Arbuscular mycorrhizal fungi (AMF) influence the turnover of soil organic carbon (SOC). The understanding of the relationship between the structure of AMF community characteristics and SOC components is limited in the Robinia pseudoacacia plantation. The present study investigated soil AMF community composition, particulate organic carbon (POC), and mineral-associated organic carbon (MAOC) in the R. pseudoacacia plantation. The results showed that MAOC and SOC increased from CK to a25 and decreased thereafter (P < 0.05). The PCoA results showed that the AMF community structure significantly differed among 15-, 25-, 35-year-old, and CK. The stability of AMF communities showed a significantly increasing trend along with stand ages (P < 0.05). The changes of soil physical and nutrient properties jointly explained 77.00% of the variations in AMF community structure. Soil texture and N availability influence SOC through AMF. MAOC content is an important factor for SOC content. Our findings construct the foundation of scientific forest management strategies, while also providing a feasible approach to enhance carbon sequestration potential by fully exploiting the function traits of AMF.

Glomalin-related soil protein (GRSP), a microbially derived carbonaceous compound, plays a crucial role in sustaining ecosystem functions and promoting the accumulation and stabilization of soil organic carbon (SOC). However, information remains limited regarding GRSP accumulation within soil aggregates during natural forest restoration on the Loess Plateau, China. Based on this, this study along the restoration time sequence (farmland as the control; Populus davidiana, 110 years; P. davidiana- Quercus liaotungensis mixed, 130 years; Q. liaotungensis, 160 years), the accumulation characteristics of GRSP were evaluated for four soil aggregate size fractions (> 5 mm, 2-5 mm, 0.25-2 mm, and < 0.25 mm) by measuring soil GRSP content, soil properties (SOC, TN, DOC, pH), extracellular enzyme activity, and microbial diversity. The results showed that forest restoration markedly enhanced total GRSP (T-GRSP), easily extractable GRSP (EE-GRSP), and difficultly extractable GRSP (DE-GRSP) contents, with T-GRSP concentrations increasing as aggregate size decreased. With forest restoration, the mean weight diameter, geometric mean diameter, and water-stable aggregate content increased significantly. The activities of beta-1,4-glucosidase, cellobiohydrolase, and N-acetyl-beta-glucosaminidase showed an increasing trend; the contents of SOC, total nitrogen (TN), and dissolved organic carbon (DOC) responded positively. The results indicated that forest restoration drives GRSP accumulation in aggregates primarily through the synergistic effects of increasing the activity of nutrient-acquiring enzymes (BG, CBH, NAG) and improving the availability of soil carbon and nitrogen (SOC, TN, DOC). Overall, these results emphasized the importance of improved aggregate stability and the combination of microbial extracellular enzymes, soil environmental factors, and GRSP during forest restoration, providing a theoretical basis for evaluating soil quality improvement through vegetation restoration from a microbial-derived carbon perspective.

The application of biochar to agricultural soils is a promising strategy for reducing greenhouse gas emissions and enhancing carbon sequestration. However, the microbially mediated mechanisms by which biochar influences soil carbon cycling remain unclear. Through a meta-analysis, we evaluated the responses of soil organic carbon (SOC) and microbial communities to biochar application, aiming to elucidate microbial regulatory roles in biochar-induced carbon sequestration processes. Our results demonstrated that biochar significantly increased all SOC fractions (mean increase: 52.4

