Spatiotemporal dynamics and driving mechanisms of carbon sequestration advantage across ecological zones: a case study from 2000 to 2023 in the Zhangjiakou–Chengde region, China

Traditional carbon sink assessments based on absolute values often fail to capture relative sequestration advantages across heterogeneous landscapes. Using the Zhangjiakou–Chengde region as a case study, this research estimated Net Ecosystem Productivity (NEP) from 2000 to 2023 and integrated it with the National Ecosystem Assessment and Ecological Security Database of China to construct relative carbon sequestration advantages (CSA) for three ecological zones: Zone A (Bashang Plateau farming–pastoral ecotone), Zone B (Yan Mountains forest zone), and Zone C (upper Yongding River agro–pastoral mosaic). Drivers were analyzed using Pearson correlation, GeoDetector, and Partial Least Squares Structural Equation Modeling (PLS-SEM). Results show that Zone A had low and unstable sinks, with over 40% exhibiting CSA variability above 50 g C/m²·a, while Zone B maintained high CSA (multi-year mean > 100 g C/m²·a) but with strong interannual variability, suggesting high capacity yet weak stability. Zone C displayed clear improvement, with 35.3% of its area enhancing. Vegetation condition (NDVI, transpiration) and precipitation emerged as dominant positive drivers, whereas evapotranspiration, land use intensity, and sensible heat flux exerted localized negative effects, with climate and hydrology influencing CSA mainly through indirect vegetation–energy pathways. Projections indicate that Zone A faces mixed evolution (34.1% both improving and degrading), Zone B carries the highest degradation risk (34.4%), and Zone C shows the greatest improvement potential (35.3%). This study establishes a CSA framework that combines spatial pattern analysis with multi-method driver identification, offering practical insights for region-specific carbon sink governance in ecologically diverse regions.

Highlights

• Developed a zonal CSA framework based on NEP and ecological security zoning.

• Identified dominant and interactive CSA drivers via Pearson, GeoDetector, and PLS-SEM.

• Detected CSA evolution trends using Theil–Sen and Hurst exponent analysis.

• Revealed high spatial heterogeneity and varying stability of CSA across zones.

• Provided decision support for region-specific carbon sink governance strategies.

1 Introduction

Climate change remains one of the most formidable environmental challenges of our time, primarily driven by rising atmospheric carbon concentrations caused by greenhouse gas emissions (Shen et al., 2024; Zhang et al., 2025). These changes have triggered complex ecological responses and feedback processes on a global scale (Liu et al., 2023). Terrestrial ecosystems act as crucial carbon sinks in the global carbon cycle, sequestering atmospheric CO₂ through photosynthesis and storing carbon in biomass and soil organic matter over extended periods (Pan et al., 2011). These carbon sinks play a pivotal role not only in mitigating global warming and achieving carbon neutrality targets but also in evaluating ecosystem service value and shaping climate policy frameworks (Grassi et al., 2017). With the implementation of the Paris Agreement and China’s dual-carbon strategy—aimed at carbon peaking and neutrality—the accurate assessment of regional ecosystem carbon sequestration capacity and its spatial—temporal variability has emerged as a key area of research in both ecological and environmental management disciplines (Yang et al., 2024; Yang et al., 2022).

Net primary productivity (NPP) and net ecosystem productivity (NEP) are the most widely used indicators for carbon sink assessment. NPP quantifies the net carbon gain by vegetation after subtracting autotrophic respiration from gross primary production and serves as a proxy for ecosystem productivity, vegetation recovery, and climate responsiveness (Heinsch et al., 2006; Running et al., 2004; Zhao and Running, 2010). NEP further subtracts soil microbial heterotrophic respiration (Rh) from NPP, making it a key indicator of net carbon exchange between terrestrial ecosystems and the atmosphere. NEP is widely used in carbon budget assessments, flux modeling, and the evaluation of ecological restoration outcomes (He B. et al., 2022; Lim et al., 2024; Meetei et al., 2025). With the increasing availability of remote sensing products and modeling techniques, regional-scale assessments of NPP and NEP have rapidly advanced (Turner et al., 2006). However, these absolute-value metrics have limitations in comparing carbon sequestration capacity across ecological units, particularly in regions characterized by complex ecosystem types and strong spatial heterogeneity (Luyssaert et al., 2007; Zheng et al., 2024). Relative advantage assessments based on ecological subregions offer a more effective approach to identifying spatial carbon sink dominance and supporting differentiated ecosystem regulation and targeted resource allocation. This concept has been widely applied in studies of spatial dominance in ecosystem services (Eigenbrod et al., 2010), land-type carbon sink comparison under urban expansion (Seto et al., 2012), and performance evaluation of ecological restoration efforts in degraded areas (Zheng et al., 2016).

