article Google Scholar
Bates, D. et al. Fitting Linear Mixed-Effect Models Using lme4. J. Stat. Softw. 67, 1–48 (2015)
Finn, D. R. et al. MicroNiche: an R package for assessing microbial niche breadth and overlap from amplicon sequencing data. FEMS Microbiol. Ecol. 96, fiaa131 (2020).
Article PubMed CAS Google Scholar
Archer, E. rfPermute: estimate permutation p-values for random forest importance metrics. Package version: 2.5-.5 https://doi.org/10.32614/CRAN.package.rfPermute (2025).
Beule, L. & Karlovsky, P. Improved normalization of species count data in ecology by scaling with ranked subsampling (SRS): application to microbial communities. PeerJ 8, e9593 (2020).
Article PubMed PubMed Central Google Scholar
Lai, J., Zhu, W., Cui, D. & Mao, L. Extension of the glmm.hp package to zero-inflated generalized linear mixed models and multiple regression. J. Plant Ecol. 16, rtad038 (2023).
Article Google Scholar
Ramond, P., Galand, P. E. & Logares, R. Microbial functional diversity and redundancy: moving forward. FEMS Microbiol. Rev. 49, fuae031 (2025).
Article PubMed CAS Google Scholar
Hurlbert, S. H. The measurement of niche overlap and some relatives. Ecology 59, 67–77 (1978).
Article Google Scholar

Download references

Acknowledgements

We wish to thank Anders Glimskär for his contribution with finding sites and providing metadata. Part of computations and data handling were enabled by resources provided by the National Academic Infrastructure for Supercomputing in Sweden (NAISS), partially funded by the Swedish Research Council through grant agreement no. 2022-06725. The following authors acknowledge funding enabling this research: TRS: C.F. Lundströms Stiftelse, CF2023-0019, Lars Hiertas Minne FO2021-0302; MB: Swedish University of Agricultural Sciences (early career grant), the Swedish Research Councils Formas (Grant 2020–00807), the Swedish Research Council (VR; Grant 2021–03724) and Novo Nordisk Foundation (NNF24OC0089849); FH & JF: were supported by the UKRI Biotechnology and Biological Sciences Research Council (BBSRC) Institute Strategic Programme Food Microbiome and Health BB/X011054/1 and its constituent project BBS/E/F/000PR13631, Decoding Biodiversity BBX011089/1 and its constituent work packages BBS/E/ER/230002 A and BBS/E/ER/230002B, FH as well as by European Research Council H2020 StG (erc-stg-948219, EPYC); JF: the BBSRC Norwich Research Park Biosciences Doctoral Training Partnership, BB/T008717/1.

Funding

Open access funding provided by Swedish University of Agricultural Sciences.

Author information

Authors and Affiliations

Department of Ecology, Swedish University of Agricultural Sciences, Ulls väg 16, Uppsala, Sweden
Tord Ranheim Sveen, Maria Viketoft, Jan Bengtsson, Joachim Strengbom & Mo Bahram
Department of Terrestrial Ecology, Netherlands Institute of Ecology, PO Box 50, 6700 AB, Wageningen, The Netherlands
Justine Lejoly & S. Emilia Hannula
Research Unit for Environmental Simulation, Helmholtz Zentrum München, German Research Center for Environmental Health, Ingolstaedter Landstraße 1, D – 85764, Neuherberg, Germany
Franz Buegger & Karin Pritsch
Gut Health and Microbes Program, Quadram Institute, Norwich Research Park, Norwich, UK
Joachim Fritscher & Falk Hildebrand
Decoding Biodiversity, Earlham Institute, Norwich Research Park, Norwich, UK
Falk Hildebrand
University of East Anglia, School of Biological Science, Norwich, UK
Falk Hildebrand
Department of Plant and Soil Sciences, University of Kentucky, Lexington, KY, USA
Ernest Osburn
Institute of Environmental Sciences, Leiden University, Leiden, The Netherlands
S. Emilia Hannula
Institute of Ecology and Earth Sciences, University of Tartu, Tartu, Estonia
Mo Bahram
Department of Agroecology, Aarhus University, Slagelse, Denmark
Mo Bahram

Authors

Tord Ranheim Sveen
View author publications
Search author on:PubMed Google Scholar
Maria Viketoft
View author publications
Search author on:PubMed Google Scholar
Jan Bengtsson
View author publications
Search author on:PubMed Google Scholar
Joachim Strengbom
View author publications
Search author on:PubMed Google Scholar
Justine Lejoly
View author publications
Search author on:PubMed Google Scholar
Franz Buegger
View author publications
Search author on:PubMed Google Scholar
Karin Pritsch
View author publications
Search author on:PubMed Google Scholar
Joachim Fritscher
View author publications
Search author on:PubMed Google Scholar
Falk Hildebrand
View author publications
Search author on:PubMed Google Scholar
Ernest Osburn
View author publications
Search author on:PubMed Google Scholar
S. Emilia Hannula
View author publications
Search author on:PubMed Google Scholar
Mo Bahram
View author publications
Search author on:PubMed Google Scholar

Contributions

T.R.S., M.V., J.B., S.E.H., J.S., and M.B. conceptualized the study. T.R.S., J.F., F.B., K.P., J.L., E.O., and F.H. collected the data. T.R.S. and M.B. analyzed the data. T.R.S., J.B., M.V., and M.B. wrote the manuscript. T.R.S., M.V., J.B., K.P., E.O., and M.B. reviewed and edited the manuscript. T.R.S. and M.B. revised the manuscript. M.B. supervised the study.

Corresponding author

Correspondence to Tord Ranheim Sveen.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Lei Deng, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Reporting Summary

Transparent Peer Review file

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Sveen, T.R., Viketoft, M., Bengtsson, J. et al. Functional diversity of soil microbial communities increases with ecosystem development. Nat Commun 16, 10408 (2025). https://doi.org/10.1038/s41467-025-66544-8

Download citation

Received: 10 January 2025
Accepted: 10 November 2025
Published: 22 November 2025
Version of record: 24 November 2025
DOI: https://doi.org/10.1038/s41467-025-66544-8