Plant community attributes play a crucial role in maintaining the functioning of terrestrial ecosystems. However, the relative contributions of different plant community attributes to ecosystem multifunctionality (EMF), and their mediating roles in linking climate and soil factors remain unclear, especially in harsh, natural alpine grasslands. We comprehensively evaluated the effects of community taxonomic and functional composition, and plant taxonomic, functional, and phylogenetic diversity, together with climate and soil parameters, on EMF (assessed using both average and threshold approaches based on eight functions) across 95 alpine grasslands covering a 3000 km transect on the Tibetan Plateau. We identified the pathways through which plant community attributes mediated EMF, as well as the threshold-dependent changes in these relationships. Results showed that plant community attributes outweighed climate and soil conditions in explaining ecosystem multifunctionality. Among multiple plant community attributes, community taxonomic composition played a larger role in maintaining mean EMF. Community taxonomic composition and functional composition (community weighted mean traits) predominantly mediated the effects of climate and soil on mean EMF. In addition, we found that plant community attributes became increasingly important at higher EMF thresholds, whereas the influence of climatic factors declined. Communities characterized by conservative resource-use strategies supported greater multifunctionality at the highest threshold. Overall, our research provides important regional insights of plant community attributes–multifunctionality relationships from alpine grasslands, and highlights urgent conservation implications for preserving and improving community composition to maximize ecosystem multifunctionality in alpine grasslands.

Increasing frequency and intensity of heat and drought events under anthropogenic climate change has triggered widespread forest mortality and canopy dieback worldwide. Previous studies have suggested that such dieback events are often preceded by a marked decline in forest resilience, yet the underlying mechanisms driving this pre-dieback decline remain unclear. Here, we combined multi-scale satellite observations of vegetation dynamics and canopy functional traits to trace resilience trajectories preceding the 2022/2023 extreme heatwave-drought dieback event in northern China’s temperate forests. We found that forests experiencing canopy dieback exhibited a pronounced increase in the standard deviation (SD) of vegetation indices starting approximately a decade prior to dieback, indicating amplified ecosystem variability. Meanwhile, the contrasting trajectories of SD and lag-1 autocorrelation (AR1) revealed a trade-off between ecosystem stability and recovery dynamics, with forests becoming less stable but maintaining faster recovery capacity before dieback. This trend of declining stability was closely associated with higher pre-dieback productivity promoted by favorable climatic conditions, such as warmer and wetter periods. This suggests that while favorable climates temporarily boost productivity, they also amplify ecosystem sensitivity to subsequent climatic extremes. Our findings reveal that apparent greening under a warming climate does not necessarily indicate recovery, but may instead serve as an early warning signal of reduced forest stability and elevated dieback risk. These insights underscore the importance of integrating high-resolution satellite observations with long-term field monitoring to disentangle the mechanisms linking enhanced productivity to declining stability and to improve predictions of forest vulnerability under accelerating climate change.

Snow exerts intricate influences on alpine ecosystems, and winter snow process is undergoing drastic change with climate warming. Yet the impacts of winter snow on vegetation growth (hereafter, snow effects) and their underlying mechanisms remain uncertain, as snow effects operate through multiple pathways. Based on remote sensing-derived snow and vegetation products, we examine the snow effects on vegetation growth in spring and summer across the Tibetan Plateau through partial correlation analysis. We then quantify the distinct pathways through which snow influences vegetation growth in spring and summer using pixel-wised piecewise-Structural Equation Modeling, and further identify the transitions of dominant ecological processes underlying these effects. We report a marked seasonal transition of snow effects on vegetation growth, with negative-to-positive shifts observed in 25% of the Tibetan Plateau from spring to summer. The negative snow effects in spring are mainly attributable to the snowmelt-date (SMD)-induced phenological pathway, through which earlier snowmelt advances spring phenology and thereby influences vegetation growth. In contrast, snow effects in summer are predominantly positive due to the reversal of snow-induced phenological effects from spring and the carryover of snow-driven soil moisture effects into summer, which promotes vegetation growth. Finally, we found that most of the dynamic global vegetation models (DGVMs) used in this study have limited ability to reproduce this seasonal transition in snow effects. These findings highlight the critical seasonal shifts in ecological processes that underpin snow effects on ecosystems, providing valuable insights into improving ecosystem models.