Leaf C:N:P stoichiometry provides a functional link between plant nutrient composition and environmental adaptation. However, whether widely distributed tree species maintain consistent leaf stoichiometric patterns across contrasting vegetation zones remains unclear. Here, we investigated the leaf C:N:P stoichiometry of Quercus variabilis, a widely distributed tree species in China, across the subtropical evergreen broad-leaved forest zone (SEF) and the warm-temperate deciduous broad-leaved forest zone (WDF). We aimed to clarify between-zone differences in leaf stoichiometric traits and the role of climate and soil constraints in shaping vegetation-zone-dependent stoichiometric variation. Our results showed that leaf C:N:P stoichiometry of Q. variabilis differed between the SEF and WDF, with significant differences detected only for leaf N and leaf C:N. Overall, leaf C varied little across zones (CV = 3.13-3.42%), leaf N and leaf C:N showed moderate variation (CV = 10.07-13.26%), whereas leaf P, leaf C:P, and leaf N:P showed greater variability (CV = 33.28-39.93%). Environmental analyses further indicated contrasting patterns between zones: leaf stoichiometry in the SEF was jointly associated with climate and soil factors and showed a positive leaf C-N relationship, whereas that in the WDF showed a stronger soil association and a negative leaf N-P relationship. These findings suggest that Q. variabilis may adjust its leaf stoichiometry in a zone-dependent manner in response to contrasting climate and soil conditions. This study provides a theoretical basis for understanding nutrient-use strategies and environmental adaptation of widely distributed tree species under heterogeneous climate and soil conditions.

Soil erosion is a primary driver of terrestrial carbon loss, particularly in the Loess Plateau of China, where intense erosion processes mobilize substantial quantities of soil organic carbon (SOC). While vegetation restoration is widely acknowledged as an effective approach to mitigate erosion and enhance SOC sequestration, its effectiveness in reducing SOC loss at the regional scale remains insufficiently quantified. This study synthesized the dataset from 183 sites comprising 1950 observations across the Loess Plateau to assess the effects of land-use changes on soil erosion and SOC loss. Additionally, key environmental factors including slope, slope length, precipitation, vegetation cover, and soil properties were analyzed to determine their relative impacts. Compared to cropland, areas undergoing vegetation restoration exhibited significantly reduced soil erosion and increased soil organic carbon content. The implementation of the "Grain for Green" Program (GGP), covering 5.6 x 106 ha of the Loess Plateau, resulted in an estimated annual reduction of approximately 1.7 Tg C of SOC erosion, accounting for about 11 % of the ecosystem carbon sequestration under the GGP of the Loess Plateau. These findings underscore the importance of tailoring restoration strategies to specific topographic and climatic conditions to maximize the dual benefits of soil conservation and carbon sequestration. The insights gained from this study provide a scientific basis for precision ecological management and policy development in the Loess Plateau and other similarly degraded regions.

Soil multifunctionality (SMF) is critical for sustaining ecosystem functions. Crop rotations are widely recognized for enhancing soil supporting and provisioning services. However, the trade-offs between these services, and the mechanisms by which long-term rotations improve SMF, remain unclear. To address this gap, we evaluated SMF in both topsoil and subsoil, as well as crop yields and economic benefits, after 39 years of contrasting crop rotation systems in dryland farmland. Our results showed that incorporating legume-based crops, particularly legume forages such as sainfoin and alfalfa, significantly increased SMF in both topsoil and subsoil (p < 0.05). Nevertheless, distinct trade-offs and synergies between economic benefits and soil functions were observed among different rotation patterns. Although cereal-based rotations produced higher short-term economic benefits, they contributed little to soil function improvement. In contrast, well-designed legume forage-based rotations promoted concurrent improvement in SMF and economic benefit. Moreover, soil carbon and nutrient cycling emerged as key drivers of SMF under long-term rotation, with strong implications for economic sustainability. Overall, our findings demonstrate that long-term crop rotation, primarily through enhanced soil nutrient availability, improves both SMF and economic benefit. Integrating forage legume-based crops into wheat rotation represents an effective strategy for strengthening soil supporting services and provisioning economic benefits in dryland agroecosystems.