Despite ongoing progress in regional carbon sink research, single-time-series or average-based metrics fail to capture intra-regional differences in relative carbon sequestration capacity and its dynamic evolution (Zeng et al., 2023; Zhao et al., 2020). Moreover, many studies still rely on basic linear correlation methods to identify driving forces, limiting their ability to reveal spatial heterogeneity and multi-path causal relationships, thereby reducing their explanatory power for complex ecological processes (Wang et al., 2021; Xu et al., 2021). Against this backdrop, the Zhangjiakou–Chengde region, located at the transitional ecotone between the Inner Mongolian Plateau and the North China Plain, exhibits diverse topography, pronounced climatic gradients, and significant land-use conflicts. As a key component of the ecological security barrier for the Beijing–Tianjin–Hebei region (Li et al., 2022), the area’s ecosystems are highly vulnerable yet ecologically diverse, underscoring the urgent need for zonal carbon sink assessments that identify evolving carbon advantages and potential degradation risks to support national ecological governance and carbon goals.

This study focuses on the Zhangjiakou–Chengde (ZC) region and its three representative ecological zones to evaluate their relative carbon sequestration advantage (CSA). Based on long-term NEP time-series data from 2000 to 2023, we characterize the spatial patterns and temporal trajectories of CSA. Specifically, this research aims to: (1) estimate NEP variations over time and construct CSA indicators to assess the relative carbon sink capacity and stability of different ecological zones; (2) identify both direct and indirect drivers of CSA using Pearson correlation analysis, GeoDetector, and partial least squares structural equation modeling (PLS-SEM); (3) apply Theil–Sen trend estimation, Mann–Kendall significance testing, and Hurst exponent analysis to determine the trend direction and persistence of CSA; and (4) classify future CSA evolution types to support fine-scale management and degradation prevention in ecological subregions. This study advances regional carbon sequestration assessment in three key aspects. First, it introduces a relative carbon sequestration advantage metric that shifts the analytical focus from absolute carbon fluxes to inter-zonal competitiveness and vulnerability. Second, it integrates spatial pattern analysis with a multi-method driver identification framework, combining correlation analysis, spatial heterogeneity detection, and structural equation modeling to reveal both direct and indirect regulatory mechanisms. Third, it translates these diagnostic insights into zone-specific governance pathways, thereby bridging theoretical assessment with practical, adaptive carbon management in ecologically heterogeneous regions.

2 Materials and methods

2.1 Study area

The ZC region (39°18′–42°37′N, 113°50′–119°15′E) is located in northern Hebei Province, adjacent to Beijing, with a total area of approximately 76,000 km². The region is characterized by diverse landforms, including plateaus, mountains, and hills, and is governed by a temperate continental monsoon climate. According to the National Ecosystem Assessment and Ecological Security Database of China (http://www.ecosystem.csdb.cn/index.jsp), and based on topographic and vegetation characteristics, the ZC region is divided into three ecological zones (Figure 1). Zone A, the Bashang Plateau agro-pastoral ecotone, features a south-to-north decreasing elevation gradient. It has an annual mean temperature ranging from −0.3 °C to 3.5 °C and receives 340–450 mm of precipitation annually. The region suffers from water scarcity, and overgrazing has led to significant grassland degradation and soil erosion. Zone B, the Northern Hebei–Yanshan Mountain deciduous broadleaf forest zone, is dominated by low mountains and hills. It receives 600–700 mm of precipitation annually, has abundant water resources, and supports a diverse array of vegetation. This area serves as a critical carbon sink in the region.

Figure 1

Figure 1. Study area.

2.2 Data sources and processing

The Net Primary Productivity (NPP) data used for NEP estimation were obtained from the MODIS satellite product (MOD17A3HGF) provided by NASA through the LAADS DAAC platform (https://ladsweb.modaps.eosdis.nasa.gov), with a spatial resolution of 1 km. To construct the framework of driving factors, this study comprehensively considered both natural conditions and anthropogenic influences on NEP, categorizing the variables into four groups: Climate and Hydrology, Vegetation and Moisture, Human Activities, and Energy Processes, as detailed in Table 1. All data cover the period from 2000 to 2023. To ensure spatial consistency among datasets with different original resolutions, all driving factors were resampled to a unified grid of 1 km × 1 km prior to analysis. Continuous variables were resampled using spatial averaging, while categorical datasets were processed using the dominant class method. This harmonization procedure minimizes scale effects and ensures that variations in CSA and its driving factors are not artifacts of spatial resolution differences. For each cell, the mean value of the raster data was extracted to reduce the influence of outliers and ensure consistency in spatial resolution, thereby facilitating robust and comparable analysis across the study area.

Table 1

Table 1. Data sources and variable descriptions.