The Qinghai-Tibet Plateau (QTP), as the world’s Third Pole, has experienced a sharp trend of warming and humidifying in recent decades within the context of global climate change. Under these circumstances, the soil nutrient cycling has been significantly modified. However, whether the soil quality had changed over recent decades on QTP remains unclear. In this study, we calculated, and drew high-resolution maps of the soil quality index (SQI) on QTP using soil property data in 1980s and 2020s. Results showed significantly an increasing trend of soil quality over the past four decades. The variation in trends of four typical ecosystems, more specifically, alpine cushion vegetation showing the most notable increase, followed by alpine meadow and alpine steppe, while forest & shrub exhibiting the smallest increase in soil quality. This implied that regions with more significant SQI changes correspond to stronger climate change. Through attribution analysis of 10 factors affecting the variation in SQI, including climate, soil hydrothermal, vegetation, and human activities, we found that the improvement in soil quality were predominantly driven by soil hydrothermal conditions. Additionally, we observed a pronounced increase in nitrogen limitation, with the factors contributing to this trend varying across different ecosystems. Our study emphasizes the warning of intensified nitrogen limitation, and, under the context of climate change, this phenomenon is likely to become increasingly severe in the future.

Formation-level vegetation mapping is pivotal for understanding ecological processes and mechanisms, as it reveals the distribution of dominant species that shape ecosystem structure and dynamics. However, fast and accurate formation-level mapping over large geographic areas is often hindered by the lack of robust mapping frameworks, limited field survey data, and unscientific or inefficient division of vegetation patches. To address these challenges, we proposed an automated mapping framework that integrates multi-source data for formation-level vegetation mapping. Our approach introduced an innovative strategy for automatically delineating vegetation patches based on slope units, improving mapping efficiency and ensuring results align more closely with actual vegetation distribution. Additionally, we developed a crowdsource-based vegetation survey system that aggregates data from diverse sensors, significantly increasing the sample size and diversity of vegetation formations. Using this framework, we successfully mapped 16 formations in Beijing with an overall accuracy of 65.7%, achieving F-scores exceeding 60% for major formations. The result indicates that Beijing’s vegetation is dominated by forests and shrublands, with the largest vegetation formation being Vitex negundo (deciduous broadleaf shrubland), covering 20% of the city in the southwestern mountains, followed by Quercus mongolica (deciduous broadleaf forest), occupying 10% in the northwestern mountains. This study provides a solid foundation for understanding Beijing’s vegetation distribution and its ecological functions. By integrating remote sensing and crowdsourced data, it demonstrates an effective approach for precise, large-scale formation-level vegetation mapping, offering valuable support for refined ecological management and interdisciplinary research.

The increasing frequency of drought and heatwave events under climate warming has led to widespread reports of severe forest mortality. However, there is still a lack of comprehensive and quantitative analysis on the shared patterns of climate anomalies, phenology, vegetation factors, soil properties, and topography linked to forest mortality pulses. In 2022 and 2023, North China experienced an unprecedented extreme heatwave-drought event, resulting in widespread forest canopy dieback (FCD). Here, we applied the latest remote sensing forest monitoring algorithms, Detection and Classification - Spectral Mixture Analysis (CCDC-SMA) and Fusion Near Real-Time (FNRT), to identify FCD during the heatwave-drought event in North China, with further validation through landscape images captured by drones. Additionally, we utilized machine learning techniques to quantify the key drivers of FCD throughout the progression of the heatwave-drought event. Our study shows that this event was highly severe, widespread, and prolonged, causing historically low anomalies in regional greenness and productivity. The 2022/2023 heatwave and drought event caused approximately 83,765 hectares of FCD, with the most severe occurrence in May 2023 (33,080 hectares), followed by June 2023 (14,190 hectares). In the early stage (September and October 2022), FCD was more severe in areas with higher forest cover and a greater proportion of plantations, suggesting that afforestation efforts may have inadvertently increased forest vulnerability. Additionally, in the mid-stage (May and June 2023), an earlier end of season (EOS) in 2022 was associated with severe FCD, suggesting that shortened growing periods and reduced carbon assimilation may have weakened tree resilience, increasing vulnerability to drought-induced mortality. These findings emphasize the complex interactions among climate extremes, forest structure, and phenology in driving FCD. We suggest future forest restoration efforts focus on structural diversity and local adaptability to reduce vulnerability to climate extremes.

