Distributions and Drivers of Soil Organic Carbon on the Tibetan Plateau: Divergent Controls Across Ecosystems

Distributions and Drivers of Soil Organic Carbon on the Tibetan Plateau: Divergent Controls Across Ecosystems | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Distributions and Drivers of Soil Organic Carbon on the Tibetan Plateau: Divergent Controls Across Ecosystems Zhigang Hu, Jens-Christian Svenning, Jiwang Tang, Yue Cheng, Wanben Wu, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6651952/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 02 Oct, 2025 Read the published version in Plant and Soil → Version 1 posted 6 You are reading this latest preprint version Abstract Background The Tibetan Plateau (TP), a vast alpine region with substantial soil organic carbon (SOC) stocks, plays a crucial role in the regional carbon cycle and climate change mitigation. However, our understanding of the spatial distribution and underlying drivers of SOC across TP remains limited. Methods We quantified the horizontal (1 km resolution) and vertical (0–200 cm depth) patterns in SOC density (SOCD) across the TP using 1,561 soil samples and a fully trained and validated Random Forest (RF) model. Twenty-six environmental variables were evaluated to determine their influence on SOCD. Correlation analysis and Structural Equation Modeling (SEM) were employed to examine the associations and causal mechanisms between these factors and SOCD. Results SOCD decreased from east to west and south to north, with mean values of SOCD ranging from 14.96 kg C m⁻² in western steppes to 29.54 kg C m⁻² in eastern forests at 0-200 cm depth. Vertically, SOCD exhibited a nonlinear pattern, initially increasing and then decreasing with depth. SOCD was highest in wetlands, followed by forests, meadows, shrublands, and steppes. Conclusions SOCD was influenced by multiple interacting environmental factors, primarily through their effects on plant productivity and soil respiration. Different ecosystems exhibit distinct regulatory mechanisms for SOCD. Forests are dominated by carbon loss through respiration, while grasslands and other ecosystems rely more on plant-derived carbon inputs. These findings enhance our understanding of distinct mechanisms regulating soil organic carbon across ecosystems, supporting improved modeling of soil carbon-climate feedbacks and informing ecosystem-specific carbon management strategies. Soil organic carbon Spatial distribution Driving factors Random Forest Tibetan Plateau Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Being the largest terrestrial organic carbon pool on Earth (Jobbágy and Jackson, 2000; Lehmann et al., 2020), soil harbors more carbon than the atmosphere and all vegetation combined (Georgiou et al., 2022). This vast pool not only regulates climate–carbon feedbacks (Bradford et al., 2016; Davidson and Janssens, 2006) but also underpins ecosystem productivity and resilience (Li et al., 2022a; Xu et al., 2019). As early as the 1880s, scientists recognized that soil organic carbon (SOC) drives both fertility and plant growth (Dokuchaev, 1883; Jobbágy and Jackson, 2000; Tian et al., 2015). Today, enhancing SOC is more than just a soil health goal—it’s a nature-based solution for mitigating climate change (Lehmann et al., 2020; Meyer et al., 2018) and sustaining critical ecosystem services (Prietzel et al., 2016; Stockmann et al., 2015). SOC content reflects the long-term balance between inputs from net primary production, such as surface litter and root exudates, and outputs from soil physicochemical processes, including decomposition and leaching of dissolved organic carbon (Jobbágy and Jackson, 2000; Li et al., 2022a; Plante et al., 2014). Climate changes, vegetation dynamics, environmental variation, and anthropogenic activities (Lehmann and Kleber, 2015) all influence the input-output processes, rendering soil organic carbon, particularly in topsoil, highly sensitive to biotic and abiotic factors (Gaitán et al., 2019; Li et al., 2022a). Even small changes in SOC can influence the broader carbon cycle and contribute to climate feedbacks (Ding et al., 2017; Ma et al., 2024; Ramesh et al., 2019). However, our comprehension of the spatial distribution of SOC and the underlying driving factors remains restricted, particularly at large regional scales. Alpine soils, characterized by high SOC density due to low temperatures that suppress decomposition and biomass turnover, play a critical role in the regional carbon balance and climate feedbacks (Davidson and Janssens, 2006; Wang et al., 2023). As the world’s “Third Pole” and the “Water Tower of Asia,” the Tibetan Plateau (TP) harbors extensive alpine ecosystems and vast permafrost soils that have accumulated substantial SOC stocks over millennia, profoundly shaping the regional carbon cycle and climate system (Genxu et al., 2008; Wang et al., 2023; Yao et al., 2012). Since 1970, the TP has warmed rapidly at 0.35°C per decade, exceeding the global average and comparable to the Arctic warming rates (Kuang and Jiao, 2016; Yao et al., 2019), making it one of the most climate-sensitive and ecologically fragile regions globally (Hu et al., 2023). Climate warming accelerates SOC decomposition and disrupts the regional carbon balance, while human activities such as overgrazing further intensify SOC loss (Jiang et al., 2024; Zhang et al., 2017; Zhu et al., 2023). Given its vast carbon storage and strong climate sensitivity, the TP is a critical region for understanding SOC distribution and its environmental drivers, with far-reaching implications for both regional carbon management and global carbon–climate feedbacks through processes like permafrost thaw and ecosystem change (Chen et al., 2022; Wang et al., 2023; Yang et al., 2008; Yao et al., 2019). However, many previous studies have focused on individual sites, specific regions, or single vegetation types, limiting their capacity to capture broader spatial trends (Ding et al., 2016; Wang et al., 2023). Given the TP’s complex topography and climatic gradients, its geographic environment exhibits pronounced spatial heterogeneity in factors such as elevation, temperature, and vegetation types (Jia et al., 2023; Wang et al., 2023; Zhang et al., 2015). This heterogeneity gives rise to substantial variations in SOC distribution and its controlling mechanisms across different ecosystems and soil profiles (Gao et al., 2023; Wang et al., 2023). For example, the influence of plant productivity, soil respiration, or moisture availability may differ markedly between alpine steppes and forests, or between the topsoil and subsoil layer (Han et al., 2022; Li et al., 2022a). As a result, findings derived from site-level studies often fail to adequately capture these differences and represent regional-scale patterns. Additionally, while key variables like temperature and precipitation are frequently examined, other potentially important drivers, such as topography, vegetation productivity or soil, are often underrepresented (Jiang et al., 2024). Since SOC dynamics result from the complex interplay of biotic and abiotic factors, examining a broader and ecologically relevant set of environmental variables is essential to disentangle their relative contributions and interactions (Wang et al., 2023), and to better inform Earth system models. Therefore, it is necessary to extensively collect regional soil samples and include a wider range of variables to compensate for the limitations of the site-specific studies. In this study, we compiled soil carbon data from 1,561 sampling sites collected between 2001 and 2021. These samples span major vegetation types across the TP, including forests, shrublands, grasslands, and wetlands. Using this extensive and spatially representative dataset, we analyzed the horizontal and vertical distribution of SOC density (SOCD) from the surface to 200 cm depth. Furthermore, we assessed the influence of 26 environmental factors—ranging from solar radiation, climate, and topography to soil properties and vegetation indices—to systematically investigate the drivers of SOCD across this highly heterogeneous alpine region. Specifically, this study aimed to: (1) quantify the spatial distribution features of SOCD across the TP in both horizontally and vertically dimensions; (2) determine the relative importance of multiple environmental drivers of SOCD, and (3) test the hypothesis that both the magnitude and environmental controls of SOCD differ across vegetation types and soil depths. Materials and methods Study area The Tibetan Plateau, commonly referred to as the “Third Pole” with an average elevation above 4000 m, covers more than a quarter of China’s total land area, rendering it the highest and largest plateau on Earth (Piao et al., 2012; Zhu et al., 2023). The annual average air temperature in most areas of the plateau remains below 0°C, with the average monthly air temperature ranging from -11.6°C in January to 9.1°C in July during the period from 2000 to 2020 (Hu et al., 2023; Wang et al., 2024). Generally, on the TP, temperature and precipitation decrease while altitude increases from southeast to northwest (Yin et al., 2013) (Fig. 1b,c). Annual average precipitation varies from 100 mm per year in the northwest to over 1000 mm per year in the southeast (Chen et al., 2021) (Fig. 1b,c). The plateau is mainly covered by steppe, meadow, forest, and shrubland (Fig. 1a). Data collection and processing Soil organic carbon data In this study, we collected and compiled 1561 soil organic carbon samples (Fig. 1a, Supplementary Table S1). Among these, 236 samples were newly acquired by our research team during field investigations in the summers of 2020 and 2021. During the investigations, sampling was conducted along the northern Tibetan grassland transect from east to west, with three replicates collected per site across a meadow-to-steppe gradient (see Supplementary Table S1 for data and sampling depths). After collection, soil samples were air-dried at ~25°C, manually disaggregated and sieved through a 2 mm mesh to remove coarse fragments. A subsample was then ground to a fine powder using a ball mill and stored in sealed containers until analysis. SOC was quantified using dry combustion with a LECO CN analyzer (LECO Corporation, USA), following the standard protocol described by Nelson and Sommers (1996), and validated through comparable studies using LECO dry combustion methods (Kowalenko, 2001; Wright and Bailey, 2001). An additional 1325 samples were compiled from 110 peer-reviewed studies conducted between 2001 and 2021 (see Supplementary Table S1 for data information and references), identified through comprehensive searches on platforms such as Google Scholar and Web of Science. These legacy datasets capture the primary vegetation and soil types of the TP with some representativeness but have uneven sampling point distribution (Fig. 1a) due to challenging geographical conditions. Sampling depths and data units were standardized across datasets, outliers were removed, and only datasets with explicitly reported sampling depths and georeferenced locations were included, with duplicated sites excluded. Samples with only SOC content were transformed into SOC density data using the following formula (Ding et al., 2019; Homann et al., 2007; Yang et al., 2010). Where density (kg m −2 ), SOC content (g kg −1 ), bulk density (g cm −3 ), thickness (cm), and volume percentage of the coarse fragment at layer i , respectively (Ding et al., 2019; Yang et al., 2008). BD and CF values were extracted from SoilGrids 2.0 (Poggio et al., 2021). We used measured data from 460 samples to establish strong linear relationships between SOCD at different depth intervals (see Supplementary Table S2 for details), enabling us to estimate missing values and obtain complete carbon density data at six depth intervals (0-10 cm, 10-20 cm, 20-30 cm, 30-50 cm, 50-100 cm, 100-200 cm) for all samples. Soil properties Soil physical properties, including bulk density (BD), coarse fragments (CF), and soil texture (sand, silt, clay), as well as soil chemical properties such as cation exchange capacity (CEC), nitrogen (N), and pH, were extracted from SoilGrids data (Poggio et al., 2021). Soil moisture (SM) data was derived from a daily soil moisture dataset over China (Li et al., 2022b), which was based on site observations and obtained from the National Tibetan Plateau Data Center (https://data.tpdc.ac.cn/home). Soil heterotrophic respiration (SRh) data, derived from the Soil Respiration Database Version 5 (SRDB-V5) using a quantile regression forest model, were extracted from a global dataset available at the Oak Ridge National Laboratory Distributed Active Archive Center (https://daac.ornl.gov/). Topography and climate Digital elevation (Elev) data (https://lta.cr.usgs.gov/GTOPO30) with a resolution of 30 meters was acquired and used for the extraction of slope and aspect via ArcGIS. Precipitation, air temperature, and radiation (Rad) data were collected from the National Meteorological Information Center (http://data.cma.cn/) and interpolated into raster data using ANUSPLIN (Hutchinson, 2004). The interpolated data explained over 90% of the actual observed data according to validation results (Li et al., 2019). The MODIS data of the land surface temperature (LST) and imagery data for calculating the land surface water index (LSWI) were obtained from the Land Processes Distributed Active Archive Center (https://lpdaac.usgs.gov/). Annual maximum and minimum temperature data were downloaded from the National Tibetan Plateau Data Center (https://data.tpdc.ac.cn/home). Evapotranspiration (ET) data (Zheng et al., 2022) and potential evapotranspiration data (Peng, 2022) were obtained from the National Tibetan Plateau Data Center (https://data.tpdc.ac.cn/home). The aridity index (AI), calculated as the ratio of potential evapotranspiration to precipitation (Thornthwaite, 1948), was used to assess regional dryness and its potential impact on SOC. Vegetation information The Normalized Difference Vegetation Index (NDVI), enhanced vegetation index (EVI), and net primary productivity (NPP) estimates were downloaded from the Land Processes Distributed Active Archive Center (https://lpdaac.usgs.gov/). Leaf area index (LAI) data was obtained from the National Centers for Environmental Information (https://www.ncei.noaa.gov/). Vegetation type data was obtained from the vegetation atlas of China, available at the Resource and Environment Science and Data Center of the Chinese Academy of Sciences (https://www.resdc.cn/Default.aspx). Data extraction For data sources in raster format, values were extracted based on the coordinates of sampling points to match the SOC data. To ensure temporal alignment between environmental variables and SOC data, variables exhibiting significant interannual variability, such as climate factors and vegetation indices, were extracted based on the year of soil sample collection. Soil physicochemical variables were extracted from gridded datasets at standard depths (0–10 cm, 10–20 cm, 20–30 cm, 30–50 cm, 50–100 cm, 100–200 cm) to match SOC data. When direct depth data were not available, we used linear interpolation between neighboring depth layers to estimate values. Dimensionality reduction of variable data To assess the relative influence of different types of variables on SOCD, the 26 environmental factors used in this study were grouped into eight categories (Fig. 6) based on their attributes: topography (Topog), soil physical properties (SPP), soil chemical properties (SCP), soil heterotrophic respiration (SRh), climate, solar radiation (Rad), vegetation-related indices (VRI), and net primary productivity (NPP). For dimensionality reduction, we first conducted individual linear regressions between the variables in each category and SOCD. The resulting regression coefficients were then used as weights to aggregate the variables into composite indices for each category (e.g., climate). This procedure yielded category-level factors while preserving the relative contribution of individual variables. Random Forest model and validation Random Forest (RF) is widely used for digitally mapping SOC (Lamichhane et al., 2019; Yang et al., 2023). In this study, RF was employed to map SOCD at different depths and evaluate the importance of factors influencing SOCD. The ‘randomForest’ package (Liaw and Wiener, 2002) in the R environment (R Core Team, 2013) was utilized. To ensure robust model performance, we randomly split the dataset into training (80%) and testing (20%, n = 312) subsets. The random split was stratified by vegetation type to preserve class distribution across subsets. Based on SOCD at different depths and corresponding soil physicochemical properties data, along with other variables, RF models were separately trained. 