Understanding the temperature sensitivity (Q10) of soil organic matter (SOM) decomposition is essential for predicting soil carbon (C) emissions under global warming. However, how land-use change alters Q10 and the underlying microbial and biochemical mechanisms remains uncertain, largely due to complex interactions among SOM quality, microbial traits, and Q10. In this study, a total of 144 topsoil and subsoil samples were collected from paired croplands and adjacent afforested lands across eight sites in China. Continuous warming incubation experiments were conducted to determine the Q10 of SOM decomposition. Results showed that warming-induced carbon (C) mineralization was substantially higher in C-poor soils than in C-rich soils, indicating greater vulnerability of C-poor soils to C loss. In topsoil, C mineralization efficiency in croplands and afforested lands of C-poor soils was 437% and 254% higher, respectively, than that in C-rich soils. Importantly, afforestation reduced Q10 in C-poor soils by 13.2% in topsoil and 15.0% in subsoil compared with croplands. This reduction in Q10 was primarily driven by decreased soil organic carbon (SOC) recalcitrance and a lower abundance ratio of recalcitrant- to labile-C degradation genes. Climate exerted a strong influence on Q10 in topsoil but had weaker effects in subsoil. These results indicate that initial soil C status fundamentally mediates microbial and biochemical regulation of SOM decomposition under warming. Therefore, incorporating initial SOC status into ecosystem models can improve predictions of soil C dynamics under climate warming, while prioritizing afforestation in C-poor soils may help mitigate warming-induced SOC loss and support soil conservation.

Microbial residue carbon (MRC) plays a critical role in soil carbon cycling and stabilization under global climate change. However, how vegetation restoration affects MRC accumulation remains unclear. In this study, we compiled 817 field observations to investigate the effects of vegetation restoration on MRC accumulation at a global scale. The results showed that vegetation restoration increased MRC, fungal residue carbon (FRC), and bacterial residue carbon (BRC) by 39.9%, 46.6%, and 34.3%, respectively, while the FRC/BRC ratio remained relatively stable. Vegetation restoration enhanced the contribution of FRC to soil organic carbon by 8.3%, highlighting its crucial role in carbon stabilization. Grassland restoration led to a greater increase in MRC (+52.1%) compared with natural succession (+36.6%) and afforestation (+24.8%). In addition, vegetation restoration in arid regions led to larger increases in both MRC (+44.1%) and FRC (+57.6%) than in humid regions (+35.3%, +35.2%, respectively). Collectively, our results reveal that fungal residues played a key role in soil organic carbon accumulation, which is primarily regulated by soil pH, the fungal-to-bacterial ratio, and nutrient availability following vegetation restoration. These findings clarify how soil pH regulates fungal pathways to facilitate carbon sequestration during vegetation restoration.

Soil biodiversity plays an important role in sustaining soil multifunctionality (SMF) by enhancing interactions in soil micro-food webs. However, how trophic interactions among soil biota regulate the soil micro-food web and consequently drive SMF under changing environments remains unclear. Here, we investigated how soil micro-food web interactions shape SMF along an elevational gradient in the Qinling Mountains, China. The results revealed that the SMF index exhibited a unimodal pattern with increasing elevation, peaking at 2000 m. Variations in SMF were primarily driven by the abundance of top Module 1, network complexity, and positive interactions within the soil micro-food web. The protistan community of Module 1 is the critical regulator of the soil micro-food web, and 42.6% of the key nodes in the protist community were predatory protists, which associated with the bacterial community change through top-down effects. This top-down regulation enhanced the positive cohesion of the soil micro-food web, thereby increasing network complexity and ultimately promoting SMF. In conclusion, our findings highlight the crucial role of soil micro-food web complexity in sustaining multiple soil functions along an elevational gradient in the Qinling Mountains.