Enhanced atmospheric deposition of reactive nitrogen (N) could alter the N cycle in terrestrial ecosystems. However, the fate of deposited N under different levels of N deposition is still not fully understood, and the long-term dynamics and redistribution of deposited N within steppe ecosystems remain uncertain. Using 15N tracers, we investigated N fate across varying N addition rates (0, 2, 5, 10, and 25 g N m-2 yr-1) over 1 month, 1 year and 8 years in a temperate steppe in Inner Mongolia, China. Under ambient N deposition, over 94 % and 87 % of added 15N was retained after 1 month and 1 year, respectively, with soil acting as the primary sink (>50 %). The N saturation effect was primarily observed in plants, with N retention stabilizing or declining above 5 g m-2 yr-1. The N addition also altered the distribution of soil N, with most added nitrate accumulating in deeper soil layers (20-30 cm), and enhancing microbial N retention in the short term. Over eight years, the overall 15N recovery dropped to less than 21 % across different N addition levels. Additionally, N addition significantly modified the inter-annual variation in N retention. Our study shows that N deposition significantly affects N retention and distribution in a steppe ecosystem. Added N was mainly retained in the soil, with plants showing signs of N saturation. Nitrate largely accumulated in deeper soil layers. The plant-litter-soil continuum faces risks of N loss under sustained high deposition, threatening ecosystem N sustainability.

Canopy uptake of NO x (NO and NO 2 ) is an important pathway for removing soil‐emitted and atmospheric NO x , especially in forested areas. Deposition velocity of NO and NO 2 varies among plant species and the existence of the compensation point, the concentration value when net exchange flux of NO or NO 2 is zero, remains uncertain. Additionally, leaf‐level uptake of NO and NO 2 in subtropical forests in China has received little attention despite this area suffering from heavy nitrogen pollution. To help address the knowledge gap, leaf‐level exchange of NO and NO 2 between needles of mature Pinus massoniana and the atmosphere was monitored in situ in a subtropical forest in Southwest China. The results showed that the total uptake of NO and NO 2 by the canopy reached 1.03 ± 0.31 kg N m −2 (Land) yr −1 , of which 93.9% was contributed by NO 2 . The NO deposition velocity was below detection limit. For NO 2 it ranged between 0.028 and 1.2 cm s −1 . Distinct NO 2 uptake was observed (0.035 cm s −1 ) during summer droughts when stomatal conductance was expected to be small. No consistent NO 2 emissions or uptake were detected at low NO 2 concentrations (<3 ppbv), indicating the need of controlled environments to conclusively determine the existence of a compensation point. Due to species‐specific NO and NO 2 deposition velocities and potential prolonged drought in East Asia, more long‐term monitoring of leaf‐level NO and NO 2 exchanges for other subtropical tree species, especially during droughts, is needed to precisely constrain the regional NO x budget.

Soil wind erosion is one of the key earth surface processes in arid and semi‐arid regions. Soil wind erosion not only leads to land desertification but also serves as an important source of fine particulate matter in the atmospheric environment. Accurate assessment of soil wind erosion and its temporal and spatial distributions is critical for planning and implementing soil conservation measures. As an important factor in wind erosion control, vegetation coverage has been included in almost all the major wind erosion models. The traditional models, however, usually overestimate wind erosion rate because they rely solely on photosynthetic vegetation coverage (PVC) but overlook nonphotosynthetic vegetation coverage (NPVC), such as fallen leaves and branches covering and protecting the soil. In the current study, field surveys, phenological data and fractional vegetation coverage (FVC) derived from the normalized difference vegetation index (NDVI) were employed to examine the temporal and spatial evolution of both PVC and NPVC in the wind erosion region on the Qinghai–Xizang Plateau (QXP). The results reveal significant variations in phonology across QXP. During 2000–2020, the growing season started on Julian Days 124–150, i.e., corresponding to the last month of spring, and ended on Julian Days 242–296, i.e., covering almost the first half of autumn, in the wind erosion‐prone areas of QXP. The vegetation greening initially began in the northern basins and southern river valleys with lower elevations and higher air temperatures, followed by the plateau areas with higher elevations and lower temperatures. Whereas, an opposite trend was manifested in the evolution of senescence. Approximately 40.7% of the area in arid, semi‐arid and extreme arid regions had never been observed greening. Owing to the combined influence of topography and climate, the vegetation coverage exhibited a decreasing trend from the southeast to the northwest of QXP. The mean annual FVC during the growing and nongrowing seasons were 36.2% and 24.4%, respectively. During the growing season, moreover, the FVC was approximately 1.04–1.37 times greater than the PVC. Regarding the interannual trend, the vegetation coverage increased from 2000 to 2020 in general. The mean annual FVC over the entire study region increased by 0.15% and 0.14% during the growing and nongrowing seasons, respectively, over the past 20 years. The temporal trend, however, varied among different areas. During the growing season, FVC remained basically unchanged in 37.2%, experienced mild improvements in 42.0% and underwent mild degradations in 20.8% of the study region. These findings hold important implications for understanding soil wind erosion processes and improving wind erosion models on QXP.