5-fold cross-validation was used to select the optimal parameters for the RF models. Statistical indices of the coefficient of determination (R 2 ), mean absolute error (MAE), and root mean square error (RMSE) were used to quantify the generalization capability of the RF models. The models for SOCD estimation at depths of 0–30 cm, 30–50 cm, 50–100 cm, and 100–200 cm achieved R² values of 0.61, 0.64, 0.66, and 0.67, respectively (mean R² = 0.65) (Fig. 2). The models showed acceptable generalization capability across depths, though a slight underestimation was observed for high SOCD values likely due to sample scarcity in that range (Fig. 2). Statistical analysis The partial correlation method was utilized to assess the correlation between pairs of variables. In exploring variations in SOCD across distinct vegetation types, the one-way analysis of variance (ANOVA) was implemented. Significance in statistical analysis was evaluated through T-tests or F-tests, with the level of significance indicated by the p-value. Piecewise fitting was conducted on SOCD at various depth intervals to uncover vertical distribution characteristics. Structural Equation Modeling was employed to test hypothesized causal relationships between environmental factors and SOCD, based on ecological knowledge and observed correlations. The model fit was assessed using common statistical criteria, including the chi-square test (χ²), and comparative fit index (CFI), with acceptable model fit indicated by χ²/df 0.9. The SEM analysis was conducted using the lavaan package in R (Rosseel, 2012). Results Spatial distribution characteristics of SOCD Vertical distribution Our findings indicate that in superficial soil layers (above 20cm), there was a slight increase in SOCD per 10 cm thickness (SOCD p10 ) with depth, whereas in deeper soil layers (below 20cm), there was a non-linear decrease in SOCD p10 (Fig. 3). According to the piecewise fitting analysis, the relationship between SOCD p10 and soil depth (Dep) is as follows (Fig. 3b): Horizontal distribution Given the minor differences in SOCD across the 0–10 cm, 10–20 cm, and 20–30 cm soil layers, as shown above, these depths were combined into a single 0–30 cm interval in subsequent analyses to enhance figure clarity and improve the overall readability of the manuscript. Based on the validated RF models and the grid data corresponding to the 26 variables used for model training, the SOCD distributions at four different depths were predicted. The results revealed that SOCD p10 exhibited similar spatial patterns at different depths across the TP, with generally higher values in the eastern and southern regions compared to the west and north(Fig. 4). These patterns were further supported by the 1561 field-measured SOC observations collected in this study (Fig. A.1a). Specifically, SOCD p10 at a depth of 200 cm exhibited an increasing trend along the longitudinal gradient, indicating higher SOCD values in eastern regions relative to western areas (Fig. A.1a). Along the latitudinal gradient, a general decline in SOCD with increasing latitude was observed, although the pattern was not entirely uniform (Fig. A.1b). Notably, higher SOCD levels were detected in the southern regions compared to the northern ones, aligning with the model predictions. Differences in SOCD distribution across vegetation types and depths The distribution of SOCD p10 across six vegetation types—wetland, forest, shrubland, meadow, steppe, and others—was meticulously analyzed across four soil depth intervals (Figs. 1a and 5). Here, “others” refers to other vegetated land types not included in the primary categories. Wetlands and forests exhibited the highest SOCD p10 , followed by meadows and shrublands, while steppes and other vegetation types had comparatively lower levels of SOCD p10 (Fig. 5). This pattern of variation in SOCD p10 among different vegetation types remained consistent across varying depth intervals (Fig. A.2). The SOCD p10 exhibited significant differences across various depth intervals, with deeper intervals generally associated with lower SOCD p10 (Fig. A.2). Notably, this pattern of divergence remained consistent across diverse vegetation types (Fig. 5). Influencing factors of SOCD Figure 6 presents the correlation analysis results between SOCD and 26 individual factors, as well as between SOCD and eight aggregated factor categories. The results indicate that most of the factors were significantly correlated with SOCD (left part of Fig. 6). Meanwhile, among the eight categories, VRI consistently showed the highest correlation with SOCD across all depth intervals (right part of Fig. 6). Specifically, all vegetation-related indices (i.e., NDVI, EVI, and LAI) showed significant positive correlations with SOCD. Following VRI, soil physical properties emerged as the second most influential category, with properties such as silt, soil moisture (SM), and bulk density (BD) showing notable associations with SOCD. Climate factors also played a role, with stronger correlations from moisture-related variables like MAP, LSWI, and ET compared to temperature-related factors such as MAT and LST. NPP and SRh, representing the input and output of SOC respectively, exhibited nearly equal strength in correlations with SOCD. Soil chemical properties and topography exhibited relatively weaker correlations with SOCD. Notably, pH negatively correlated with SOCD, while N showed a positive correlation. Elevation and slope also impacted SOCD, with elevation having a negative effect and slope a positive one. Additionally, SOCD was negatively correlated with solar radiation. Vertically, the correlation strength between SOCD and the eight categories of factors generally diminished with increasing soil depth, particularly for vegetation-related indices and climate factors. Structural Equation Modeling (SEM) results, which included the eight categories of factors, indicate that NPP (path coefficient = 1.74) and SRh (path coefficient = 2.41) were the dominant factors influencing SOCD (Fig. 7). Rad, Climate, SCP, SPP, Topog, and VRI primarily affected SOCD either directly or indirectly through their impact on NPP and SRh. For instance, Rad affected VRI and NPP by influencing the climate, subsequently impacting SOCD. Similarly, climate influenced SOCD through its effects on SPP, SCP, VRI, and NPP. Additionally, SRh was primarily influenced by SOCD, followed by climate and soil properties. Variations in factors influencing SOCD across vegetation types and depths The correlation between SOCD and influencing factors differed significantly among vegetation types (Fig. 8a). Some factors correlated positively with SOCD in certain vegetation types but negatively with others. For instance, LST correlated positively with SOCD in steppes and wetlands but negatively with other types. Additionally, the strength of correlations, as indicated by correlation coefficients and their significance levels, also varied among vegetation types. The relative importance of factors influencing SOCD varies markedly across different vegetation types (Fig. 8b, Fig. A.3). Tmin and Rad are the most important factors for forest and meadow, respectively, while LAI emerges as the dominant factor for other vegetation types, particularly steppe. Except for shrubland, most vegetation types, especially meadows, and steppes, had a considerable number of factors with importance exceeding 10%, highlighting the significant role multiple factors play in explaining SOCD variation (Fig. 8b, Fig. A.3). Correspondingly, the RF models explained 40–60% of SOCD variation across vegetation types, with the highest explanatory power in steppes and the lowest in shrublands (Fig. A.3). As depth increased, the number of important factors decreased across all vegetation types (Fig. A.3), indicating fewer factors explaining SOCD variation in deeper layers. Discussion Distribution features of SOCD In this study, we analyzed the three-dimensional distribution of SOCD across different vegetation types and depth intervals. The expansive TP, marked by considerable geographical diversity encompassing variations in climate, topography, vegetation, and landscape (Jia et al., 2023; Wang et al., 2023), leads to spatial heterogeneity in SOCD distribution (Gao et al., 2023). Our findings demonstrate that the SOCD on the TP exhibits a spatial pattern characterized by a gradual decrease from southeast to northwest, as supported by both sample data analysis (Fig. A.1) and simulation results (Fig. 3). This pattern aligns with previous studies (Gao et al., 2023; Han et al., 2022; Yang et al., 2023; Yang et al., 2010). Vertically, SOCD exhibited a nonlinear distribution across the 0-200 cm depth, slightly increasing within the 0-20 cm surface layers and decreasing at greater depths (Fig. 4). The overall decreasing trend is consistent with previous studies on the TP (Ding et al., 2016; Wang et al., 2023) and other regions (Zhang et al., 2018). Similar trends in surface layers were observed in other regions (Rose et al., 2017). This depth-dependent SOCD distribution highlights the concentration of biologically mediated carbon inputs in topsoil. Our findings also reveal that SOCD varied notably across different vegetation types, and this ranking remained consistent across different depth intervals, with the highest values observed in wetlands, followed by forests, meadows, shrublands, steppes, and other vegetation types (Fig. 5, Fig. A.2). Wetlands exhibited the highest SOCD due to limited decomposition under water-logged conditions (Jiang et al., 2019), covering only 2.6% of the plateau but representing over 6% of the total SOC stock within the upper 100 cm of soil (Chen et al., 2022). In contrast, forests on the plateau have higher net ecosystem productivity but also high soil respiration rates (Chen et al., 2022), resulting in them containing only 8% of the total SOC stock on the TP and having a lower SOCD than wetlands (Chen et al., 2022; Jiang et al., 2019). This study confirms that forests exhibit higher SOCD than grasslands (meadows and steppes), which aligns with findings at the national scale (Xu et al., 2019), though contrasts with localized studies (Wang et al., 2023). Among grassland types, meadows had higher SOCD than steppes, in agreement with prior research (Chen et al., 2022; Ding et al., 2017; Ding et al., 2016). These patterns suggest that vegetation type significantly influences soil carbon storage, which is relevant for optimizing land management and enhancing ecosystem carbon sinks (Post and Kwon, 2000). Effects of environmental factors on SOCD Net primary production and soil heterotrophic respiration, representing the carbon inputs and outputs of SOC, are the most direct environmental factors influencing SOCD in this study (Fig. 7). Among the eight categories of factors, vegetation-related indices had the strongest correlation with SOCD (Fig. 6). This strong relationship likely because vegetation indices reflect vegetation status and biomass, which are closely related to carbon inputs from plant litter and root turnover (Gao et al., 2013; Ma et al., 2024; Schloss et al., 1999), suggesting that vegetation indices can serve as effective indicators of soil carbon storage potential (Kunkel et al., 2022). Soil physical properties were the second most influential factors affecting SOCD, followed by climatic factors, NPP, soil heterotrophic respiration, soil chemical properties, topography and solar radiation (Fig. 6). SPP impacted SOCD by modulating vegetation growth (indicated by VRI) (Saporetti-Junior et al., 2012), which subsequently affects carbon inputs (indicated by NPP) (Fig.7). Additionally, SPP had a significant effect on soil respiration (Chen et al., 2014), which relates to the carbon output of SOC. For instance, soil moisture showed a significant positive correlation with VRI and NPP, while it negatively correlated with heterotrophic respiration (Fig. 6). Similarly, SCP, as a crucial aspect of soil properties, also affected SOC's input and output. For instance, enhances photosynthesis and biomass production, increasing carbon input to the soil, while also stimulating microbial activity and decomposition—both of which influence SOCD (Janssens et al., 2010; Leghari et al., 2016). Climate factors primarily influenced SOCD indirectly by regulating key ecological processes, including soil respiration and vegetation activity (Fig.7). For example, temperature simultaneously enhanced plant productivity (Zhong et al., 2019) and microbial decomposition (Ma et al., 2022), while precipitation shaped soil moisture regimes (Sehler et al., 2019) and promoted biomass accumulation (Zhang et al., 2021). Topographic and geomorphic factors modulated these climate–vegetation–soil interactions (Fig. 7). For instance, altitude affected climate patterns (Grabherr et al., 2010), such as temperature, which in turn affected plant growth and respiration rates. Additionally, solar radiation influenced the carbon cycle both directly through its effect on photosynthesis (Ballare et al., 2011) and indirectly via its role in climate modulation (Lean and Rind, 1998) (Fig. 7). Collectively, these findings indicate that SOCD is shaped by the synergistic effects of vegetation dynamics, soil characteristics, and climate conditions. These environmental drivers interact through coupled pathways that control carbon inputs via biomass production and outputs through microbial respiration. The integration of above- and belowground processes highlights the importance of a systems-level perspective for understanding the mechanisms governing soil carbon dynamics across ecosystems (Crowther et al., 2016; Luo et al., 2016). Divergent controls of SOCD in different ecosystems The impact of environmental factors on SOCD varies across regions and land use types (Wang et al., 2021; Wang et al., 2023). Our findings (Fig. 8) show that the same environmental factor can exert opposite effects depending on vegetation types. For example, increasing altitude tends to enhance SOCD in forests and shrublands but decreases it in grasslands and wetlands. Similarly, higher temperatures (MAT, LST) promote soil carbon accumulation in steppes and wetlands while reducing it in forests and shrublands. These contrasting responses are largely shaped by the vegetation's geographic distribution and the ecological traits of major vegetation types. Grasslands and wetlands, primarily distributed at higher altitudes of the northwestern TP (Hu et al., 2023) (Fig. 1), suffer from temperature constraints that limit plant productivity and carbon inputs (Ma et al., 2018; Piao et al., 2019). As elevation rises and temperature drops, plant productivity diminishes (Wang et al., 2020), leading to decreased soil carbon input and ultimately lowering SOCD (Teron et al., 2024). In contrast, forests and shrublands are mainly distributed in lower, warmer southeastern regions (Chen et al., 2020) (Fig. 1), where elevation-driven cooling reduces respiration and decomposition rates (Davidson and Janssens, 2006; Ma et al., 2022), thereby lowering carbon output and increasing SOCD (Xu et al., 2010). This divergent response arises from ecosystem-specific thresholds in carbon cycling, especially in how soils stabilize or release organic carbon (Schmidt et al., 2011). Forests and shrublands typically possess higher microbial carbon-use efficiency and more mineral-associated organic matter, enabling better carbon retention in cooler climates (Liang et al., 2017). In contrast, grasslands and wetlands in higher regions depend more on continuous biomass input and are thus more sensitive to temperature-driven changes in productivity (Clemmensen et al., 2013; Crowther et al., 2016). Collectively, these mechanisms explain why cooling increases SOCD in warm-region forests and shrublands by reducing carbon output more than input, but decreases SOCD in cold-region grasslands and wetlands by limiting input more than output Building on these ecosystem-specific mechanisms, we further found that both NPP and SRh were negatively correlated with SOCD in forests, but positively correlated in other vegetation types (Fig. 8). These contrasting correlations reinforce the notion that carbon cycling is governed by biome-specific regulatory processes. In forests, the contribution of carbon input (indicated by NPP) to SOCD may be outweighed by carbon losses via respiration. For instance, rising temperatures may enhance NPP, but—as discussed above—this warming-driven increase is likely modest. Moreover, much of the additional NPP may be sequestered in aboveground biomass (Aragão et al., 2009), limiting its contribution to soil carbon input. Furthermore, forest soils often have compact structure and deeper roots that promote microbial respiration, enhancing carbon loss (Qi and Yang, 2017). As a result, warming exerts a stronger influence on accelerating SOC decomposition than on enhancing carbon input, indicating that in forested regions, soil respiration (carbon output) largely determines the dynamics of SOCD. Conversely, in non-forest ecosystems such as grasslands, carbon inputs from NPP may surpass carbon losses through respiration. As discussed earlier, warming appears to boost soil carbon inputs in grassland systems, likely due to their typically higher belowground carbon allocation and faster organic matter turnover into soil (Herold et al., 2014; Trumbore et al., 1995). This is partly because grassland soils typically have looser structure, shallower roots, and distinct microbial communities, which favor fast carbon incorporation into soil (Juarez et al., 2013). However, the capacity of warming to accelerate carbon decomposition may be constrained—potentially