Long-term agricultural cultivation accelerates soil erosion and exacerbates water and soil organic matter loss. Vegetation restoration is widely recognized as an effective strategy for improving soil carbon (C) sequestration. However, the effects of vegetation restoration following farmland abandonment on soil organic carbon stock (SOCS) and its fractions across soil profiles remain poorly understood. Here, we investigated the dynamics of SOCS, particulate organic carbon stock (POCS), mineral-associated organic carbon stock (MOCS) and their driving factors across a 0-100 cm soil profile over a 170-year chronosequence following farmland abandonment on the Loess Plateau of China. The results showed that 170 years of vegetation restoration increased SOCS and POCS by 41.5% and 81.2%, respectively. In the 20-100 cm layer, SOCS and POCS initially declined and began to increase after approximately 30 years of vegetation restoration. MOCS remained unchanged in the 0-60 cm layer, whereas increased in the 60-100 cm layer after vegetation restoration. The 40-60 cm layer served as a "transitional zone", marking a shift in dominant sequestration mechanisms from POCS-dominated accumulation in the topsoil (0-40 cm) to MOCS-dominated stabilization in the subsoil (60-100 cm). SOCS and POCS were driven by plant-derived inputs, while MOCS was mainly controlled by soil physical properties and microbial activity. Our findings demonstrate that long-term vegetation restoration effectively enhances soil organic carbon and particulate organic carbon accumulation in surface soils while promoting the stabilization of mineral-associated organic carbon in subsoil, highlighting the importance of vegetation restoration for sustainable soil C management.

Microbial carbon use efficiency (CUE), a key link between microbial metabolism and ecosystem carbon cycling, is influenced by litter input, nutrient availability, and microbial diversity during vegetation succession. Although microbial carbon use efficiency (CUE) has been widely studied, most existing work has focused on short-term disturbances, laboratory incubations, or surface soils, leaving the depth-specific mechanisms of CUE regulation during long-term vegetation succession insufficiently understood. Here, this study investigated an similar to 170-year vegetation succession chronosequence from grassland to climax forest, analyzing microbial community structure, function, and CUE in topsoil (0-20 cm) and subsoil (20-40 cm) using amplicon sequencing, high-throughput qPCR microarrays, and biogeochemical stoichiometric modeling. The results revealed topsoil CUE remained lower than subsoil CUE consistently during vegetation succession. Topsoil CUE showed an initial increase, followed by a decline and rebound, while subsoil CUE increased initially, then steadily declined. In topsoil, CUE was mainly driven by nutrient availability, indirectly influenced by litter quality, microbial diversity, and microbial products (e.g., biomass, extracellular enzymes). In subsoil, CUE was more directly regulated by nutrients and functional gene abundance (e.g., rbcL, nosZ2, nirS3, nirK2). CUE was shaped by resource inputs (e.g., dissolved organic C and N, litter C:N) and environmental conditions (pH, moisture). Despite pronounced shifts in microbial composition, CUE remained relatively stable in early and late succession. These findings highlight depth-specific microbial processes controlling CUE and provide mechanistic insights into soil carbon stabilization during long-term vegetation succession.

Nitrogen (N) deposition alters the composition and release of plant root exudates, thereby influencing the dynamics of soil organic carbon (SOC). However, the effects of specific root exudate compounds on SOC fractions under different N levels remain unclear. In this study, we conducted an incubation experiment to investigate how oxalic acid, citric acid, D-tryptophan, D(+) maltose, and p-hydroxybenzoic acid interact with different N addition levels (0, 90, 180, and 270 kg N ha−1) in affecting SOC fractions. The results revealed that N addition altered SOC through changes in soil physicochemical properties and microbial activity, reduced soil microbial biomass C, and increased CO2 emissions. The addition of root exudates resulted in an average 2.5

ABSTRACT Litter decomposition and microbial activity are important factors affecting soil organic carbon (SOC) sequestration in the terrestrial ecosystems. However, how litter and microorganisms influence soil carbon (C) sequestration remains under debate, which hampers the understanding of C sequestration mechanisms. This study conducted a microcosm experiment by adding 13 C‐labeled litter from the grazed and fenced grasslands to clarify litter‐derived C dynamics and trace the C flow from litter to SOC pools. The results showed that the litter decomposition rate and litter‐derived C incorporated in dissolved organic C (DOC) in the fenced grasslands was lower than that in the grazed grasslands. Meanwhile, litter decomposition in the fenced grasslands formed a larger proportion of mineral‐associated organic C (MAOC, 8%) and microbial biomass C (MBC, 1%), compared with the grazed grasslands (3% and 0.3%). Hydrolase was positively correlated with the litter‐derived C in both particulate organic C (POC) and MAOC, whereas the oxidase: hydrolase ratio showed a negative correlation. Collectively, fenced grassland increased the percentage of litter‐derived C incorporation into both labile and stable SOC pools because of the higher exoenzymes than the grazed grasslands. These findings highlight the key roles of microbial exoenzymes in litter decomposition and demonstrate that microbial activity is conducive to SOC formation.