The leaching of dissolved organic nitrogen (DON) signals N loss in forest ecosystems, connecting terrestrial and aquatic ecosystems. The response of DON to varied inorganic N (Nin) deposition remains unclear. A 16-year continuous monitoring of DON in throughfall, soil water, and stream water was conducted under a field N manipulation (10-year Nin addition and subsequent cessation of Nin addition) in a subtropical forest in China. Under reference conditions, the average organic N (Nor) deposition was 23.0 ± 4.9 kg N ha-1 yr-1, with approximately one-third leaching at a soil depth of 30 cm, a value that was higher than those reported for temperate forests. Over a 16-year period, DON concentrations tended to increase in throughfall, soil solutions and stream water. The long-term increase in DON concentrations in soil solutions was primarily driven by increasing Nor deposition and indirectly influenced by decreasing acid deposition. Due to a doubling of monthly Nin input, either as sodium nitrate or as ammonium nitrate, the soil DON leaching doubled, causing the soil to transition from a DON sink to a DON source. After Nin addition ceased, the DON leaching returned to natural levels, and the soil reverted to being a DON sink. This study reveals the importance of Nor deposition and decreasing acid deposition as drivers of long-term trends in DON concentrations in soil solutions, in addition to elucidating the response patterns of DON to increasing and decreasing Nin deposition.

Soil wind erosion is one of the major ecological and environmental issues in arid, semi-arid and partly semihumid regions, and serves as an important driver of the migration and turnover of soil organic matter (SOM). The soil organic matter loss via wind erosion is typically calculated based on SOM of the bulk soil sample collected in the source areas, which usually remarkably deviates from the actual SOM budget. To address this discrepancy, it is imperative to examine SOM differentiation across particle size fractions and its spatial distribution in wind erosion regions. A total of 40 soil samples were collected in the wind erosion region of northeastern China and dry-sieved into eight size fractions, i.e., 0-63, 63-100, 100-150, 150-200, 200-250, 250-300, 300-400, and 400-880 mu m, and SOM was measured for each fraction. The particle size-specific SOM predictive models were developed, and the spatial distribution of SOM in each fraction was simulated for the wind erosion region of northeastern China. No matter for which size fraction, the predictive model exhibited strong performance and stability based on the three key factors, i.e., SOM in bulk soil (SOMbulk), soil clay content (CLAY) and temperature (T), which was capable of explaining over 75 % of the SOM variation. The simulated spatial mean SOM showed a general decreasing trend from the 0-63 mu m to the 300-400 mu m fractions, followed by a slight increase in the 400-880 mu m fraction. Specifically, the area proportion of SOM larger than 30 g center dot kg- 1 decreased markedly from 45.44 % in the 0-63 mu m fraction to 16.07 % in the 300-400 mu m fraction, whereas the proportion of SOM below 10 g center dot kg- 1 increased from 8.50 % to 22.21 %. These changes were mainly distributed in western Hulunbuir City, western Xilin Gol League, and the Horqin sandy land, where soils are subject to wind erosion, characterized by low clay content and weak aggregation. In contrast, the area proportion of SOM exceeding 20 g center dot kg- 1 rose from 37.96 % in the 300-400 mu m fraction to 43.57 % in the 400-880 mu m fraction. This trend was primarily spotted in the semi-humid zones, where well-developed aggregates promoted SOM accumulation. The findings provide a crucial foundation for accurate quantification of wind erosion-driven SOM loss and in-depth understanding of land degradation mechanisms in northeastern China.