due to the relatively limited availability of labile carbon substrates in these soils (Zhang et al., 2005), especially in the northwest grassland regions (Fig. 4). Thus, in non-forest ecosystems, like grasslands, SOCD is primarily governed by carbon input (indicated by NPP). This stands in contrast to forest ecosystems, highlighting biome-specific mechanisms driving variation in soil organic carbon across ecosystems. Limitations and implications This study compiled over 1,500 SOCD samples to simulate its distribution across the TP. However, the region’s complex terrain resulted in uneven spatial coverage, and sample sizes varied across vegetation types—forests and shrublands were underrepresented compared to grasslands—potentially introducing bias and increasing uncertainty. While these limitations affect spatial representativeness and interpretative depth, our analysis of SOCD drivers focuses on measured rather than simulated data. Moreover, the lack of comparative experiments constrains the empirical strength of some mechanistic explanations, despite extensive literature support. Despite these limitations, the study provides a high-resolution SOCD distribution dataset covering the entire plateau, offering essential data support for future assessments of regional carbon sinks and management strategies. The depth-dependent SOCD equation we developed offers a scalable approach for estimating subsoil carbon stocks, a component often overlooked in regional and global carbon inventories. Importantly, our results reveal that environmental factors influence SOCD in a context-specific manner, with variations across different vegetation types. This highlights the need for incorporating vegetation-specific parameters, such as root biomass allocation and microbial turnover, into carbon cycle models to improve their accuracy. Understanding these variations can refine model predictions and better capture the dynamics of soil carbon in diverse ecosystems. From a management perspective, our findings support the development of targeted carbon sequestration strategies. For example, enhancing vegetation productivity in alpine meadows may be more effective, while focusing on improving soil physical conditions could benefit alpine steppes. These insights can inform region-specific policies for carbon neutrality and land-based climate mitigation efforts. Conclusions This study provides the first comprehensive assessment of both horizontal and vertical patterns of SOCD and its environmental drivers across diverse vegetation types and soil depths (0-200cm) on the TP, based on analysis of over 1561 soil samples. Spatially, SOCD exhibited a decline from humid southeast to the arid northwest, aligning with variations in productivity, moisture, and organic matter stabilization potential. Vertically, SOCD followed a nonlinear distribution, peaking at approximately 2.58 kg C m⁻² per 10 cm within the 0-20 cm surface layers before decreasing at greater depths, indicating the concentration of biologically mediated carbon inputs in topsoil. SOCD varied by vegetation type, with the highest levels in wetlands (29.54 kg C m⁻²), followed by forests (27.06 kg C m⁻²), meadows (21.78 kg C m⁻²), shrublands (21.47 kg C m⁻²), and steppes (14.96 kg C m⁻²) at depth of 0–200 cm. The SOCD on the TP was influenced by a combination of environmental factors, including vegetation productivity, soil respiration, climate, topography, soil properties, and solar radiation. Overall, SOCD is primarily driven by vegetation growth and soil respiration—reflected by standardized path coefficients of 1.74 and –2.41, respectively—which control carbon inputs and outputs. Environmental factors modulate these processes to varying degrees across vegetation types, with their influence diminishing with soil depth. In forests, SOCD was predominantly controlled by carbon losses through soil respiration, while in grasslands and other vegetation types, carbon inputs via plant productivity were the primary determinant. Mechanistically, these differences may be attributed to variations in soil structure, microbial composition, and root depth across ecosystems. These findings elucidate divergent mechanisms regulating soil organic carbon across ecosystems, enabling more accurate modeling of soil carbon–climate interactions through ecosystem-specific calibrations of respiration rates, carbon input pathways, and decomposition parameters. Furthermore, the study highlights the need for region-specific management strategies for carbon sequestration to better support carbon neutrality and land-based climate mitigation efforts. Declarations CRediT authorship contribution statement Z. Hu: Writing – original draft, Writing – review & editing, Investigation, Methodology, Formal analysis, Data curation, Visualization, Validation, Software, Conceptualization. X. Zhang and J.C. Svenning: Supervision, Conceptualization, Resources, Writing – review & editing, Funding acquisition. B. Niu: Funding acquisition, Conceptualization. J. Tang: Investigation, Data curation. Y. Yue, W. Wu, M. Ni, Y. Yang: Writing – review & editing. Data availability Data will be made available on request. Acknowledgements This research was jointly supported by the Science and technology Projects of Xizang Autonomous Region, China (grant numbers, XZ202501ZY0087, XZ202501ZY0118), the West Light Foundation of the Chinese Academy of Sciences (grant number, xbzg-zdsys-202202), and the Second Tibetan Plateau Scientific Expedition and Research Program (STEP) (grant number 2019QZKK1002). We also consider this work a contribution to Center for Ecological Dynamics in a Novel Biosphere (ECONOVO), funded by Danish National Research Foundation (grant DNRF173 to JCS). The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. References Aragão L, Malhi Y, Metcalfe D, Silva-Espejo JE, Jiménez E, Navarrete D, et al. Above-and below-ground net primary productivity across ten Amazonian forests on contrasting soils. Biogeosciences 2009; 6: 2759-2778. Ballare CL, Caldwell MM, Flint SD, Robinson SA, Bornman JF. Effects of solar ultraviolet radiation on terrestrial ecosystems. Patterns, mechanisms, and interactions with climate change. Photochemical & Photobiological Sciences 2011; 10: 226-241. Bradford MA, Wieder WR, Bonan GB, Fierer N, Raymond PA, Crowther TW. Managing uncertainty in soil carbon feedbacks to climate change. Nature Climate Change 2016; 6: 751-758. Chen A, Huang L, Liu Q, Piao S. Optimal temperature of vegetation productivity and its linkage with climate and elevation on the Tibetan Plateau. Global Change Biology 2021; 27: 1942-1951. Chen H, Ju P, Zhu Q, Xu X, Wu N, Gao Y, et al. Carbon and nitrogen cycling on the Qinghai–Tibetan Plateau. Nature Reviews Earth & Environment 2022; 3: 701-716. Chen J, Yan F, Lu Q. Spatiotemporal variation of vegetation on the Qinghai–Tibet Plateau and the influence of climatic factors and human activities on vegetation trend (2000–2019). Remote Sensing 2020; 12: 3150. Chen S, Zou J, Hu Z, Chen H, Lu Y. Global annual soil respiration in relation to climate, soil properties and vegetation characteristics: Summary of available data. Agricultural and Forest Meteorology 2014; 198: 335-346. Clemmensen K, Bahr A, Ovaskainen O, Dahlberg A, Ekblad A, Wallander H, et al. Roots and associated fungi drive long-term carbon sequestration in boreal forest. Science 2013; 339: 1615-1618. Crowther TW, Todd-Brown KE, Rowe CW, Wieder WR, Carey JC, Machmuller MB, et al. Quantifying global soil carbon losses in response to warming. Nature 2016; 540: 104-108. Davidson EA, Janssens IA. Temperature sensitivity of soil carbon decomposition and feedbacks to climate change. Nature 2006; 440: 165-173. Ding J, Chen L, Ji C, Hugelius G, Li Y, Liu L, et al. Decadal soil carbon accumulation across Tibetan permafrost regions. Nature Geoscience 2017; 10: 420-424. Ding J, Li F, Yang G, Chen L, Zhang B, Liu L, et al. The permafrost carbon inventory on the Tibetan Plateau: a new evaluation using deep sediment cores. Global Change Biology 2016; 22: 2688-2701. Ding J, Wang T, Piao S, Smith P, Zhang G, Yan Z, et al. The paleoclimatic footprint in the soil carbon stock of the Tibetan permafrost region. Nature Communications 2019; 10: 4195. Dokuchaev VV. Russian chernozem. St. Petersburg: Evdokimova, 1883. Gaitán JJ, Maestre FT, Bran DE, Buono GG, Dougill AJ, Garcia Martinez G, et al. Biotic and abiotic drivers of topsoil organic carbon concentration in drylands have similar effects at regional and global scales. Ecosystems 2019; 22: 1445-1456. Gao M, Xu R, Huang J, Su B, Jiang S, Shi P, et al. Increase of carbon storage in the Qinghai-Tibet Plateau: Perspective from land-use change under global warming. Journal of Cleaner Production 2023; 414: 137540. Gao Y, Zhou X, Wang Q, Wang C, Zhan Z, Chen L, et al. Vegetation net primary productivity and its response to climate change during 2001–2008 in the Tibetan Plateau. Science of the Total Environment 2013; 444: 356-362. Genxu W, Yuanshou L, Yibo W, Qingbo W. Effects of permafrost thawing on vegetation and soil carbon pool losses on the Qinghai–Tibet Plateau, China. Geoderma 2008; 143: 143-152. Georgiou K, Jackson RB, Vindušková O, Abramoff RZ, Ahlström A, Feng W, et al. Global stocks and capacity of mineral-associated soil organic carbon. Nature Communications 2022; 13: 3797. Grabherr G, Gottfried M, Pauli H. Climate change impacts in alpine environments. Geography Compass 2010; 4: 1133-1153. Han D, Hu Z, Wang X, Wang T, Chen A, Weng Q, et al. Shift in controlling factors of carbon stocks across biomes on the Qinghai-Tibetan Plateau. Environmental Research Letters 2022; 17: 074016. Herold N, Schöning I, Michalzik B, Trumbore S, Schrumpf M. Controls on soil carbon storage and turnover in German landscapes. Biogeochemistry 2014; 119: 435-451. Homann PS, Kapchinske JS, Boyce A. Relations of mineral-soil C and N to climate and texture: regional differences within the conterminous USA. Biogeochemistry 2007; 85: 303-316. Hu Z, Niu B, Tang J, Zhang Y, Xiang M, Zhang X. Has the Dominant Climatic Driver for the Carbon Budget of Alpine Grassland Shifted from Temperature to Precipitation on the Qinghai–Tibet Plateau? Remote Sensing 2023; 15: 2492. Hutchinson M. ANUSPLIN Version 4.3. Centre for Resource and Environmental Studies. The Australian National University, Canberra, Australia 2004. Janssens I, Dieleman W, Luyssaert S, Subke J-A, Reichstein M, Ceulemans R, et al. Reduction of forest soil respiration in response to nitrogen deposition. Nature geoscience 2010; 3: 315-322. Jia Z, Wang X, Feng X, Ma J, Wang X, Zhang X, et al. Exploring the spatial heterogeneity of ecosystem services and influencing factors on the Qinghai Tibet Plateau. Ecological Indicators 2023; 154: 110521. Jiang L, Chen H, Zhu Q, Yang Y, Li M, Peng C, et al. Assessment of frozen ground organic carbon pool on the Qinghai-Tibet Plateau. Journal of soils and sediments 2019; 19: 128-139. Jiang M, Li H, Zhang W, Liu J, Zhang Q. Effects of climate change and grazing on the soil organic carbon stock of alpine wetlands on the Tibetan Plateau from 2000 to 2018. Catena 2024; 238: 107870. Jobbágy EG, Jackson RB. The vertical distribution of soil organic carbon and its relation to climate and vegetation. Ecological applications 2000; 10: 423-436. Juarez S, Nunan N, Duday A-C, Pouteau V, Schmidt S, Hapca S, et al. Effects of different soil structures on the decomposition of native and added organic carbon. European journal of soil biology 2013; 58: 81-90. Kowalenko CG. Assessment of Leco CNS-2000 analyzer for simultaneously measuring total carbon, nitrogen, and sulphur in soil. Communications in Soil Science and Plant Analysis 2001; 32: 2065-2078. Kuang X, Jiao JJ. Review on climate change on the Tibetan Plateau during the last half century. Journal of Geophysical Research: Atmospheres 2016; 121: 3979-4007. Kunkel V, Wells T, Hancock G. Modelling soil organic carbon using vegetation indices across large catchments in eastern Australia. Science of The Total Environment 2022; 817: 152690. Lamichhane S, Kumar L, Wilson B. Digital soil mapping algorithms and covariates for soil organic carbon mapping and their implications: A review. Geoderma 2019; 352: 395-413. Lean J, Rind D. Climate forcing by changing solar radiation. Journal of Climate 1998; 11: 3069-3094. Leghari SJ, Wahocho NA, Laghari GM, HafeezLaghari A, MustafaBhabhan G, HussainTalpur K, et al. Role of nitrogen for plant growth and development: A review. Advances in Environmental Biology 2016; 10: 209-219. Lehmann J, Hansel CM, Kaiser C, Kleber M, Maher K, Manzoni S, et al. Persistence of soil organic carbon caused by functional complexity. Nature Geoscience 2020; 13: 529-534. Lehmann J, Kleber M. The contentious nature of soil organic matter. Nature 2015; 528: 60-68. Li H, Wu Y, Liu S, Xiao J, Zhao W, Chen J, et al. Decipher soil organic carbon dynamics and driving forces across China using machine learning. Global Change Biology 2022a; 28: 3394-3410. Li M, Wu JS, Song CQ, He YT, Niu B, Fu G, et al. Temporal Variability of Precipitation and Biomass of Alpine Grasslands on the Northern Tibetan Plateau. Remote Sensing 2019; 11. Li Q, Shi G, Shangguan W, Nourani V, Li J, Li L, et al. A 1 km daily soil moisture dataset over China using in situ measurement and machine learning. Earth System Science Data 2022b; 14: 5267-5286. Liang C, Schimel JP, Jastrow JD. The importance of anabolism in microbial control over soil carbon storage. Nature microbiology 2017; 2: 1-6. Liaw A, Wiener M. Classification and regression by randomForest. R news 2002; 2: 18-22. Luo Y, Ahlström A, Allison SD, Batjes NH, Brovkin V, Carvalhais N, et al. Toward more realistic projections of soil carbon dynamics by Earth system models. Global Biogeochemical Cycles 2016; 30: 40-56. Ma H, Peng M, Yang Z, Yang K, Zhao C, Li K, et al. Spatial distribution and driving factors of soil organic carbon in the Northeast China Plain: Insights from latest monitoring data. Science of the Total Environment 2024; 911: 168602. Ma M, Yuan W, Dong J, Zhang F, Cai W, Li H. Large-scale estimates of gross primary production on the Qinghai-Tibet plateau based on remote sensing data. International Journal of Digital Earth 2018; 11: 1166-1183. Ma X, He P, Zeng Y, Ma J, Wu X. A global-drive analysis of ecosystem respiration in the Arctic and Third Pole. Ecological Indicators 2022; 145: 109668. Meyer RS, Cullen BR, Whetton PH, Robertson FA, Eckard RJ. Potential impacts of climate change on soil organic carbon and productivity in pastures of south eastern Australia. Agricultural Systems 2018; 167: 34-46. Peng S. 1km Monthly Potential Evapotranspiration Dataset in China (1990–2022). National Tibetan Plateau / Third Pole Environment Data Center, Beijing, China, 2022. Piao S, Niu B, Zhu J, Zhang X, Wang T, Wang S, et al. Responses and feedback of the Tibetan Plateau's alpine ecosystem to climate change. Chinese Science Bulletin 2019; 64: 2842-2855. Piao S, Tan K, Nan H, Ciais P, Fang J, Wang T, et al. Impacts of climate and CO2 changes on the vegetation growth and carbon balance of Qinghai–Tibetan grasslands over the past five decades. Global and Planetary Change 2012; 98: 73-80. Plante A, Conant R, Freedman B. Soil organic matter dynamics, climate change effects. Global Environmental Change, edited by: Freedman, B 2014: 317-323. Poggio L, De Sousa LM, Batjes NH, Heuvelink GB, Kempen B, Ribeiro E, et al. SoilGrids 2.0: producing soil information for the globe with quantified spatial uncertainty. Soil 2021; 7: 217-240. Post WM, Kwon KC. Soil carbon sequestration and land‐use change: processes and potential. Global change biology 2000; 6: 317-327. Prietzel J, Zimmermann L, Schubert A, Christophel D. Organic matter losses in German Alps forest soils since the 1970s most likely caused by warming. Nature Geoscience 2016; 9: 543-548. Qi L, Yang J. Microbial community composition regulates SOC decomposition response to forest conversion in a Chinese temperate forest. Ecological research 2017; 32: 163-172. R Core Team R. R: A language and environment for statistical computing. 