Land use changes accelerate the carbon (C) cycle, yet their impact on soil inorganic C (SIC), particularly pedogenic carbonates (PC), is often overlooked. Here, we investigated the response of soil C pools to land use changes in the Loess Plateau, and elucidated the PC formation mechanisms. For broader context, a global meta-analysis was conducted focusing on the joint response of soil organic C (SOC) and SIC to land use changes. Converting croplands to forests resulted in a trade-off between SOC gains and SIC losses at both local and global scales. The SIC loss in forests was primarily derived from lithogenic carbonate (LC, 16 Mg C ha-1). These LC losses were partly compensated by the neoformation of PC in soils at 40-100 cm depth. Converting croplands to orchards increased SIC stocks, primarily due to PC formation (23 Mg C ha-1). More than 51 % of the CO2 fixed during PC formation originated from microbial respiration. Organic fertilizers or plant residues input stimulated microbial activity, generating more CO2 and subsequently HCO3-. The HCO3-then formed PC in the presence of Ca2+ and Mg2+ in soil, which are prerequisites for PC formation. Organic fertilizer-derived Ca2+ and Mg2+ contributed to PC accumulation in orchards. In forests, the porous soil structure facilitated Ca2+, Mg2+, and HCO3-leaching into the deep soil layers, thereby generating more PC at 40-100 cm depth. This study revealed the instability of SIC, and first quantified soil CO2 fixed by land use-induced PC. To establish accurate carbon sink accounting methodologies, future research should clarify the contribution of SIC dynamics to atmospheric CO2 removal.

Large-scale ecological restoration has substantially enhanced vegetation greening and carbon sequestration across China, yet increasing atmospheric dryness may intensify conflicts between carbon gain and water consumption. Understanding the long-term dynamics of vegetation carbon-water coupling is therefore critical for evaluating the sustainability of ecological restoration under climate change. Here, we integrated long-term satellite observations, interpretable machine learning, and structural equation modelling to quantified the spatiotemporal evolution of carbon use efficiency (CUE), water use efficiency (WUE), and their coupling relationships across China from 1982 to 2018, and to identify the dominant drivers and pathways regulating these dynamics. We found that ecological restoration promoted vegetation carbon sequestration, with a marked shift in CUE from a declining to an increasing trend around 2000 and continued greening projected for 58% of the study area. In contrast, WUE exhibited an inverted-U trajectory and is projected to decline across 64.47% of vegetated regions. The trade-off between CUE and WUE intensified substantially over the past four decades, increasing by 58.33%, particularly in humid and semi-arid ecosystems. Atmospheric dryness, represented by increasing vapor pressure deficit (VPD), together with hydrothermal conditions dominated by solar radiation, emerged as the primary drivers of carbon-water decoupling and exhibited pronounced nonlinear threshold responses. Carbon-water trade-offs were strongest in regions characterized by increasing VPD (-0.0017–0.0036), low solar radiation (<0.17 W m-2), and elevations below 1,150 m. Moreover, changes in WUE exerted a stronger influence on coupling dynamics than changes in CUE, highlighting the central role of water regulation in shaping ecosystem responses to climate change. These findings provide a mechanistic basis for incorporating carbon-water trade-offs into climate-adaptive ecological restoration and sustainable water-resource management.