Dasiphora fruticosa widely distributes in subalpine and alpine regions in the world, and is originated from Qinghai-Tibet Plateau. Further exploration of influence of environmental factors on plant functional traits of Dasiphora fruticosa in Qinghai-Tibet Plateau is essential to predict the growth and distribution under climate change more accurately. The Dasiphora fruticosa shrub on 24 plots were sampled at the altitude gradient of 2550-5200 meters above sea level on Qinghai-Tibet Plateau. Totally 13 plant functional traits of Dasiphora fruticosa were measured, including morphological traits (plant height, crown width, specific leaf area and leaf dry matter content) and stoichiometric traits (carbon, nitrogen and phosphorus content in leaves, flowers and stems). The results showed that morphological traits of Dasiphora fruticosa changed significantly along altitude. With the increase of altitude, plant height and crown width tended to be shorter. Leaf dry matter content also decreased along altitude. The stoichiometric traits varied along altitude, and were different in organs. Specifically, the phosphorus content in organs increased significantly along altitude. The morphological traits had large coefficient of variation. Soil properties were the main drivers of most of plant functional traits. Soil properties significantly directly affected the morphological traits and carbon and phosphorus contents while mean annual precipitation significantly indirectly affected them by affecting soil nutrients. Dasiphora fruticosa adapts to diverse habitats by adjusting its morphological traits and phosphorus content in organs. Soil properties have a stronger influence and act as a direct filter on plant functional traits of Dasiphora fruticosa in alpine regions.

Over the past few decades, terrestrial ecosystems have experienced rising atmospheric nitrogen (N) deposition, which further impacts the global carbon (C) budget through soil microbial respiration (MR). However, the effects of N deposition on MR are rarely characterized in deep soil (depth > 10 cm) rather than in surface soil (0–10 cm). This study attempted to elucidate how N deposition regulates MR along the soil profile and its underlying mechanism. We collected soil samples and determined MR across three soil layers (shallow, medium, and deep) from a decade-long and five-level N addition experiment in a temperate steppe in Inner Mongolia. We further used structural equation modeling to explore how long-term N addition regulates MR through various biotic (plant attributes and microbial community structure) and abiotic (soil properties) factors across the three soil layers. The overall response of MR to N addition varied with soil depth, shifting from stimulation in the shallow soil layer (standardized total effect of 0.36) to inhibition in the medium and deep soil layers (-0.34 and − 0.31). The identified direct and indirect pathways by which N addition regulates MR significantly differed across soil layers. Our results found that the N addition effect on soil C decomposition varied across different soil layers and involved distinct mechanisms in the temperate grassland. As soil depth increases, the suppressive effect of N deposition on MR provides evidence that increasing N deposition may contribute to C accrual in deep soil in grassland ecosystems.

Alpine meadow occupies similar to 46.7 % of the grassland area and stores more than 60 % of the soil organic carbon (SOC) in the Qinghai-Tibet Plateau, with 3.19 Pg C in the top 30 cm. Parent material can highly influence SOC stocks via various mineral components with divergent adsorption abilities. However, previous studies mainly concentrated on the relation between SOC and parent material in temperate and tropical forest ecosystems, evidences from different environments are needed to gain a holistic understanding. In this study, we explored how the functional distinct SOC fractions (the particulate (POC), mineral-associated (MAOC), and hot-water extractable carbon (HWEOC)) vary among four types of parent materials (monzonitic granite (MG), slate (SL), muscovite schist (SH), and diorite (DI)) by measurements from 16 sites on the Qinghai-Tibet Plateau. We also assessed the plant and geochemical factors in controlling the divergent SOC fractions. We found that MAOC and HWEOC vary strongly, while POC remain stable among parent materials due to sufficient plant input and slow decompositions rate. Moreover, our results revealed that geochemical factors exert dominant controls on each SOC fraction, with soil pH and texture showing a significant influence. Soils developed from the SL displayed the highest MAOC and HWEOC compared with that formed on the other three parent materials due to finer soil texture as well as relative high multivalent Ca and Mg. Furthermore, root traits increased its control on SOC fractions in relatively deeper soil depth (10-30 cm). Taken together, these results provided evidence of the key role of parent material in controlling the divergent SOC fractions in the less weathered alpine regions, emphasizing that parent material and soil geochemistry should be adequately considered in the biogeochemical and Earth system models.