2013. Ramesh T, Bolan NS, Kirkham MB, Wijesekara H, Kanchikerimath M, Rao CS, et al. Soil organic carbon dynamics: Impact of land use changes and management practices: A review. Advances in agronomy 2019; 156: 1-107. Rose A, Shakurideen O, Kolawole M, Davies N, Ayejuyo O, Friday A, et al. An investigation of the role of vegetation as a determinant of the vertical distribution of soil organic carbon. Journal of Emerging Trends in Engineering and Applied Sciences 2017; 8: 43-53. Rosseel Y. lavaan: An R package for structural equation modeling. Journal of statistical software 2012; 48: 1-36. Saporetti-Junior AW, Schaefer CEGR, de Souza AL, Soares MP, Araújo DSD, Meira-Neto JAA. Influence of soil physical properties on plants of the Mussununga ecosystem, Brazil. Folia Geobotanica 2012; 47: 29-39. Schloss A, Kicklighter D, Kaduk J, Wittenberg U, Intercomparison TPOTPNM. Comparing global models of terrestrial net primary productivity (NPP): comparison of NPP to climate and the Normalized Difference Vegetation Index (NDVI). Global Change Biology 1999; 5: 25-34. Schmidt MW, Torn MS, Abiven S, Dittmar T, Guggenberger G, Janssens IA, et al. Persistence of soil organic matter as an ecosystem property. Nature 2011; 478: 49-56. Sehler R, Li J, Reager J, Ye H. Investigating relationship between soil moisture and precipitation globally using remote sensing observations. Journal of Contemporary Water Research & Education 2019; 168: 106-118. Stockmann U, Padarian J, McBratney A, Minasny B, de Brogniez D, Montanarella L, et al. Global soil organic carbon assessment. Global Food Security 2015; 6: 9-16. Teron G, Bordoloi R, Paul A, Singha LB, Tripathi OP. Effect of altitude on soil physico-chemical properties and microbial biomass carbon in the Eaglenest Wildlife Sanctuary of Arunachal Pradesh. Geology, Ecology, and Landscapes 2024: 1-19. Thornthwaite CW. An approach toward a rational classification of climate. Geographical review 1948; 38: 55-94. Tian H, Lu C, Yang J, Banger K, Huntzinger DN, Schwalm CR, et al. Global patterns and controls of soil organic carbon dynamics as simulated by multiple terrestrial biosphere models: Current status and future directions. Global Biogeochemical Cycles 2015; 29: 775-792. Trumbore SE, Davidson EA, Barbosa de Camargo P, Nepstad DC, Martinelli LA. Belowground cycling of carbon in forests and pastures of Eastern Amazonia. Global biogeochemical cycles 1995; 9: 515-528. Wang X, Hu Z, Zhang Z, Tang J, Niu B. Altitude-Shifted Climate Variables Dominate the Drought Effects on Alpine Grasslands over the Qinghai–Tibetan Plateau. Sustainability 2024; 16: 6697. Wang Y, Huang L, Jia X, Zhu Y. Distribution characteristics and controls of soil organic carbon at different spatial scales in China′ s Loess Plateau. Journal of environmental management 2021; 293: 112943. Wang Y, Peng D, Shen M, Xu X, Yang X, Huang W, et al. Contrasting effects of temperature and precipitation on vegetation greenness along elevation gradients of the Tibetan Plateau. Remote Sensing 2020; 12: 2751. Wang Z, Huang L, Shao Ma. Spatial variations and influencing factors of soil organic carbon under different land use types in the alpine region of Qinghai-Tibet Plateau. Catena 2023; 220. Wright AF, Bailey JS. Organic carbon, total carbon, and total nitrogen determinations in soils of variable calcium carbonate contents using a Leco CN-2000 dry combustion analyzer. Communications in Soil Science and Plant Analysis 2001; 32: 3243-3258. Xu L, Yu G, He N. Increased soil organic carbon storage in Chinese terrestrial ecosystems from the 1980s to the 2010s. Journal of Geographical Sciences 2019; 29: 49-66. Xu Z, Wan C, Xiong P, Tang Z, Hu R, Cao G, et al. Initial responses of soil CO 2 efflux and C, N pools to experimental warming in two contrasting forest ecosystems, Eastern Tibetan Plateau, China. Plant and Soil 2010; 336: 183-195. Yang R, Huang L, Zhang X, Zhu C, Xu L. Mapping the distribution, trends, and drivers of soil organic carbon in China from 1982 to 2019. Geoderma 2023; 429: 116232. Yang Y, Fang J, Ma W, Smith P, Mohammat A, Wang S, et al. Soil carbon stock and its changes in northern China's grasslands from 1980s to 2000s. Global Change Biology 2010; 16: 3036-3047. Yang Y, Fang J, Tang Y, Ji C, Zheng C, He J, et al. Storage, patterns and controls of soil organic carbon in the Tibetan grasslands. Global change biology 2008; 14: 1592-1599. Yao T, Thompson LG, Mosbrugger V, Zhang F, Ma Y, Luo T, et al. Third Pole Environment (TPE). Environmental Development 2012; 3: 52-64. Yao T, Xue Y, Chen D, Chen F, Thompson L, Cui P, et al. Recent third pole’s rapid warming accompanies cryospheric melt and water cycle intensification and interactions between monsoon and environment: Multidisciplinary approach with observations, modeling, and analysis. Bulletin of the American Meteorological society 2019; 100: 423-444. Yin Y, Wu S, Zhao D, Zheng D, Pan T. Modeled effects of climate change on actual evapotranspiration in different eco-geographical regions in the Tibetan Plateau. Journal of Geographical Sciences 2013; 23: 195-207. Zhang W, Zhang F, Qi J, Hou F. Modeling impacts of climate change and grazing effects on plant biomass and soil organic carbon in the Qinghai–Tibetan grasslands. Biogeosciences 2017; 14: 5455-5470. Zhang W, Zhu W, Hu S. Soil resource availability impacts microbial response to organic carbon and inorganic nitrogen inputs. Journal of Environmental Sciences (China) 2005; 17: 705-710. Zhang X, Liu M, Zhao X, Li Y, Zhao W, Li A, et al. Topography and grazing effects on storage of soil organic carbon and nitrogen in the northern China grasslands. Ecological Indicators 2018; 93: 45-53. Zhang X, Yang Y, Piao S, Bao W, Wang S, Wang G, et al. Ecological change on the Tibetan Plateau. Chinese Science Bulletin 2015; 60: 3048-3056. Zhang Y, Hu Q, Zou F. Spatio-temporal changes of vegetation net primary productivity and its driving factors on the Qinghai-Tibetan Plateau from 2001 to 2017. Remote Sensing 2021; 13: 1566. Zheng C, Jia L, Hu G. Global land surface evapotranspiration monitoring by ETMonitor model driven by multi-source satellite earth observations. Journal of Hydrology 2022; 613: 128444. Zhong L, Ma Y, Xue Y, Piao S. Climate Change Trends and Impacts on Vegetation Greening Over the Tibetan Plateau. Journal of Geophysical Research: Atmospheres 2019; 124: 7540-7552. Zhu Z, Wang H, Harrison SP, Prentice IC, Qiao S, Tan S. Optimality principles explaining divergent responses of alpine vegetation to environmental change. Global Change Biology 2023; 29: 126-142. Supplementary Files Appendices.docx SupplementaryMaterial.xlsx Cite Share Download PDF Status: Published Journal Publication published 02 Oct, 2025 Read the published version in Plant and Soil → Version 1 posted Editorial decision: Major revisions 10 Jun, 2025 Reviewers agreed at journal 15 May, 2025 Reviewers invited by journal 14 May, 2025 Editor invited by journal 14 May, 2025 Editor assigned by journal 13 May, 2025 First submitted to journal 13 May, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6651952","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":456335640,"identity":"da7f63da-a439-46fc-856d-9c2b88f68876","order_by":0,"name":"Zhigang Hu","email":"data:image/png;base64,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","orcid":"","institution":"Aarhus Universitet","correspondingAuthor":true,"prefix":"","firstName":"Zhigang","middleName":"","lastName":"Hu","suffix":""},{"id":456335641,"identity":"82dff325-3a84-43a2-945b-60abdc9ccb61","order_by":1,"name":"Jens-Christian Svenning","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jens-Christian","middleName":"","lastName":"Svenning","suffix":""},{"id":456335642,"identity":"41c9019e-f12c-4d4d-87c3-621328ba3d49","order_by":2,"name":"Jiwang Tang","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jiwang","middleName":"","lastName":"Tang","suffix":""},{"id":456335643,"identity":"7e38e928-68de-4242-9e91-ad9fee122a51","order_by":3,"name":"Yue Cheng","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Yue","middleName":"","lastName":"Cheng","suffix":""},{"id":456335644,"identity":"b87012b0-0f0a-401c-a5ac-728de8e3bf6d","order_by":4,"name":"Wanben Wu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Wanben","middleName":"","lastName":"Wu","suffix":""},{"id":456335645,"identity":"86a71d0d-1d93-4e28-924f-042cb45ebfc7","order_by":5,"name":"Ming Ni","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Ming","middleName":"","lastName":"Ni","suffix":""},{"id":456335646,"identity":"40461e77-6202-4a9b-9e0b-f61432bb3209","order_by":6,"name":"Yingbo Yang","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Yingbo","middleName":"","lastName":"Yang","suffix":""},{"id":456335647,"identity":"d234c3fe-b42c-410e-8586-364883efd9ba","order_by":7,"name":"Xianzhou Zhang","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Xianzhou","middleName":"","lastName":"Zhang","suffix":""},{"id":456335648,"identity":"3e6b8f56-91a5-424e-814b-fcba94cfe635","order_by":8,"name":"Ben Niu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Ben","middleName":"","lastName":"Niu","suffix":""}],"badges":[],"createdAt":"2025-05-13 06:19:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6651952/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6651952/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s11104-025-07845-1","type":"published","date":"2025-10-02T15:58:04+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":82884250,"identity":"a0949f9d-db58-472c-9810-172335780cf7","added_by":"auto","created_at":"2025-05-16 11:33:30","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1604483,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of vegetation types and soil organic carbon samples on the Tibetan Plateau (a). Temperature and precipitation trends are shown along longitude and latitude (c, d). Data source: Resource and Environment Science and Data Center of the Chinese Academy of Sciences; National Meteorological Information Center.\u003c/p\u003e","description":"","filename":"image1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6651952/v1/c8f64fa0407fc5bf4d0d7e3e.jpeg"},{"id":82884249,"identity":"8625dc47-ac2c-4ff7-bd23-eac2709b7fa8","added_by":"auto","created_at":"2025-05-16 11:33:30","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":146172,"visible":true,"origin":"","legend":"\u003cp\u003eComparison between the observed data-based soil organic carbon density (SOCD) (x-axis) and the SOCD predicted via Random Forest models (y-axis). The dashed gray line shows the 1:1 line. The bands represent 95% confidence.\u003c/p\u003e","description":"","filename":"image2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6651952/v1/84f014c6173154560fe3e492.jpeg"},{"id":82884252,"identity":"e788d699-da9e-4fcf-9330-4317c458c705","added_by":"auto","created_at":"2025-05-16 11:33:30","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":145021,"visible":true,"origin":"","legend":"\u003cp\u003eThe vertical pattern of the soil organic carbon density per 10cm thickness (SOCD\u003csub\u003ep10\u003c/sub\u003e) (a) and its relationship with soil depth (b). The purple line in the left panel (a) represents the trend of mean SOCD\u003csub\u003ep10\u003c/sub\u003e. The depth shown in the right panel (b) corresponds to the midpoint of depth intervals, with the corresponding SOCD\u003csub\u003e p10\u003c/sub\u003e value representing the average for all soil samples within that depth interval.\u003c/p\u003e","description":"","filename":"image3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6651952/v1/782e9f31e26c4505da88d24d.jpeg"},{"id":82884948,"identity":"4d0b9702-28bf-4487-b5bc-4dfa8a3f09b3","added_by":"auto","created_at":"2025-05-16 11:41:30","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":579685,"visible":true,"origin":"","legend":"\u003cp\u003eThe spatial pattern of the soil organic carbon density across the Tibetan Plateau based on Random Forest models.\u003c/p\u003e","description":"","filename":"image4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6651952/v1/04bbe8c3d86a27bbe6497a92.jpeg"},{"id":82884951,"identity":"d58cb44f-0854-48c2-b62d-fb187a93991a","added_by":"auto","created_at":"2025-05-16 11:41:30","extension":"jpeg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":142542,"visible":true,"origin":"","legend":"\u003cp\u003eThe distribution differences in soil organic carbon density per 10cm thickness (SOCD\u003csub\u003ep10\u003c/sub\u003e) across various vegetation types and depths. The gradient cyan histograms represent the mean SOCD\u003csub\u003ep10\u003c/sub\u003e across different depths. The histograms correspond to the secondary y-axis.\u003c/p\u003e","description":"","filename":"image5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6651952/v1/6e40153c87f0110aa0fcd65c.jpeg"},{"id":82884949,"identity":"d78e2463-2af2-4547-8d27-d26d44d60ebf","added_by":"auto","created_at":"2025-05-16 11:41:30","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":188733,"visible":true,"origin":"","legend":"\u003cp\u003eRelationship between soil organic carbon density (SOCD\u003csub\u003ep10\u003c/sub\u003e) and influencing factors. S30, S50, S100, and S200 represent SOCD\u003csub\u003ep10\u003c/sub\u003e for soil depth intervals of 0-30cm, 30-50cm, 50-100cm, and 100-200cm, respectively. Elev, elevation; CF, coarse fragments; BD, bulk density; SM, soil moisture; N, nitrogen; CEC, cation exchange capacity; SRh, soil heterotrophic respiration; MAP, mean annual precipitation; MAT mean annual temperature; Tmax, maximum temperature; Tmin, minimum temperature; LST, land surface temperature; LSWI, land surface water index; AI, aridity index; ET, evapotranspiration; Rad, solar radiation; NDVI, normalized difference vegetation index; EVI, enhanced vegetation index; LAI, leaf area index; NPP, net primary productivity; Topog, topography; SPP, soil physical properties; SCP, soil chemical properties; VRI, vegetation-related indices.\u003c/p\u003e","description":"","filename":"image6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6651952/v1/fb084273a51619dd263e4404.jpg"},{"id":82884254,"identity":"07fcc897-db14-492d-bfc4-9e5f14039d6b","added_by":"auto","created_at":"2025-05-16 11:33:30","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":120889,"visible":true,"origin":"","legend":"\u003cp\u003eStructural equation model (SEM) showing the relationship between soil organic carbon density (SOCD) and eight categories of factors. Abbreviations accorded to Figure 6. Dash lines indicate insignificance (p \u0026gt; 0.05). Red arrows indicate negative relationships. For categories with multiple variables, connected paths do not reflect positive or negative relationships, as their values do not represent true variable values, unlike single-variable categories (i.e., Rad, NPP, SRh, SOCD).\u003c/p\u003e","description":"","filename":"image7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6651952/v1/5a244b4dba26754427a04be4.jpg"},{"id":82884258,"identity":"6e88d8d1-bde7-4554-8e56-a55e843ea608","added_by":"auto","created_at":"2025-05-16 11:33:30","extension":"jpeg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":396169,"visible":true,"origin":"","legend":"\u003cp\u003eContributions of influencing factors to mean soil organic carbon density at 0–200 cm depth, and differences across vegetation types based on correlations and Random Forest models. The abbreviations accorded to Figure 6.\u003c/p\u003e","description":"","filename":"image8.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6651952/v1/45259b5f24157119c6881279.jpeg"},{"id":92884108,"identity":"0b79e069-85e7-4d5f-8297-c61c2ec139fb","added_by":"auto","created_at":"2025-10-06 16:12:23","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3848887,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6651952/v1/1421ade7-d3d3-4d5a-891d-aed4003d0c7d.pdf"},{"id":82885810,"identity":"eee809aa-9903-4646-a708-4d1708d6dbfa","added_by":"auto","created_at":"2025-05-16 11:49:30","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":607343,"visible":true,"origin":"","legend":"","description":"","filename":"Appendices.docx","url":"https://assets-eu.researchsquare.com/files/rs-6651952/v1/f7d6628bb02fd8bccece8984.docx"},{"id":82884952,"identity":"f4120d16-cdd0-46be-98aa-f0399c09f812","added_by":"auto","created_at":"2025-05-16 11:41:30","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":436113,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6651952/v1/445fb61d638f2b61aa0968d5.xlsx"}],"financialInterests":"","formattedTitle":"Distributions and Drivers of Soil Organic Carbon on the Tibetan Plateau: Divergent Controls Across Ecosystems","fulltext":[{"header":"Introduction","content":"\u003cp\u003eBeing the largest terrestrial organic carbon pool on Earth (Jobb\u0026aacute;gy and Jackson, 2000; Lehmann et al., 2020), soil harbors more carbon than the atmosphere and all vegetation combined (Georgiou et al., 2022). This vast pool not only regulates climate\u0026ndash;carbon feedbacks (Bradford et al., 2016; Davidson and Janssens, 2006) but also underpins ecosystem productivity and resilience (Li et al., 2022a; Xu et al., 2019). As early as the 1880s, scientists recognized that soil organic carbon (SOC) drives both fertility and plant growth (Dokuchaev, 1883; Jobb\u0026aacute;gy and Jackson, 2000; Tian et al., 2015). Today, enhancing SOC is more than just a soil health goal\u0026mdash;it\u0026rsquo;s a nature-based solution for mitigating climate change (Lehmann et al., 2020; Meyer et al., 2018) and sustaining critical ecosystem services (Prietzel et al., 2016; Stockmann et al., 2015).\u003c/p\u003e\n\u003cp\u003eSOC content reflects the long-term balance between inputs from net primary production, such as surface litter and root exudates, and outputs from soil physicochemical processes, including decomposition and leaching of dissolved organic carbon (Jobb\u0026aacute;gy and Jackson, 2000; Li et al., 2022a; Plante et al., 2014). Climate changes, vegetation dynamics, environmental variation, and anthropogenic activities (Lehmann and Kleber, 2015) all influence the input-output processes, rendering soil organic carbon, particularly in topsoil, highly sensitive to biotic and abiotic factors (Gait\u0026aacute;n et al., 2019; Li et al., 2022a). Even small changes in SOC can influence the broader carbon cycle and contribute to climate feedbacks (Ding et al., 2017; Ma et al., 2024; Ramesh et al., 2019). However, our comprehension of the spatial distribution of SOC and the underlying driving factors remains restricted, particularly at large regional scales.