Deep soil organic carbon is a crucial but overlooked sink in large-scale vegetation restoration, as assessments restricted to shallow soils systematically underestimate the true carbon sequestration. Here, a systematic synthesis of 0-500 cm soil organic carbon database for the Loess Plateau was compiled, and machine learning was applied to reconstruct historical carbon sequestration since the Grain for Green Program. Based on this, future carbon sequestration potentials were projected for 2030 and 2060 under various emission scenarios (SSP126, SSP245, and SSP585) and quantified the spatial differentiation of climate vulnerability along the aridity gradient. The results estimated that the Grain for Green Program has generated a cumulative increase of 198.5 Tg soil organic carbon (0-500 cm) across 5.26 × 106 ha of restored land, of which about 61% occurred in 0-200 cm soils. Future 0-500 cm soil carbon sequestration potential continues to increase, with gains of 1.5-7.5 Tg during 2020-2030 and 7.9-14.0 Tg during 2030-2060, reaching 215.3 Tg by 2060 under the high emission scenario SSP585. Soil depth and clay content regulated the pattern of Climate vulnerability in the whole Loess Plateau. Herein, the semi-arid and semi-humid zones forms a climate-resilience core, where carbon sequestration remains stable or continues to increase under future climate change. These findings provided a robust quantitative baseline for the true carbon sequestration capacity of dry land ecosystems and offer an operational scientific basis for land use conversion to support carbon neutrality under global climate change.

Increasing aridity reduces soil biodiversity, and destabilizes microbial networks. However, systematic comparisons of microbial community structure and co-occurrence patterns in natural versus artificial ecosystems across climatic aridity gradients remain scarce. This study examined bacterial communities across 12 grassland and farmland sites spanning three climatic zones of the Loess Plateau in China, exploring the distribution of sensitive functional groups and their responses to aridity. Results showed that both grassland and farmland ecosystems exhibited lower bacterial network complexity and stability with increasing aridity gradient. Farmland communities demonstrated heightened sensitivity to drought than grassland communities, as indicated by steeper declines in species richness with both latitude and elevation. Indicator species analysis identified 861 and 829 sensitive ASVs (Amplicon Sequencing Variants) in grasslands and farmlands, respectively, with 159 shared between ecosystems. At the genus level, most sensitive Actinobacteriota (Nocardioides, Microlunatus) were negatively correlated with soil organic carbon and total nitrogen (P < 0.05). Multiple sensitive groups of Proteobacteria (Achromobacter, Lysobacter, Steroidobacter) and Nitrospirota (Nitrospira) were positively correlated with soil moisture and macronutrient contents, as well as the functional potentials of carbon, nitrogen and phosphorus cycling genes (P < 0.05). In terms of their relative abundance and quantity, the sensitive bacterial groups were highly susceptible to loss as drought severity increased. This study underscores that climatic zonation explained three times more variation in soil bacterial communities across the Loess Plateau than land-use pattern. The findings provide new insights into the mechanisms linking bacterial communities and functions in dryland soils.

Rhizosphere metabolites play a pivotal role in plant-soil-microbe interactions, yet the absence of standardised sampling method presents challenges for data comparability and a unified understanding of rhizochemistry dynamics. In this study, the rhizosphere metabolites of spring wheat (Triticum aestivum L.) were systematically evaluated using three commonly used sampling methods - rhizosphere soil extraction, hydroponic collection, and in situ sampling. We assessed the differences in metabolite abundance, diversity, composition, and metabolic pathways among the three methods. The results revealed that rhizosphere soil extraction method is simple and repeatable, but it tends to underestimate the rare metabolic features compared to the other two methods. The hydroponic collection method offered a more comprehensive rhizochemical profile, however, it may not reflect the actual soil environment in which plants grow. The in situ method mostly captured the rhizochemical dynamics under natural conditions but was technically complex and less reproducible. Pathway enrichment of differential metabolic features showed a distinct pattern, with hydroponic emphasising plant-centric primary metabolism, and rhizosphere soil extraction and in situ methods reflecting pathways more associated with environmental interaction and ecological adaptation. Overall, our results highlight that the sampling methods significantly influences the abundance and composition of the detected rhizochemical profile, thus the rhizosphere metabolites sampling methods should be considered according to the experimental objectives of future research.