To understand the effects of atmospheric sulfur (S), nitrogen (N), and heavy metal (HM) deposition on streamwater chemistry in Buha River (BHR), the upstream of Qinghai Lake (QHL) on Qinghai-Tibet Plateau (QTP), water, mosses, and soil were collected from three representative small watersheds. Mosses had low S contents (0.09 +/- 0.05%), slightly high N, zinc (Zn), and lead (Pb) contents (1.36 +/- 0.46%, 78.7 +/- 19.4 mgkg(-1), and 10.5 +/- 0.83 mgkg(-1)), and quite high copper (Cu) contents (27.3 +/- 4.65 mgkg(-1)), compared with the global background values. The delta N-15 in mosses and soil Pb isotopic composition revealed the contribution of coal combustion to soil pollution via atmospheric deposition. The average SO42- and NO3- concentrations of streamwater were 21.9 and 1.74 mgL-1. Elevated concentrations and increasing ratios of SO42-/Cl- and NO3-/Cl- compared with historical levels indicated a significant impact of atmospheric deposition. However, the average Zn, Pb, and Cu concentrations in streamwater were 6.99 +/- 2.36, 0.52 +/- 0.87, and 2.22 +/- 1.50 mu gL-1, respectively, with very limited effects of HM deposition. The high pH of streamwater provided by the chemical weathering of carbonate rocks prevented water acidification and transport of HMs from soil to surface waters. Given QTP's ecological fragility and potential climate changes, the significance of long-distance atmospheric transportation to surface water still mattered.

Aims Nitrogen (N) deposition alters plant stoichiometry and the state of nutrient limitation in forest ecosystems. In subtropical forests of China, responses of plants stoichiometry to chronically elevated and recently declined N deposition remain unclear. Methods A 14-year investigation of needle elements, including N, phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg) and aluminum (Al), was carried out under field N manipulation (10-year N addition and subsequent cessation of N addition) in a typical, subtropical Masson pine ( Pinus massoniana Lamb.) forest in southwest China. Results Needle N and P concentrations were significantly elevated by N addition even after N addition ceased while needle N:P remained constant among treatments. Needle N:K and P:K were significantly elevated by N addition, but were no longer different from reference plots after N addition ceased. Masson pine’s annual relative growth rate (RGR) was significantly decreased by N addition but the differences among treatments became insignificant among treatments after N addition ceased. Additionally, the annual RGR was negatively correlated with needle Al concentration. Needle Al exhibited a trend of first rising and then decreasing over the 14-year experiment. Conclusions Under the background of high ambient N deposition and Ca deposition, chronic N addition does not aggravate P limitation but can result in K limitation in a N-saturated subtropical forest. With the recovery of subtropical forest from soil acidification and N saturation, the elevated needle N and P induced by N addition will persist for at least four years and most likely for much longer in the future, but the risk of K limitation is expected to diminish. Graphical abstract

In response to climate change, China is making great efforts to increase the green area for carbon sequestration. Road verges, as marginal land with favorable conditions for plant growth and ease of transportation, can be used for biomass production, but the biomass production and carbon sequestration potential have not been assessed. Here, we mapped the biomass production potential of road verges in China by combining a biomass model and Geographic Information System and then evaluated the effect of road runoff and CO2 fertilization on the production according to the runoff coefficient and vehicle emission inventory. Nationwide, road verges can produce 15.86 Mt C yr-1 of biomass. Road runoff contributes to a biomass production of 1.26 Mt C yr-1 through increasing soil water availability, which mainly occurs in arid regions. The CO2 fertilization effect by vehicle emission is considerable in Eastern and Southern China, contributing to a production of 0.09 Mt C yr-1. Life cycle assessment shows that major road verges in China have a carbon sequestration potential of 6.87 Mt C yr-1 currently. Our results revealed that road verges can make a significant contribution to carbon neutrality under proper management.

Abstract Background and aims Over the past few decades, terrestrial ecosystems have experienced rising atmospheric nitrogen (N) deposition, which further impacts the global carbon (C) budget through soil microbial respiration (MR). However, the effects of N deposition on MR are rarely characterized in subsoil (depth > 10 cm) rather than in topsoil (0–10 cm). This study attempted to elucidate how N deposition regulates MR along the soil profile and its underlying mechanism. Methods We collected soil samples and determined MR across three soil layers (shallow, medium, and deep) from a decade-long and five-level N addition experiment in a temperate steppe in Inner Mongolia. We further used structural equation modeling to explore how long-term N addition regulates MR through various biotic (plant attributes and microbial community structure) and abiotic (soil properties) factors across the three soil layers. Results The overall response of MR to N addition varied with soil depth, shifting from stimulation in the shallow soil layer (standardized total effect of 0.36) to inhibition in the medium and deep soil layers (-0.34 and − 0.31). The identified direct and indirect pathways by which N addition regulates MR significantly differed across soil layers. Conclusion As soil depth increases, the suppressive effect of N deposition on MR provides evidence that increasing N deposition may contribute to C accrual in the subsoil in grassland ecosystems.