\u003c/p\u003e\n\u003cp\u003eAlpine soils, characterized by high SOC density due to low temperatures that suppress decomposition and biomass turnover, play a critical role in the regional carbon balance and climate feedbacks (Davidson and Janssens, 2006; Wang et al., 2023). As the world\u0026rsquo;s \u0026ldquo;Third Pole\u0026rdquo; and the \u0026ldquo;Water Tower of Asia,\u0026rdquo; the Tibetan Plateau (TP) harbors extensive alpine ecosystems and vast permafrost soils that have accumulated substantial SOC stocks over millennia, profoundly shaping the regional carbon cycle and climate system (Genxu et al., 2008; Wang et al., 2023; Yao et al., 2012). Since 1970, the TP has warmed rapidly at 0.35\u0026deg;C per decade, exceeding the global average and comparable to the Arctic warming rates (Kuang and Jiao, 2016; Yao et al., 2019), making it one of the most climate-sensitive and ecologically fragile regions globally (Hu et al., 2023). Climate warming accelerates SOC decomposition and disrupts the regional carbon balance, while human activities such as overgrazing further intensify SOC loss (Jiang et al., 2024; Zhang et al., 2017; Zhu et al., 2023). Given its vast carbon storage and strong climate sensitivity, the TP is a critical region for understanding SOC distribution and its environmental drivers, with far-reaching implications for both regional carbon management and global carbon\u0026ndash;climate feedbacks through processes like permafrost thaw and ecosystem change (Chen et al., 2022; Wang et al., 2023; Yang et al., 2008; Yao et al., 2019).\u003c/p\u003e\n\u003cp\u003eHowever, many previous studies have focused on individual sites, specific regions, or single vegetation types, limiting their capacity to capture broader spatial trends (Ding et al., 2016; Wang et al., 2023). Given the TP\u0026rsquo;s complex topography and climatic gradients, its geographic environment exhibits pronounced spatial heterogeneity in factors such as elevation, temperature, and vegetation types (Jia et al., 2023; Wang et al., 2023; Zhang et al., 2015). This heterogeneity gives rise to substantial variations in SOC distribution and its controlling mechanisms across different ecosystems and soil profiles (Gao et al., 2023; Wang et al., 2023). For example, the influence of plant productivity, soil respiration, or moisture availability may differ markedly between alpine steppes and forests, or between the topsoil and subsoil layer (Han et al., 2022; Li et al., 2022a). As a result, findings derived from site-level studies often fail to adequately capture these differences and represent regional-scale patterns. Additionally, while key variables like temperature and precipitation are frequently examined, other potentially important drivers, such as topography, vegetation productivity or soil, are often underrepresented (Jiang et al., 2024). Since SOC dynamics result from the complex interplay of biotic and abiotic factors, examining a broader and ecologically relevant set of environmental variables is essential to disentangle their relative contributions and interactions (Wang et al., 2023), and to better inform Earth system models. Therefore, it is necessary to extensively collect regional soil samples and include a wider range of variables to compensate for the limitations of the site-specific studies.\u003c/p\u003e\n\u003cp\u003eIn this study, we compiled soil carbon data from 1,561 sampling sites collected between 2001 and 2021. These samples span major vegetation types across the TP, including forests, shrublands, grasslands, and wetlands. Using this extensive and spatially representative dataset, we analyzed the horizontal and vertical distribution of SOC density (SOCD) from the surface to 200 cm depth. Furthermore, we assessed the influence of 26 environmental factors\u0026mdash;ranging from solar radiation, climate, and topography to soil properties and vegetation indices\u0026mdash;to systematically investigate the drivers of SOCD across this highly heterogeneous alpine region. Specifically, this study aimed to: (1) quantify the spatial distribution features of SOCD across the TP in both horizontally and vertically dimensions; (2) determine the relative importance of multiple environmental drivers of SOCD, and (3) test the hypothesis that both the magnitude and environmental controls of SOCD differ across vegetation types and soil depths.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cp\u003e\u003cem\u003eStudy area\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe Tibetan Plateau, commonly referred to as the \u0026ldquo;Third Pole\u0026rdquo; with an average elevation above 4000 m, covers more than a quarter of China\u0026rsquo;s total land area, rendering it the highest and largest plateau on Earth (Piao et al., 2012; Zhu et al., 2023). The annual average air temperature in most areas of the plateau remains below 0\u0026deg;C, with the average monthly air temperature ranging from -11.6\u0026deg;C in January to 9.1\u0026deg;C in July during the period from 2000 to 2020 (Hu et al., 2023; Wang et al., 2024). Generally, on the TP, temperature and precipitation decrease while altitude increases from southeast to northwest (Yin et al., 2013) (Fig. 1b,c). Annual average precipitation varies from 100 mm per year in the northwest to over 1000 mm per year in the southeast (Chen et al., 2021) (Fig. 1b,c). The plateau is mainly covered by steppe, meadow, forest, and shrubland (Fig. 1a).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eData collection and processing\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eSoil organic carbon data\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eIn this study, we collected and compiled 1561 soil organic carbon samples (Fig. 1a, Supplementary Table S1). Among these, 236 samples were newly acquired by our research team during field investigations in the summers of 2020 and 2021. During the investigations, sampling was conducted along the northern Tibetan grassland transect from east to west, with three replicates collected per site across a meadow-to-steppe gradient (see Supplementary Table S1 for data and sampling depths). After collection, soil samples were air-dried at ~25\u0026deg;C, manually disaggregated and sieved through a 2 mm mesh to remove coarse fragments. A subsample was then ground to a fine powder using a ball mill and stored in sealed containers until analysis. SOC was quantified using dry combustion with a LECO CN analyzer (LECO Corporation, USA), following the standard protocol described by Nelson and Sommers (1996), and validated through comparable studies using LECO dry combustion methods (Kowalenko, 2001; Wright and Bailey, 2001).\u003c/p\u003e\n\u003cp\u003eAn additional 1325 samples were compiled from 110 peer-reviewed studies conducted between 2001 and 2021 (see Supplementary Table S1 for data information and references), identified through comprehensive searches on platforms such as Google Scholar and Web of Science. These legacy datasets capture the primary vegetation and soil types of the TP with some representativeness but have uneven sampling point distribution (Fig. 1a) due to challenging geographical conditions. Sampling depths and data units were standardized across datasets, outliers were removed, and only datasets with explicitly reported sampling depths and georeferenced locations were included, with duplicated sites excluded. Samples with only SOC content were transformed into SOC density data using the following formula (Ding et al., 2019; Homann et al., 2007; Yang et al., 2010).\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAATQAAABCCAYAAADZjSr3AAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsMAAA7DAcdvqGQAAAsnSURBVHhe7Z3Ji9TOH4bj7+42HgUPLgdRUERxwAUUHJeDFxEVQQQPbuBRQfGmuBwFdQQP3lTQi6CoF8ENxwUUBAUZEbeTC/oPzK+fmny+1pSV7nR3Mt0d3wdCtkpSy6fe1JpMGKmRCCFEBfhfuhZCiJ5HgiaEqAwSNCFEZZCgCSEqgwRNCFEZJGhCiMogQRNCVAYJmhCiMkjQSuLVq1fJ7Nmzk61btyZXrlxJ+vr6kiVLlqRnhRBlIEEriS9fviRnzpxJ7t69m/z69St59+5d8vPnz/SsEKIMJGglsWHDhuTTp0+uhLZ79+5kaGgoGRgYSM8KIcpAglYi9+7dS1auXOm2Hz165LYvXLjg9oUQxSNBKxGqm2vWrEn3kmTfvn1Jf39/uieEKBp9bUMIURlUQhOiB3j48KFrj6X3vCrcunXLhYmwFQYlNCFE93L58uWRxYsXj3z79i098oeXL19mniuDBw8ejGzZsoVanVumTp06cvjw4fTs6Hk7Fy5r165NXf1heHjY+Z8wFoFKaEJ0MZReaHu9fft2Mm3atPToKKdPn04WLlyYPH/+/K9zZcB4yhUrViSLFi1KagLqllmzZo0ZjrR8+fKkJnDueE1f/ltqYpasXr06dfWHmTNnurARxiJKnxK0NqDHcsKECYUsGKcQIceOHUtOnDjxl2AdOXIkmT9/flIrEaVHyuX9+/fJtm3bklOnTiUHDx50/mHZvHlzsmDBgtTVKC9evEhqpbh0bxTEDP/G4D579+5NDh06lB5pA4ppojUo5lNcJhopejdb7PevrxlKelSIUaiOYRuss7AqXtnUBGekVupK9+qDf6gKNwPuG4U1DyqhtQFvlosXL7ptit3bt29323nh+qtXr6Z7Qozl6dOnbk21rJN8//49OX/+/F+lrhg09ANVYat9UJpshJXy3r5969atIkFrExJicHDQbd+5c6fpgbMYK+0NQoR8/PjRtT0Vwbp168Y0ccQW3MR48+aNW69fv96t68EA8lqtY0zb2a5du9KzjXn9+nW61RoStAJgapO9vfbs2dN04yaT2IUoExreTWSyFtzEePz4sVvT4N8IBpP7U/y4p1/CpPG/zNkyErSCOHv27H8lrU2bNrliel5IdBpahQhpxo7KYvLkyelWfeg4oMd12bJl6ZG/OXfunCsAlIUErSBoD7t+/brbHh4eTo4ePeq2hWiVGTNmOIGox+/fv90aMSkLm77n98QzhCOsWVy7ds2tJ02a5NYhVGn5jFY9snpCc1MraooCobeSaGUparCg+Dep18uJbTFQ1WyNXnb2m+1dzAvPo5fTnsd2vQG1MWxQbgy7vt1eTs3lLAHeRHQQ1IzMvWE73UslehdsiTFcVWiSsLazWJWTnlDGr2W14+VFglYCtHvMmTPHDeWgx+fZs2fpGSGag5kCGzdudB8IpVmjl2He5o4dO9w23ws0qC6TT/jcVjhIt1kKb0MjM9OTQV3Z7w4OGzepg/NJanPDNVkNoKFb6u4cM7i/nbOlkxN5MbwbN264bUpo7c4CyBunNoHZd5MVB43itEzsmf6CX8LnE28xdzbWqWyIz/D5/tKoPagI6FmkIT2W3r3IgQMHkunTp6d7o7bNCAHC2K6YOSihFQn1eEa/W12YOj31ex9GHXPM2phwS508Vr8O3dr9/PYpRtwTFN8N9+JYWW0KeaCNAT+w0EbQKnnilLDzHNrwiA8W4iB0B3nitB712mq4982bN9O9OJzHrxYe/Mp1HAsh3JwD3Fsb5eDgoDuWB9IhK2wc99uCfHgGdslzQxvDH8RDIwir+T+EOMxzD7D2p07ac9FYmNrJGyGFC5qf6DEwEtyEgTBD9eE+MbdhhuI87iyDAAZIJs0ypvGCDInf8At+agWurxenxAVuiEMfixc//vLGaT24B+EJ3RPXhLdROPEn7nzMXyGxsONXhCYv+DMm2FnhMAiPxVMowhwP4zsGceGLspHlJ9EehQsahkYC+uLiQyLGRCYmaNwrjyDZmzQEw2/0BuSZjZY8hpsF8UCYuU+sBJqHRnHKfWPhjwla3jhtRCgG3DOPmAH+9eMUsYhlbu6N/8Nwx2ylEaGANBKzEEpxsTjOQyhqErPyKFzQSCwSHoMLE8zecjEjIrF9gzG3rBtBBoll0jyCNh5YWJrNhEa9OCWzcDxWBcMt50wQmonTPJgoEP95xQy4Bn/Ywn5YYoSsF1UrggYmJNgK65gdZkH4sqqmeTBRI654tsSsHAoXNCDxrP3IT7h6hojh+qLUjNFmGQjXt2OERWGlNL9U0ixZcRorhRnhS6JVIagHL4ys58dARHBv4kfckMljL556LyrEoRUsDppJC/zKNe2+CCyt8rxkcfcvL61SiqAZJJyfeOG+QQYlEH6myHIbYhkkrJbYPRu9hS0C6y3tCBGQ+fKEJQ9hvGSJFPER+j1vnCKCeUovlKAQ6mZKaFwTipGlVXh97EVl4fJLpHn9a7bSbAkty3/AOcLfCK7FnyqhlUupgobh+BnIDN/HEjrMaLjNk/li1RIM1TJap0FQ8EsouK0Sxinhj2U2yzj+8bxxmgcypC8K+CuPqOGHsNRsJU8fEx8/3rg3zyC9Gz0nxGzChCT0fz0sbK1i/uY+EPpFFEdhgoZR+oZK8ZxE84vpJCBGaiUxEpaEjmWE0C3nEYdQpPwMghszVDJus0ZfNFbFaNVw88RpWBJjn4wTy6x54pTzYRyHWByH988jar7/cWeC7IcTwpIc17Aflsby+DdLQLLC4YMfcRN7ERB3PL/R9fjbxMzI8pNoj0IFDWMjgVlIRD/jGRiBuWPNflYG4J4kun9P3wC4v52zBcPrBiMhTISvUWarR944tcyOG+KLzJNVImwUp6SHX52LQRxnZWKeHfMj8Gx7btbzgXv74TZ3MVvJ41+em2UTHOd8FoTV/MCzQojLehAXoZgZhDMmlKJ1Sq1y/ssgZGTKLLHuVvB3vRJHt9FJ//Lcdl5Yonj0+aASYAoPn9a+dOlSz82/4wN9EydOLPZfiSXSSf8+efLE/QFpvKaMFQFT4Zg+FptGZVPsbGoX0+hCd1xvUw37+vpyfV57XEmFTRQEVT2qIbHqSS+A36kG9UrJspP+parL87uhiSMP2CRZPivbE49U7YlL7Jgahl8l5jjhtSq0tRF3k61L0AoGg2ARopugnZD2PBOhEOtc8ts/rcPG2mMRbvb9l4c1rXQLqnIWCF+HqCW+/uQkuo7jx4+P+WRPiP1haunSpW4N8+bNc+uvX7+69f37991nfvxmFKrc2Hy3IEErCNpw+FEqP4XVBx1Fr8EfpiDW5ms/SeG7ZVltwt3S5ipBKwAaTnfu3Om+69TKDyD4kicNrUJ0I3l+kkLHTDcgQSuA/fv3Jz9+/HB/fmoFvtQpRCcx0Yr1flrVc8qUKdHzUMjHGQtAgtYmNkSDL9S2MkSDojrXq5oqOkl/f79bDw0NuTV8/vzZrefOnevWq1atcl9g9kXtw4cPrl2ta0g7B0QL0PtDNzbR2O7Sq8M8RG9BLyb2Zj2XPtZDb8M22PZnUZi9+8M22M+aGdIJJGhtQJd1KEytLhI0USYMuWBMmdkbQsS+P8vCpmLZecQsHN+HeNm0NNaNpp2NN/rrkxCiMqgNTQhRGSRo4wAdB8yLE0KUi6qcQojKoBJaydjPcjv102Mh/iVUQhsH+MwKA2+FEOWiElrJUDIbGBhI94QQZSJBK5le/AigEL2KBG0cOHnyZLolhCgTtaEJISqDSmhCiMogQRNCVAYJmhCiMkjQhBCVQYImhKgMEjQhRGWQoAkhKkKS/B8DlCYn1s6yTwAAAABJRU5ErkJggg==\" width=\"308\" height=\"66\"\u003e\u003c/p\u003e\n\u003cp\u003eWhere \u003cimg src=\"data:image/png;base64,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\" width=\"258\" height=\"24\"\u003e density (kg m\u003csup\u003e\u0026minus;2\u003c/sup\u003e), SOC content (g kg\u003csup\u003e\u0026minus;1\u003c/sup\u003e), bulk density (g cm\u003csup\u003e\u0026minus;3\u003c/sup\u003e), thickness (cm), and volume percentage of the coarse fragment at layer \u003cem\u003ei\u003c/em\u003e, respectively (Ding et al., 2019; Yang et al., 2008). BD and CF values were extracted from SoilGrids 2.0 (Poggio et al., 2021).\u003c/p\u003e\n\u003cp\u003eWe used measured data from 460 samples to establish strong linear relationships between SOCD at different depth intervals (see Supplementary Table S2 for details), enabling us to estimate missing values and obtain complete carbon density data at six depth intervals (0-10 cm, 10-20 cm, 20-30 cm, 30-50 cm, 50-100 cm, 100-200 cm) for all samples. \u003c/p\u003e\n\u003cp\u003e\u003cem\u003eSoil properties\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eSoil physical properties, including bulk density (BD), coarse fragments (CF), and soil texture (sand, silt, clay), as well as soil chemical properties such as cation exchange capacity (CEC), nitrogen (N), and pH, were extracted from SoilGrids data (Poggio et al., 2021). Soil moisture (SM) data was derived from a daily soil moisture dataset over China (Li et al., 2022b), which was based on site observations and obtained from the National Tibetan Plateau Data Center (https://data.tpdc.ac.cn/home). Soil heterotrophic respiration (SRh) data, derived from the Soil Respiration Database Version 5 (SRDB-V5) using a quantile regression forest model, were extracted from a global dataset available at the Oak Ridge National Laboratory Distributed Active Archive Center (https://daac.ornl.gov/).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eTopography\u003c/em\u003e\u003cem\u003e and climate\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eDigital elevation (Elev) data (https://lta.cr.usgs.gov/GTOPO30) with a resolution of 30 meters was acquired and used for the extraction of slope and aspect via ArcGIS. Precipitation, air temperature, and radiation (Rad) data were collected from the National Meteorological Information Center (http://data.cma.cn/) and interpolated into raster data using ANUSPLIN (Hutchinson, 2004). The interpolated data explained over 90% of the actual observed data according to validation results (Li et al., 2019). The MODIS data of the land surface temperature (LST) and imagery data for calculating the land surface water index (LSWI) were obtained from the Land Processes Distributed Active Archive Center (https://lpdaac.usgs.gov/). Annual maximum and minimum temperature data were downloaded from the National Tibetan Plateau Data Center (https://data.tpdc.ac.cn/home). Evapotranspiration (ET) data (Zheng et al., 2022) and potential evapotranspiration data (Peng, 2022) were obtained from the National Tibetan Plateau Data Center (https://data.tpdc.ac.cn/home). The aridity index (AI), calculated as the ratio of potential evapotranspiration to precipitation (Thornthwaite, 1948), was used to assess regional dryness and its potential impact on SOC.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eVegetation information\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe Normalized Difference Vegetation Index (NDVI), enhanced vegetation index (EVI), and net primary productivity (NPP) estimates were downloaded from the Land Processes Distributed Active Archive Center (https://lpdaac.usgs.gov/). Leaf area index (LAI) data was obtained from the National Centers for Environmental Information (https://www.ncei.noaa.gov/). Vegetation type data was obtained from the vegetation atlas of China, available at the Resource and Environment Science and Data Center of the Chinese Academy of Sciences (https://www.resdc.cn/Default.aspx). \u003c/p\u003e\n\u003cp\u003e\u003cem\u003eData extraction\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eFor data sources in raster format, values were extracted based on the coordinates of sampling points to match the SOC data. To ensure temporal alignment between environmental variables and SOC data, variables exhibiting significant interannual variability, such as climate factors and vegetation indices, were extracted based on the year of soil sample collection. Soil physicochemical variables were extracted from gridded datasets at standard depths (0\u0026ndash;10 cm, 10\u0026ndash;20 cm, 20\u0026ndash;30 cm, 30\u0026ndash;50 cm, 50\u0026ndash;100 cm, 100\u0026ndash;200 cm) to match SOC data. When direct depth data were not available, we used linear interpolation between neighboring depth layers to estimate values.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eDimensionality reduction of variable data\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTo assess the relative influence of different types of variables on SOCD, the 26 environmental factors used in this study were grouped into eight categories (Fig. 6) based on their attributes: topography (Topog), soil physical properties (SPP), soil chemical properties (SCP), soil heterotrophic respiration (SRh), climate, solar radiation (Rad), vegetation-related indices (VRI), and net primary productivity (NPP). For dimensionality reduction, we first conducted individual linear regressions between the variables in each category and SOCD. The resulting regression coefficients were then used as weights to aggregate the variables into composite indices for each category (e.g., climate). This procedure yielded category-level factors while preserving the relative contribution of individual variables.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eRandom Forest model and validation\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eRandom Forest (RF) is widely used for digitally mapping SOC (Lamichhane et al., 2019; Yang et al., 2023). In this study, RF was employed to map SOCD at different depths and evaluate the importance of factors influencing SOCD. The \u0026lsquo;randomForest\u0026rsquo; package (Liaw and Wiener, 2002) in the R environment (R Core Team, 2013) was utilized. To ensure robust model performance, we randomly split the dataset into training (80%) and testing (20%, n = 312) subsets. The random split was stratified by vegetation type to preserve class distribution across subsets. Based on SOCD at different depths and corresponding soil physicochemical properties data, along with other variables, RF models were separately trained. 5-fold cross-validation was used to select the optimal parameters for the RF models. Statistical indices of the coefficient of determination (R\u003csup\u003e2\u003c/sup\u003e), mean absolute error (MAE), and root mean square error (RMSE) were used to quantify the generalization capability of the RF models. The models for SOCD estimation at depths of 0\u0026ndash;30 cm, 30\u0026ndash;50 cm, 50\u0026ndash;100 cm, and 100\u0026ndash;200 cm achieved R\u0026sup2; values of 0.61, 0.64, 0.66, and 0.67, respectively (mean R\u0026sup2; = 0.65) (Fig. 2). The models showed acceptable generalization capability across depths, though a slight underestimation was observed for high SOCD values likely due to sample scarcity in that range (Fig. 2).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eStatistical analysis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe partial correlation method was utilized to assess the correlation between pairs of variables. In exploring variations in SOCD across distinct vegetation types, the one-way analysis of variance (ANOVA) was implemented. Significance in statistical analysis was evaluated through T-tests or F-tests, with the level of significance indicated by the p-value. Piecewise fitting was conducted on SOCD at various depth intervals to uncover vertical distribution characteristics. Structural Equation Modeling was employed to test hypothesized causal relationships between environmental factors and SOCD, based on ecological knowledge and observed correlations. The model fit was assessed using common statistical criteria, including the chi-square test (\u0026chi;\u0026sup2;), and comparative fit index (CFI), with acceptable model fit indicated by \u0026chi;\u0026sup2;/df \u0026lt; 3, CFI \u0026gt; 0.9. The SEM analysis was conducted using the lavaan package in R (Rosseel, 2012). \u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cem\u003eSpatial distribution characteristics of SOCD\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eVertical distribution\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eOur findings indicate that in superficial soil layers (above 20cm), there was a slight increase in SOCD per 10 cm thickness (SOCD\u003csub\u003ep10\u003c/sub\u003e) with depth, whereas in deeper soil layers (below 20cm), there was a non-linear decrease in SOCD\u003csub\u003ep10\u003c/sub\u003e (Fig. 3). According to the piecewise fitting analysis, the relationship between SOCD\u003csub\u003ep10\u003c/sub\u003e and soil depth (Dep) is as follows (Fig. 3b):\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"data:image/png;base64,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\" width=\"382\" height=\"67\"\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eHorizontal distribution\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eGiven the minor differences in SOCD across the 0\u0026ndash;10 cm, 10\u0026ndash;20 cm, and 20\u0026ndash;30 cm soil layers, as shown above, these depths were combined into a single 0\u0026ndash;30 cm interval in subsequent analyses to enhance figure clarity and improve the overall readability of the manuscript. Based on the validated RF models and the grid data corresponding to the 26 variables used for model training, the SOCD distributions at four different depths were predicted. The results revealed that SOCD\u003csub\u003ep10\u003c/sub\u003e exhibited similar spatial patterns at different depths across the TP, with generally higher values in the eastern and southern regions compared to the west and north(Fig. 4). These patterns were further supported by the 1561 field-measured SOC observations collected in this study (Fig. A.1a). Specifically, SOCD\u003csub\u003ep10\u003c/sub\u003e at a depth of 200 cm exhibited an increasing trend along the longitudinal gradient, indicating higher SOCD values in eastern regions relative to western areas (Fig. A.1a). Along the latitudinal gradient, a general decline in SOCD with increasing latitude was observed, although the pattern was not entirely uniform (Fig. A.1b). Notably, higher SOCD levels were detected in the southern regions compared to the northern ones, aligning with the model predictions.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eDifferences in SOCD distribution across vegetation types and depths\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe distribution of SOCD\u003csub\u003ep10\u003c/sub\u003e across six vegetation types\u0026mdash;wetland, forest, shrubland, meadow, steppe, and others\u0026mdash;was meticulously analyzed across four soil depth intervals (Figs. 1a and 5). Here, \u0026ldquo;others\u0026rdquo; refers to other vegetated land types not included in the primary categories. Wetlands and forests exhibited the highest SOCD\u003csub\u003ep10\u003c/sub\u003e, followed by meadows and shrublands, while steppes and other vegetation types\u0026nbsp;had\u0026nbsp;comparatively lower levels of SOCD\u003csub\u003ep10\u003c/sub\u003e (Fig. 5). This pattern of variation in SOCD\u003csub\u003ep10\u003c/sub\u003e among different vegetation types remained consistent across varying depth intervals (Fig. A.2). The SOCD\u003csub\u003ep10\u003c/sub\u003e exhibited significant differences across various depth intervals, with deeper intervals generally associated with lower SOCD\u003csub\u003ep10\u003c/sub\u003e (Fig. A.2). Notably, this pattern of divergence remained consistent across diverse vegetation types (Fig. 5).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eInfluencing factors of SOCD\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eFigure 6 presents the correlation analysis results between SOCD and 26 individual factors, as well as between SOCD and eight aggregated factor categories. The results indicate that most of the factors were significantly correlated with SOCD (left part of Fig. 6). Meanwhile, among the eight categories, VRI consistently showed the highest correlation with SOCD across all depth intervals (right part of Fig. 6). Specifically, all vegetation-related indices (i.e., NDVI, EVI, and LAI) showed significant positive correlations with SOCD. Following VRI, soil physical properties emerged as the second most influential category, with properties such as silt, soil moisture (SM), and bulk density (BD) showing notable associations with SOCD. Climate factors also played a role, with stronger correlations from moisture-related variables like MAP, LSWI, and ET compared to temperature-related factors such as MAT and LST. NPP and SRh, representing the input and output of SOC respectively, exhibited nearly equal strength in correlations with SOCD. Soil chemical properties and topography exhibited relatively weaker correlations with SOCD. Notably, pH negatively correlated with SOCD, while N showed a positive correlation. Elevation and slope also impacted SOCD, with elevation having a negative effect and slope a positive one. Additionally, SOCD was negatively correlated with solar radiation. Vertically, the correlation strength between SOCD and the eight categories of factors generally diminished with increasing soil depth, particularly for vegetation-related indices and climate factors.\u003c/p\u003e\n\u003cp\u003eStructural Equation Modeling (SEM) results, which included the eight categories of factors, indicate that NPP (path coefficient = 1.74) and SRh (path coefficient = 2.41) were the dominant factors influencing SOCD (Fig. 7). Rad, Climate, SCP, SPP, Topog, and VRI primarily affected SOCD either directly or indirectly through their impact on NPP and SRh. For instance, Rad affected VRI and NPP by influencing the climate, subsequently impacting SOCD. Similarly, climate influenced SOCD through its effects on SPP, SCP, VRI, and NPP. Additionally, SRh was primarily influenced by SOCD, followed by climate and soil properties.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eVariations in factors influencing SOCD across vegetation types and depths\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe correlation between SOCD and influencing factors differed significantly among vegetation types (Fig. 8a). Some factors correlated positively with SOCD in certain vegetation types but negatively with others. For instance, LST correlated positively with SOCD in steppes and wetlands but negatively with other types. Additionally, the strength of correlations, as indicated by correlation coefficients and their significance levels, also varied among vegetation types. The relative importance of factors influencing SOCD varies markedly across different vegetation types (Fig. 8b, Fig. A.3). Tmin and Rad are the most important factors for forest and meadow, respectively, while LAI emerges as the dominant factor for other vegetation types, particularly steppe. Except for shrubland, most vegetation types, especially meadows, and steppes, had a considerable number of factors with importance exceeding 10%, highlighting the significant role multiple factors play in explaining SOCD variation (Fig. 8b, Fig. A.3). Correspondingly, the RF models explained 40\u0026ndash;60% of SOCD variation across vegetation types, with the highest explanatory power in steppes and the lowest in shrublands (Fig. A.3). As depth increased, the number of important factors decreased across all vegetation types (Fig. A.3), indicating fewer factors explaining SOCD variation in deeper layers.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003e\u003cem\u003eDistribution features of SOCD\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eIn this study, we analyzed the three-dimensional distribution of SOCD across different vegetation types and depth intervals. The expansive TP, marked by considerable geographical diversity encompassing variations in climate, topography, vegetation, and landscape (Jia et al., 2023; Wang et al., 2023), leads to spatial heterogeneity in SOCD distribution (Gao et al., 2023). Our findings demonstrate that the SOCD on the TP exhibits a spatial pattern characterized by a gradual decrease from southeast to northwest, as supported by both sample data analysis (Fig. A.1) and simulation results (Fig. 3). This pattern aligns with previous studies\u0026nbsp;(Gao et al., 2023; Han et al., 2022; Yang et al., 2023; Yang et al., 2010). Vertically, SOCD exhibited a nonlinear distribution across the 0-200 cm depth, slightly increasing within the 0-20 cm surface layers and decreasing at greater depths (Fig. 4). The overall decreasing trend is consistent with previous studies on the TP\u0026nbsp;(Ding et al., 2016; Wang et al., 2023)\u0026nbsp;and other regions\u0026nbsp;(Zhang et al., 2018). Similar trends in surface layers were observed in other regions\u0026nbsp;(Rose et al., 2017). This depth-dependent SOCD distribution highlights the concentration of biologically mediated carbon inputs in topsoil.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Our findings also reveal that SOCD varied notably across different vegetation types, and this ranking remained consistent across different depth intervals, with the highest values observed in wetlands, followed by forests, meadows, shrublands, steppes, and other vegetation types (Fig. 5, Fig. A.2). Wetlands exhibited the highest SOCD due to limited decomposition under water-logged conditions (Jiang et al., 2019), covering only 2.6% of the plateau but representing over 6% of the total SOC stock within the upper 100 cm of soil (Chen et al., 2022). In contrast, forests on the plateau have higher net ecosystem productivity but also high soil respiration rates (Chen et al., 2022), resulting in them containing only 8% of the total SOC stock on the TP and having a lower SOCD than wetlands (Chen et al., 2022; Jiang et al., 2019). This study confirms that forests exhibit higher SOCD than grasslands (meadows and steppes), which aligns with findings at the national scale (Xu et al., 2019), though contrasts with localized studies (Wang et al., 2023). Among grassland types, meadows had higher SOCD than steppes, in agreement with prior research (Chen et al., 2022; Ding et al., 2017; Ding et al., 2016). These patterns suggest that vegetation type significantly influences soil carbon storage, which is relevant for optimizing land management and enhancing ecosystem carbon sinks\u0026nbsp;(Post and Kwon, 2000).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eEffects of environmental factors on SOCD\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eNet primary production and soil heterotrophic respiration, representing the carbon inputs and outputs of SOC, are the most direct environmental factors influencing SOCD in this study (Fig. 7). Among the eight categories of factors, vegetation-related indices had the strongest correlation with SOCD (Fig. 6). This strong relationship likely because vegetation indices reflect vegetation status and biomass, which are closely related to carbon inputs from plant litter and root turnover (Gao et al., 2013; Ma et al., 2024; Schloss et al., 1999), suggesting that vegetation indices can serve as effective indicators of soil carbon storage potential (Kunkel et al., 2022).\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Soil physical properties were the second most influential factors affecting SOCD, followed by climatic factors, NPP, soil heterotrophic respiration, soil chemical properties, topography and solar radiation (Fig. 6). SPP impacted SOCD by modulating vegetation growth (indicated by VRI) (Saporetti-Junior et al., 2012), which subsequently affects carbon inputs (indicated by NPP) (Fig.7). Additionally, SPP had a significant effect on soil respiration (Chen et al., 2014), which relates to the carbon output of SOC. For\u0026nbsp;instance, soil moisture showed a significant positive correlation with VRI and NPP, while it negatively correlated with heterotrophic respiration (Fig. 6). Similarly, SCP, as a crucial aspect of soil properties, also affected SOC\u0026apos;s input and output. For instance, enhances photosynthesis and biomass production, increasing carbon input to the soil, while also stimulating microbial activity and decomposition\u0026mdash;both of which influence SOCD\u0026nbsp;(Janssens et al., 2010; Leghari et al., 2016).\u003c/p\u003e\n\u003cp\u003eClimate factors primarily influenced SOCD indirectly by regulating key ecological processes, including soil respiration and vegetation activity (Fig.7). For example, temperature simultaneously enhanced plant productivity\u0026nbsp;(Zhong et al., 2019) and microbial decomposition (Ma et al., 2022), while precipitation shaped soil moisture regimes (Sehler et al., 2019) and promoted biomass accumulation (Zhang et al., 2021). Topographic and geomorphic factors modulated these climate\u0026ndash;vegetation\u0026ndash;soil interactions (Fig. 7). For instance, altitude affected climate patterns (Grabherr et al., 2010), such as temperature, which in turn affected plant growth and respiration rates. Additionally, solar radiation influenced the carbon cycle both directly through its effect on photosynthesis (Ballare et al., 2011) and indirectly via its role in climate modulation (Lean and Rind, 1998) (Fig. 7).\u003c/p\u003e\n\u003cp\u003eCollectively, these findings indicate that SOCD is shaped by the synergistic effects of vegetation dynamics, soil characteristics, and climate conditions. These environmental drivers interact through coupled pathways that control carbon inputs via biomass production and outputs through microbial respiration. The integration of above- and belowground processes highlights the importance of a systems-level perspective for understanding the mechanisms governing soil carbon dynamics across ecosystems (Crowther et al., 2016; Luo et al., 2016).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eDivergent controls of SOCD in different ecosystems\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;The impact of environmental factors on SOCD varies across regions and land use types (Wang et al., 2021; Wang et al., 2023). Our findings (Fig. 8) show that the same environmental factor can exert opposite effects depending on vegetation types. For example, increasing altitude tends to enhance SOCD in forests and shrublands but decreases it in grasslands and wetlands. Similarly, higher temperatures (MAT, LST) promote soil carbon accumulation in steppes and wetlands while reducing it in forests and shrublands. These contrasting responses are largely shaped by the vegetation\u0026apos;s geographic distribution and the ecological traits of major vegetation types. Grasslands and wetlands, primarily distributed at higher altitudes of the northwestern TP (Hu et al., 2023) (Fig. 1), suffer from temperature constraints that limit plant productivity and carbon inputs (Ma et al., 2018; Piao et al., 2019). As elevation rises and temperature drops, plant productivity diminishes (Wang et al., 2020), leading to decreased soil carbon input and ultimately lowering SOCD (Teron et al., 2024). In contrast, forests and shrublands are mainly distributed in lower, warmer southeastern regions (Chen et al., 2020) (Fig. 1), where elevation-driven cooling reduces respiration and decomposition rates (Davidson and Janssens, 2006; Ma et al., 2022), thereby lowering carbon output and increasing SOCD (Xu et al., 2010).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThis divergent response arises from ecosystem-specific thresholds in carbon cycling, especially in how soils stabilize or release organic carbon (Schmidt et al., 2011). Forests and shrublands typically possess higher microbial carbon-use efficiency and more mineral-associated organic matter, enabling better carbon retention in cooler climates (Liang et al., 2017). In contrast, grasslands and wetlands in higher regions depend more on continuous biomass input and are thus more sensitive to temperature-driven changes in productivity\u0026nbsp;(Clemmensen et al., 2013; Crowther et al., 2016). Collectively, these mechanisms explain why cooling increases SOCD in warm-region forests and shrublands by reducing carbon output more than input, but decreases SOCD in cold-region grasslands and wetlands by limiting input more than output\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Building on these ecosystem-specific mechanisms, we further found that both NPP and SRh were negatively correlated with SOCD in forests, but positively correlated in other vegetation types (Fig. 8). These contrasting correlations reinforce the notion that carbon cycling is governed by biome-specific regulatory processes.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn forests, the contribution of carbon input (indicated by NPP) to SOCD may be outweighed by carbon losses via respiration. For instance, rising temperatures may enhance NPP, but\u0026mdash;as discussed above\u0026mdash;this warming-driven increase is likely modest. Moreover, much of the additional NPP may be sequestered in aboveground biomass (Arag\u0026atilde;o et al., 2009), limiting its contribution to soil carbon input. Furthermore, forest soils often have compact structure and deeper roots that promote microbial respiration, enhancing carbon loss (Qi and Yang, 2017). As a result, warming exerts a stronger influence on accelerating SOC decomposition than on enhancing carbon input, indicating that in forested regions, soil respiration (carbon output) largely determines the dynamics of SOCD.\u003c/p\u003e\n\u003cp\u003eConversely, in non-forest ecosystems such as grasslands, carbon inputs from NPP may surpass carbon losses through respiration. As discussed earlier, warming appears to boost soil carbon inputs in grassland systems, likely due to their typically higher belowground carbon allocation and faster organic matter turnover into soil (Herold et al., 2014; Trumbore et al., 1995). This is partly because grassland soils typically have looser structure, shallower roots, and distinct microbial communities, which favor fast carbon incorporation into soil (Juarez et al., 2013). However, the capacity of warming to accelerate carbon decomposition may be constrained\u0026mdash;potentially due to the relatively limited availability of labile carbon substrates in these soils (Zhang et al., 2005), especially in the northwest grassland regions (Fig. 4). Thus, in non-forest ecosystems, like grasslands, SOCD is primarily governed by carbon input (indicated by NPP). This stands in contrast to forest ecosystems, highlighting biome-specific mechanisms driving variation in soil organic carbon across ecosystems.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eLimitations and implications\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThis study compiled over 1,500 SOCD samples to simulate its distribution across the TP. However, the region\u0026rsquo;s complex terrain resulted in uneven spatial coverage, and sample sizes varied across vegetation types\u0026mdash;forests and shrublands were underrepresented compared to grasslands\u0026mdash;potentially introducing bias and increasing uncertainty. While these limitations affect spatial representativeness and interpretative depth, our analysis of SOCD drivers focuses on measured rather than simulated data. Moreover, the lack of comparative experiments constrains the empirical strength of some mechanistic explanations, despite extensive literature support.\u003c/p\u003e\n\u003cp\u003eDespite these limitations, the study provides a high-resolution SOCD distribution dataset covering the entire plateau, offering essential data support for future assessments of regional carbon sinks and management strategies. The depth-dependent SOCD equation we developed offers a scalable approach for estimating subsoil carbon stocks, a component often overlooked in regional and global carbon inventories.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eImportantly, our results reveal that environmental factors influence SOCD in a context-specific manner, with variations across different vegetation types. This highlights the need for incorporating vegetation-specific parameters, such as root biomass allocation and microbial turnover, into carbon cycle models to improve their accuracy. Understanding these variations can refine model predictions and better capture the dynamics of soil carbon in diverse ecosystems.\u003c/p\u003e\n\u003cp\u003eFrom a management perspective, our findings support the development of targeted carbon sequestration strategies. For example, enhancing vegetation productivity in alpine meadows may be more effective, while focusing on improving soil physical conditions could benefit alpine steppes. These insights can inform region-specific policies for carbon neutrality and land-based climate mitigation efforts.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThis study provides the first comprehensive assessment of both horizontal and vertical patterns of SOCD and its environmental drivers across diverse vegetation types and soil depths (0-200cm) on the TP, based on analysis of over 1561 soil samples. Spatially, SOCD exhibited a decline from humid southeast to the arid northwest, aligning with variations in productivity, moisture, and organic matter stabilization potential. Vertically, SOCD followed a nonlinear distribution, peaking at approximately 2.58 kg C m⁻\u0026sup2; per 10 cm within the 0-20 cm surface layers before decreasing at greater depths, indicating the concentration of biologically mediated carbon inputs in topsoil. SOCD varied by vegetation type, with the highest levels in wetlands (29.54 kg C m⁻\u0026sup2;), followed by forests (27.06 kg C m⁻\u0026sup2;), meadows (21.78 kg C m⁻\u0026sup2;), shrublands (21.47 kg C m⁻\u0026sup2;), and steppes (14.96 kg C m⁻\u0026sup2;) at depth of 0\u0026ndash;200 cm. The SOCD on the TP was influenced by a combination of environmental factors, including vegetation productivity, soil respiration, climate, topography, soil properties, and solar radiation. Overall, SOCD is primarily driven by vegetation growth and soil respiration\u0026mdash;reflected by standardized path coefficients of 1.74 and \u0026ndash;2.41, respectively\u0026mdash;which control carbon inputs and outputs. Environmental factors modulate these processes to varying degrees across vegetation types, with their influence diminishing with soil depth. In forests, SOCD was predominantly controlled by carbon losses through soil respiration, while in grasslands and other vegetation types, carbon inputs via plant productivity were the primary determinant. Mechanistically, these differences may be attributed to variations in soil structure, microbial composition, and root depth across ecosystems. These findings elucidate divergent mechanisms regulating soil organic carbon across ecosystems, enabling more accurate modeling of soil carbon\u0026ndash;climate interactions through ecosystem-specific calibrations of respiration rates, carbon input pathways, and decomposition parameters. Furthermore, the study highlights the need for region-specific management strategies for carbon sequestration to better support carbon neutrality and land-based climate mitigation efforts.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eCRediT authorship contribution statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eZ. Hu:\u0026nbsp;\u003c/strong\u003eWriting \u0026ndash; original draft, Writing \u0026ndash; review \u0026amp; editing, Investigation, Methodology, Formal analysis, Data curation, Visualization, Validation, Software, Conceptualization. \u003cstrong\u003eX. Zhang and J.C. Svenning:\u0026nbsp;\u003c/strong\u003eSupervision, Conceptualization, Resources, Writing \u0026ndash; review \u0026amp; editing, Funding acquisition. \u003cstrong\u003eB. Niu:\u0026nbsp;\u003c/strong\u003eFunding acquisition, Conceptualization. \u003cstrong\u003eJ. Tang:\u0026nbsp;\u003c/strong\u003eInvestigation, Data curation. \u003cstrong\u003eY. Yue, W. Wu, M. Ni, Y. Yang:\u003c/strong\u003e Writing \u0026ndash; review \u0026amp; editing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData will be made available on request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was jointly supported by the Science and technology Projects of Xizang Autonomous Region, China (grant numbers, XZ202501ZY0087, XZ202501ZY0118), the West Light Foundation of the Chinese Academy of Sciences (grant number, xbzg-zdsys-202202), and the Second Tibetan Plateau Scientific Expedition and Research Program (STEP) (grant number 2019QZKK1002). We also consider this work a contribution to Center for Ecological Dynamics in a Novel Biosphere (ECONOVO), funded by Danish National Research Foundation (grant DNRF173 to JCS).\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eArag\u0026atilde;o L, Malhi Y, Metcalfe D, Silva-Espejo JE, Jim\u0026eacute;nez E, Navarrete D, et al. Above-and below-ground net primary productivity across ten Amazonian forests on contrasting soils. Biogeosciences 2009; 6: 2759-2778.\u003c/li\u003e\n\u003cli\u003eBallare CL, Caldwell MM, Flint SD, Robinson SA, Bornman JF. Effects of solar ultraviolet radiation on terrestrial ecosystems. Patterns, mechanisms, and interactions with climate change. Photochemical \u0026amp; Photobiological Sciences 2011; 10: 226-241.\u003c/li\u003e\n\u003cli\u003eBradford MA, Wieder WR, Bonan GB, Fierer N, Raymond PA, Crowther TW. Managing uncertainty in soil carbon feedbacks to climate change. Nature Climate Change 2016; 6: 751-758.\u003c/li\u003e\n\u003cli\u003eChen A, Huang L, Liu Q, Piao S. Optimal temperature of vegetation productivity and its linkage with climate and elevation on the Tibetan Plateau. Global Change Biology 2021; 27: 1942-1951.\u003c/li\u003e\n\u003cli\u003eChen H, Ju P, Zhu Q, Xu X, Wu N, Gao Y, et al. Carbon and nitrogen cycling on the Qinghai\u0026ndash;Tibetan Plateau. Nature Reviews Earth \u0026amp; Environment 2022; 3: 701-716.\u003c/li\u003e\n\u003cli\u003eChen J, Yan F, Lu Q. Spatiotemporal variation of vegetation on the Qinghai\u0026ndash;Tibet Plateau and the influence of climatic factors and human activities on vegetation trend (2000\u0026ndash;2019). Remote Sensing 2020; 12: 3150.\u003c/li\u003e\n\u003cli\u003eChen S, Zou J, Hu Z, Chen H, Lu Y. Global annual soil respiration in relation to climate, soil properties and vegetation characteristics: Summary of available data. Agricultural and Forest Meteorology 2014; 198: 335-346.\u003c/li\u003e\n\u003cli\u003eClemmensen K, Bahr A, Ovaskainen O, Dahlberg A, Ekblad A, Wallander H, et al. Roots and associated fungi drive long-term carbon sequestration in boreal forest. Science 2013; 339: 1615-1618.\u003c/li\u003e\n\u003cli\u003eCrowther TW, Todd-Brown KE, Rowe CW, Wieder WR, Carey JC, Machmuller MB, et al. Quantifying global soil carbon losses in response to warming. Nature 2016; 540: 104-108.\u003c/li\u003e\n\u003cli\u003eDavidson EA, Janssens IA. Temperature sensitivity of soil carbon decomposition and feedbacks to climate change. Nature 2006; 440: 165-173.\u003c/li\u003e\n\u003cli\u003eDing J, Chen L, Ji C, Hugelius G, Li Y, Liu L, et al. Decadal soil carbon accumulation across Tibetan permafrost regions. Nature Geoscience 2017; 10: 420-424.\u003c/li\u003e\n\u003cli\u003eDing J, Li F, Yang G, Chen L, Zhang B, Liu L, et al. The permafrost carbon inventory on the Tibetan Plateau: a new evaluation using deep sediment cores. Global Change Biology 2016; 22: 2688-2701.\u003c/li\u003e\n\u003cli\u003eDing J, Wang T, Piao S, Smith P, Zhang G, Yan Z, et al. The paleoclimatic footprint in the soil carbon stock of the Tibetan permafrost region. Nature Communications 2019; 10: 4195.\u003c/li\u003e\n\u003cli\u003eDokuchaev VV. Russian chernozem. St. Petersburg: Evdokimova, 1883.\u003c/li\u003e\n\u003cli\u003eGait\u0026aacute;n JJ, Maestre FT, Bran DE, Buono GG, Dougill AJ, Garcia Martinez G, et al. Biotic and abiotic drivers of topsoil organic carbon concentration in drylands have similar effects at regional and global scales. Ecosystems 2019; 22: 1445-1456.\u003c/li\u003e\n\u003cli\u003eGao M, Xu R, Huang J, Su B, Jiang S, Shi P, et al. Increase of carbon storage in the Qinghai-Tibet Plateau: Perspective from land-use change under global warming. Journal of Cleaner Production 2023; 414: 137540.\u003c/li\u003e\n\u003cli\u003eGao Y, Zhou X, Wang Q, Wang C, Zhan Z, Chen L, et al. Vegetation net primary productivity and its response to climate change during 2001\u0026ndash;2008 in the Tibetan Plateau. Science of the Total Environment 2013; 444: 356-362.\u003c/li\u003e\n\u003cli\u003eGenxu W, Yuanshou L, Yibo W, Qingbo W. Effects of permafrost thawing on vegetation and soil carbon pool losses on the Qinghai\u0026ndash;Tibet Plateau, China. Geoderma 2008; 143: 143-152.\u003c/li\u003e\n\u003cli\u003eGeorgiou K, Jackson RB, Vindu\u0026scaron;kov\u0026aacute; O, Abramoff RZ, Ahlstr\u0026ouml;m A, Feng W, et al. Global stocks and capacity of mineral-associated soil organic carbon. Nature Communications 2022; 13: 3797.\u003c/li\u003e\n\u003cli\u003eGrabherr G, Gottfried M, Pauli H. Climate change impacts in alpine environments. Geography Compass 2010; 4: 1133-1153.\u003c/li\u003e\n\u003cli\u003eHan D, Hu Z, Wang X, Wang T, Chen A, Weng Q, et al. Shift in controlling factors of carbon stocks across biomes on the Qinghai-Tibetan Plateau. Environmental Research Letters 2022; 17: 074016.\u003c/li\u003e\n\u003cli\u003eHerold N, Sch\u0026ouml;ning I, Michalzik B, Trumbore S, Schrumpf M. Controls on soil carbon storage and turnover in German landscapes. Biogeochemistry 2014; 119: 435-451.\u003c/li\u003e\n\u003cli\u003eHomann PS, Kapchinske JS, Boyce A. Relations of mineral-soil C and N to climate and texture: regional differences within the conterminous USA. Biogeochemistry 2007; 85: 303-316.\u003c/li\u003e\n\u003cli\u003eHu Z, Niu B, Tang J, Zhang Y, Xiang M, Zhang X. Has the Dominant Climatic Driver for the Carbon Budget of Alpine Grassland Shifted from Temperature to Precipitation on the Qinghai\u0026ndash;Tibet Plateau? Remote Sensing 2023; 15: 2492.\u003c/li\u003e\n\u003cli\u003eHutchinson M. ANUSPLIN Version 4.3. Centre for Resource and Environmental Studies. The Australian National University, Canberra, Australia 2004.\u003c/li\u003e\n\u003cli\u003eJanssens I, Dieleman W, Luyssaert S, Subke J-A, Reichstein M, Ceulemans R, et al. Reduction of forest soil respiration in response to nitrogen deposition. Nature geoscience 2010; 3: 315-322.\u003c/li\u003e\n\u003cli\u003eJia Z, Wang X, Feng X, Ma J, Wang X, Zhang X, et al. Exploring the spatial heterogeneity of ecosystem services and influencing factors on the Qinghai Tibet Plateau. Ecological Indicators 2023; 154: 110521.\u003c/li\u003e\n\u003cli\u003eJiang L, Chen H, Zhu Q, Yang Y, Li M, Peng C, et al. Assessment of frozen ground organic carbon pool on the Qinghai-Tibet Plateau. Journal of soils and sediments 2019; 19: 128-139.\u003c/li\u003e\n\u003cli\u003eJiang M, Li H, Zhang W, Liu J, Zhang Q. Effects of climate change and grazing on the soil organic carbon stock of alpine wetlands on the Tibetan Plateau from 2000 to 2018. Catena 2024; 238: 107870.\u003c/li\u003e\n\u003cli\u003eJobb\u0026aacute;gy EG, Jackson RB. The vertical distribution of soil organic carbon and its relation to climate and vegetation. Ecological applications 2000; 10: 423-436.\u003c/li\u003e\n\u003cli\u003eJuarez S, Nunan N, Duday A-C, Pouteau V, Schmidt S, Hapca S, et al. Effects of different soil structures on the decomposition of native and added organic carbon. European journal of soil biology 2013; 58: 81-90.\u003c/li\u003e\n\u003cli\u003eKowalenko CG. Assessment of Leco CNS-2000 analyzer for simultaneously measuring total carbon, nitrogen, and sulphur in soil. Communications in Soil Science and Plant Analysis 2001; 32: 2065-2078.\u003c/li\u003e\n\u003cli\u003eKuang X, Jiao JJ. Review on climate change on the Tibetan Plateau during the last half century. Journal of Geophysical Research: Atmospheres 2016; 121: 3979-4007.\u003c/li\u003e\n\u003cli\u003eKunkel V, Wells T, Hancock G. Modelling soil organic carbon using vegetation indices across large catchments in eastern Australia. Science of The Total Environment 2022; 817: 152690.\u003c/li\u003e\n\u003cli\u003eLamichhane S, Kumar L, Wilson B. Digital soil mapping algorithms and covariates for soil organic carbon mapping and their implications: A review. Geoderma 2019; 352: 395-413.\u003c/li\u003e\n\u003cli\u003eLean J, Rind D. Climate forcing by changing solar radiation. Journal of Climate 1998; 11: 3069-3094.\u003c/li\u003e\n\u003cli\u003eLeghari SJ, Wahocho NA, Laghari GM, HafeezLaghari A, MustafaBhabhan G, HussainTalpur K, et al. Role of nitrogen for plant growth and development: A review. Advances in Environmental Biology 2016; 10: 209-219.\u003c/li\u003e\n\u003cli\u003eLehmann J, Hansel CM, Kaiser C, Kleber M, Maher K, Manzoni S, et al. Persistence of soil organic carbon caused by functional complexity. Nature Geoscience 2020; 13: 529-534.\u003c/li\u003e\n\u003cli\u003eLehmann J, Kleber M. The contentious nature of soil organic matter. Nature 2015; 528: 60-68.\u003c/li\u003e\n\u003cli\u003eLi H, Wu Y, Liu S, Xiao J, Zhao W, Chen J, et al. Decipher soil organic carbon dynamics and driving forces across China using machine learning. Global Change Biology 2022a; 28: 3394-3410.\u003c/li\u003e\n\u003cli\u003eLi M, Wu JS, Song CQ, He YT, Niu B, Fu G, et al. Temporal Variability of Precipitation and Biomass of Alpine Grasslands on the Northern Tibetan Plateau. Remote Sensing 2019; 11.\u003c/li\u003e\n\u003cli\u003eLi Q, Shi G, Shangguan W, Nourani V, Li J, Li L, et al. A 1 km daily soil moisture dataset over China using in situ measurement and machine learning. Earth System Science Data 2022b; 14: 5267-5286.\u003c/li\u003e\n\u003cli\u003eLiang C, Schimel JP, Jastrow JD. The importance of anabolism in microbial control over soil carbon storage. Nature microbiology 2017; 2: 1-6.\u003c/li\u003e\n\u003cli\u003eLiaw A, Wiener M. Classification and regression by randomForest. R news 2002; 2: 18-22.\u003c/li\u003e\n\u003cli\u003eLuo Y, Ahlstr\u0026ouml;m A, Allison SD, Batjes NH, Brovkin V, Carvalhais N, et al. Toward more realistic projections of soil carbon dynamics by Earth system models. Global Biogeochemical Cycles 2016; 30: 40-56.\u003c/li\u003e\n\u003cli\u003eMa H, Peng M, Yang Z, Yang K, Zhao C, Li K, et al. Spatial distribution and driving factors of soil organic carbon in the Northeast China Plain: Insights from latest monitoring data. Science of the Total Environment 2024; 911: 168602.\u003c/li\u003e\n\u003cli\u003eMa M, Yuan W, Dong J, Zhang F, Cai W, Li H. Large-scale estimates of gross primary production on the Qinghai-Tibet plateau based on remote sensing data. International Journal of Digital Earth 2018; 11: 1166-1183.\u003c/li\u003e\n\u003cli\u003eMa X, He P, Zeng Y, Ma J, Wu X. A global-drive analysis of ecosystem respiration in the Arctic and Third Pole. Ecological Indicators 2022; 145: 109668.\u003c/li\u003e\n\u003cli\u003eMeyer RS, Cullen BR, Whetton PH, Robertson FA, Eckard RJ. Potential impacts of climate change on soil organic carbon and productivity in pastures of south eastern Australia. Agricultural Systems 2018; 167: 34-46.\u003c/li\u003e\n\u003cli\u003ePeng S. 1km Monthly Potential Evapotranspiration Dataset in China (1990\u0026ndash;2022). National Tibetan Plateau / Third Pole Environment Data Center, Beijing, China, 2022.\u003c/li\u003e\n\u003cli\u003ePiao S, Niu B, Zhu J, Zhang X, Wang T, Wang S, et al. Responses and feedback of the Tibetan Plateau\u0026apos;s alpine ecosystem to climate change. Chinese Science Bulletin 2019; 64: 2842-2855.\u003c/li\u003e\n\u003cli\u003ePiao S, Tan K, Nan H, Ciais P, Fang J, Wang T, et al. Impacts of climate and CO2 changes on the vegetation growth and carbon balance of Qinghai\u0026ndash;Tibetan grasslands over the past five decades. Global and Planetary Change 2012; 98: 73-80.\u003c/li\u003e\n\u003cli\u003ePlante A, Conant R, Freedman B. Soil organic matter dynamics, climate change effects. Global Environmental Change, edited by: Freedman, B 2014: 317-323.\u003c/li\u003e\n\u003cli\u003ePoggio L, De Sousa LM, Batjes NH, Heuvelink GB, Kempen B, Ribeiro E, et al. SoilGrids 2.0: producing soil information for the globe with quantified spatial uncertainty. Soil 2021; 7: 217-240.\u003c/li\u003e\n\u003cli\u003ePost WM, Kwon KC. Soil carbon sequestration and land‐use change: processes and potential. Global change biology 2000; 6: 317-327.\u003c/li\u003e\n\u003cli\u003ePrietzel J, Zimmermann L, Schubert A, Christophel D. Organic matter losses in German Alps forest soils since the 1970s most likely caused by warming. Nature Geoscience 2016; 9: 543-548.\u003c/li\u003e\n\u003cli\u003eQi L, Yang J. Microbial community composition regulates SOC decomposition response to forest conversion in a Chinese temperate forest. Ecological research 2017; 32: 163-172.\u003c/li\u003e\n\u003cli\u003eR Core Team R. R: A language and environment for statistical computing. 2013.\u003c/li\u003e\n\u003cli\u003eRamesh T, Bolan NS, Kirkham MB, Wijesekara H, Kanchikerimath M, Rao CS, et al. Soil organic carbon dynamics: Impact of land use changes and management practices: A review. Advances in agronomy 2019; 156: 1-107.\u003c/li\u003e\n\u003cli\u003eRose A, Shakurideen O, Kolawole M, Davies N, Ayejuyo O, Friday A, et al. An investigation of the role of vegetation as a determinant of the vertical distribution of soil organic carbon. Journal of Emerging Trends in Engineering and Applied Sciences 2017; 8: 43-53.\u003c/li\u003e\n\u003cli\u003eRosseel Y. lavaan: An R package for structural equation modeling. Journal of statistical software 2012; 48: 1-36.\u003c/li\u003e\n\u003cli\u003eSaporetti-Junior AW, Schaefer CEGR, de Souza AL, Soares MP, Ara\u0026uacute;jo DSD, Meira-Neto JAA. Influence of soil physical properties on plants of the Mussununga ecosystem, Brazil. Folia Geobotanica 2012; 47: 29-39.\u003c/li\u003e\n\u003cli\u003eSchloss A, Kicklighter D, Kaduk J, Wittenberg U, Intercomparison TPOTPNM. Comparing global models of terrestrial net primary productivity (NPP): comparison of NPP to climate and the Normalized Difference Vegetation Index (NDVI). Global Change Biology 1999; 5: 25-34.\u003c/li\u003e\n\u003cli\u003eSchmidt MW, Torn MS, Abiven S, Dittmar T, Guggenberger G, Janssens IA, et al. Persistence of soil organic matter as an ecosystem property. Nature 2011; 478: 49-56.\u003c/li\u003e\n\u003cli\u003eSehler R, Li J, Reager J, Ye H. Investigating relationship between soil moisture and precipitation globally using remote sensing observations. Journal of Contemporary Water Research \u0026amp; Education 2019; 168: 106-118.\u003c/li\u003e\n\u003cli\u003eStockmann U, Padarian J, McBratney A, Minasny B, de Brogniez D, Montanarella L, et al. Global soil organic carbon assessment. Global Food Security 2015; 6: 9-16.\u003c/li\u003e\n\u003cli\u003eTeron G, Bordoloi R, Paul A, Singha LB, Tripathi OP. Effect of altitude on soil physico-chemical properties and microbial biomass carbon in the Eaglenest Wildlife Sanctuary of Arunachal Pradesh. Geology, Ecology, and Landscapes 2024: 1-19.\u003c/li\u003e\n\u003cli\u003eThornthwaite CW. An approach toward a rational classification of climate. Geographical review 1948; 38: 55-94.\u003c/li\u003e\n\u003cli\u003eTian H, Lu C, Yang J, Banger K, Huntzinger DN, Schwalm CR, et al. Global patterns and controls of soil organic carbon dynamics as simulated by multiple terrestrial biosphere models: Current status and future directions. Global Biogeochemical Cycles 2015; 29: 775-792.\u003c/li\u003e\n\u003cli\u003eTrumbore SE, Davidson EA, Barbosa de Camargo P, Nepstad DC, Martinelli LA. Belowground cycling of carbon in forests and pastures of Eastern Amazonia. Global biogeochemical cycles 1995; 9: 515-528.\u003c/li\u003e\n\u003cli\u003eWang X, Hu Z, Zhang Z, Tang J, Niu B. Altitude-Shifted Climate Variables Dominate the Drought Effects on Alpine Grasslands over the Qinghai\u0026ndash;Tibetan Plateau. Sustainability 2024; 16: 6697.\u003c/li\u003e\n\u003cli\u003eWang Y, Huang L, Jia X, Zhu Y. Distribution characteristics and controls of soil organic carbon at different spatial scales in China\u0026prime; s Loess Plateau. Journal of environmental management 2021; 293: 112943.\u003c/li\u003e\n\u003cli\u003eWang Y, Peng D, Shen M, Xu X, Yang X, Huang W, et al. Contrasting effects of temperature and precipitation on vegetation greenness along elevation gradients of the Tibetan Plateau. Remote Sensing 2020; 12: 2751.\u003c/li\u003e\n\u003cli\u003eWang Z, Huang L, Shao Ma. Spatial variations and influencing factors of soil organic carbon under different land use types in the alpine region of Qinghai-Tibet Plateau. Catena 2023; 220.\u003c/li\u003e\n\u003cli\u003eWright AF, Bailey JS. Organic carbon, total carbon, and total nitrogen determinations in soils of variable calcium carbonate contents using a Leco CN-2000 dry combustion analyzer. Communications in Soil Science and Plant Analysis 2001; 32: 3243-3258.\u003c/li\u003e\n\u003cli\u003eXu L, Yu G, He N. Increased soil organic carbon storage in Chinese terrestrial ecosystems from the 1980s to the 2010s. Journal of Geographical Sciences 2019; 29: 49-66.\u003c/li\u003e\n\u003cli\u003eXu Z, Wan C, Xiong P, Tang Z, Hu R, Cao G, et al. Initial responses of soil CO 2 efflux and C, N pools to experimental warming in two contrasting forest ecosystems, Eastern Tibetan Plateau, China. Plant and Soil 2010; 336: 183-195.\u003c/li\u003e\n\u003cli\u003eYang R, Huang L, Zhang X, Zhu C, Xu L. Mapping the distribution, trends, and drivers of soil organic carbon in China from 1982 to 2019. Geoderma 2023; 429: 116232.\u003c/li\u003e\n\u003cli\u003eYang Y, Fang J, Ma W, Smith P, Mohammat A, Wang S, et al. Soil carbon stock and its changes in northern China\u0026apos;s grasslands from 1980s to 2000s. Global Change Biology 2010; 16: 3036-3047.\u003c/li\u003e\n\u003cli\u003eYang Y, Fang J, Tang Y, Ji C, Zheng C, He J, et al. Storage, patterns and controls of soil organic carbon in the Tibetan grasslands. Global change biology 2008; 14: 1592-1599.\u003c/li\u003e\n\u003cli\u003eYao T, Thompson LG, Mosbrugger V, Zhang F, Ma Y, Luo T, et al. Third Pole Environment (TPE). Environmental Development 2012; 3: 52-64.\u003c/li\u003e\n\u003cli\u003eYao T, Xue Y, Chen D, Chen F, Thompson L, Cui P, et al. Recent third pole\u0026rsquo;s rapid warming accompanies cryospheric melt and water cycle intensification and interactions between monsoon and environment: Multidisciplinary approach with observations, modeling, and analysis. Bulletin of the American Meteorological society 2019; 100: 423-444.\u003c/li\u003e\n\u003cli\u003eYin Y, Wu S, Zhao D, Zheng D, Pan T. Modeled effects of climate change on actual evapotranspiration in different eco-geographical regions in the Tibetan Plateau. Journal of Geographical Sciences 2013; 23: 195-207.\u003c/li\u003e\n\u003cli\u003eZhang W, Zhang F, Qi J, Hou F. Modeling impacts of climate change and grazing effects on plant biomass and soil organic carbon in the Qinghai\u0026ndash;Tibetan grasslands. Biogeosciences 2017; 14: 5455-5470.\u003c/li\u003e\n\u003cli\u003eZhang W, Zhu W, Hu S. Soil resource availability impacts microbial response to organic carbon and inorganic nitrogen inputs. Journal of Environmental Sciences (China) 2005; 17: 705-710.\u003c/li\u003e\n\u003cli\u003eZhang X, Liu M, Zhao X, Li Y, Zhao W, Li A, et al. Topography and grazing effects on storage of soil organic carbon and nitrogen in the northern China grasslands. Ecological Indicators 2018; 93: 45-53.\u003c/li\u003e\n\u003cli\u003eZhang X, Yang Y, Piao S, Bao W, Wang S, Wang G, et al. Ecological change on the Tibetan Plateau. Chinese Science Bulletin 2015; 60: 3048-3056.\u003c/li\u003e\n\u003cli\u003eZhang Y, Hu Q, Zou F. Spatio-temporal changes of vegetation net primary productivity and its driving factors on the Qinghai-Tibetan Plateau from 2001 to 2017. Remote Sensing 2021; 13: 1566.\u003c/li\u003e\n\u003cli\u003eZheng C, Jia L, Hu G. Global land surface evapotranspiration monitoring by ETMonitor model driven by multi-source satellite earth observations. Journal of Hydrology 2022; 613: 128444.\u003c/li\u003e\n\u003cli\u003eZhong L, Ma Y, Xue Y, Piao S. Climate Change Trends and Impacts on Vegetation Greening Over the Tibetan Plateau. Journal of Geophysical Research: Atmospheres 2019; 124: 7540-7552.\u003c/li\u003e\n\u003cli\u003eZhu Z, Wang H, Harrison SP, Prentice IC, Qiao S, Tan S. Optimality principles explaining divergent responses of alpine vegetation to environmental change. Global Change Biology 2023; 29: 126-142.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"plant-and-soil","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"plso","sideBox":"Learn more about [Plant and Soil](https://www.springer.com/journal/11104)","snPcode":"11104","submissionUrl":"https://submission.nature.com/new-submission/11104/3","title":"Plant and Soil","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Soil organic carbon, Spatial distribution, Driving factors, Random Forest, Tibetan Plateau","lastPublishedDoi":"10.21203/rs.3.rs-6651952/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6651952/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cem\u003eBackground \u003c/em\u003eThe Tibetan Plateau (TP), a vast alpine region with substantial soil organic carbon (SOC) stocks, plays a crucial role in the regional carbon cycle and climate change mitigation. However, our understanding of the spatial distribution and underlying drivers of SOC across TP remains limited.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eMethods\u003c/em\u003e We quantified the horizontal (1 km resolution) and vertical (0–200 cm depth) patterns in SOC density (SOCD) across the TP using 1,561 soil samples and a fully trained and validated Random Forest (RF) model. Twenty-six environmental variables were evaluated to determine their influence on SOCD. Correlation analysis and Structural Equation Modeling (SEM) were employed to examine the associations and causal mechanisms between these factors and SOCD.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eResults \u003c/em\u003e\u0026nbsp;SOCD decreased from east to west and south to north, with mean values of \u0026nbsp;SOCD ranging from 14.96 kg C m⁻² in western steppes to 29.54 kg C m⁻² in eastern forests at 0-200 cm depth. Vertically, SOCD exhibited a nonlinear pattern, initially increasing and then decreasing with depth. SOCD was highest in wetlands, followed by forests, meadows, shrublands, and steppes.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eConclusions \u003c/em\u003eSOCD was influenced by multiple interacting environmental factors, primarily through their effects on plant productivity and soil respiration. Different ecosystems exhibit distinct regulatory mechanisms for SOCD. Forests are dominated by carbon loss through respiration, while grasslands and other ecosystems rely more on plant-derived carbon inputs. These findings enhance our understanding of distinct mechanisms regulating soil organic carbon across ecosystems, supporting improved modeling of soil carbon-climate feedbacks and informing ecosystem-specific carbon management strategies.\u003c/p\u003e","manuscriptTitle":"Distributions and Drivers of Soil Organic Carbon on the Tibetan Plateau: Divergent Controls Across Ecosystems","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-16 11:33:26","doi":"10.21203/rs.3.rs-6651952/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Major revisions","date":"2025-06-10T11:24:34+00:00","index":"","fulltext":""},{"type":"reviewerAgreed","content":"","date":"2025-05-15T22:10:26+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-05-14T06:40:28+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"Plant and Soil","date":"2025-05-14T05:58:19+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-05-13T09:23:02+00:00","index":"","fulltext":""},{"type":"submitted","content":"Plant and Soil","date":"2025-05-13T05:07:29+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"plant-and-soil","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"plso","sideBox":"Learn more about [Plant and Soil](https://www.springer.com/journal/11104)","snPcode":"11104","submissionUrl":"https://submission.nature.com/new-submission/11104/3","title":"Plant and Soil","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"b3b1ed6d-d2a5-4aa4-a4a3-b3e1685abf7b","owner":[],"postedDate":"May 16th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-10-06T16:07:48+00:00","versionOfRecord":{"articleIdentity":"rs-6651952","link":"https://doi.org/10.1007/s11104-025-07845-1","journal":{"identity":"plant-and-soil","isVorOnly":false,"title":"Plant and Soil"},"publishedOn":"2025-10-02 15:58:04","publishedOnDateReadable":"October 2nd, 2025"},"versionCreatedAt":"2025-05-16 11:33:26","video":"","vorDoi":"10.1007/s11104-025-07845-1","vorDoiUrl":"https://doi.org/10.1007/s11104-025-07845-1","workflowStages":[]},"version":"v1","identity":"rs-6651952","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6651952","identity":"rs-6651952","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}