2.3 Method

The methodological framework of this study is illustrated in Figure 2 (1) NEP from 2000 to 2023 was estimated using MODIS-derived NPP data combined with a heterotrophic soil respiration model (Rh). Based on the national ecosystem assessment zoning system, the ZC region was categorized into three ecological zones: A, B, and C. (2) The CSA index was developed for each zone using NEP as the core metric, and Z-score standardization was applied to evaluate the relative intensity and stability of CSA across zones. (3) A multi-method approach involving Pearson correlation, GeoDetector analysis, and PLS-SEM was used to identify the key drivers of CSA, considering temporal dynamics, spatial differentiation, and underlying causal mechanisms. (4) The Theil–Sen trend estimation and Hurst exponent were jointly applied to forecast the direction and persistence of CSA changes in each ecological zone.

Figure 2

Figure 2. Technical flow chart.

2.3.1 Estimation of NEP

NEP reflects the net carbon flux between terrestrial ecosystems and the atmosphere, effectively indicating the rate of change in ecosystem carbon storage (Zhao et al., 2019). A substantial portion of NPP contributes to increasing the pool of organic carbon in ecosystems, while the remainder, entering the soil as litter, is either decomposed by microbial activity and released back into the atmosphere or stored as soil organic matter. Accordingly, NEP is defined as the difference between NPP and heterotrophic respiration (Rh) (Raich et al., 1991), as shown in Equation 1:

$NEP\left(x,t\right)=NPP\left(x,t\right)-R_h\left(x,t\right)(1)$

Where $NEP\left(x,t\right)$ denotes the NEP at grid cell x in year t (g C/m² a¹); $NPP\left(x,t\right)$ means the NPP at grid cell x in year t (g C/m² a¹); and $R_h\left(x,t\right)$ means the R_h at grid cell x in year t (g C/m² a¹).

Rh was estimated using a soil microbial respiration model (Pei et al., 2009), which has been successfully applied in various regions and shown reliable performance (Qi et al., 2024; Zou et al., 2022). The model is defined in Equation 2:

$R_h\left(x,t\right)=0.22\times (\exp \left(0.0912T\left(x,t\right)+\ln \left(0.3145P\left(x,t\right)+1\right)\right)\times 30\times 46.5\%(2)$

where $T\left(x,t\right)$ is the mean annual temperature (°C) at grid cell x in year t, and $P\left(x,t\right)$ is the annual precipitation (mm) at grid cell x in year.

2.3.2 Calculation of carbon sequestration advantage (CSA)

To assess the relative carbon sequestration potential across different ecological zones, this study introduces the concept of CSA, which quantifies the extent to which the NEP of a given zone outperforms that of other zones. The core idea is to compare the NEP value of each grid cell within a target ecological zone against the average NEP of all other zones. CSA is defined as a relative, inter-zonal comparative indicator that quantifies the extent to which the carbon sequestration capacity of a given ecological zone exceeds or falls below that of other zones within the same region. Unlike absolute carbon flux metrics such as NEP or NPP, CSA does not represent ecosystem carbon balance per se, but rather captures the spatial competitiveness and comparative advantage of carbon sequestration under heterogeneous ecological conditions. This approach, adapted from previous work on identifying spatial advantages in ecosystem services (Peng et al., 2023), has proven effective in informing spatial management strategies through interregional comparisons. The CSA is calculated using Equation 3:

$CS{A}_i\left(x,t\right)=NE{P}_i\left(x,t\right)-{\overline{NEP}}_{-i}\left(t\right)(3)$

where $CS{A}_i\left(x,t\right)$ represents CSA of grid cell x in ecological zone i in year t (g C/m² a¹), $NE{P}_i\left(x,t\right)$ represents NEP of grid cell x in ecological zone i in year t (g C/m² a¹), and ${\overline{NEP}}_{-i}\left(t\right)$ denotes the mean NEP across all ecological zones except zone i in year t (g C/m² a¹).

2.3.3 Spatial pattern matching analysis of CSA

The spatial variability of carbon sequestration functions within each ecological zone was assessed by calculating the standard deviation of CSA values. A larger standard deviation indicates greater fluctuation in carbon sequestration performance, while a smaller value suggests more stability. To examine the spatial coordination between CSA magnitude and its variability, both CSA and its standard deviation were normalized using Z-score transformation and plotted as a scatterplot, with standardized CSA on the x-axis and standardized variability on the y-axis. The position of each point in the scatterplot quadrants reflects the spatial matching characteristics of CSA. The Z-score transformations are defined by Equations 4, 5:

$Z_{CS{A}_i}\left(x\right)=\frac{CS{A}_i\left(x\right)-{\overline{CSA}}_i}{S_{CS{A}_i}}(4)$

$Z_{S{D\left(CSA\right)}_i}\left(x\right)=\frac{S{D\left(CSA\right)}_i\left(x\right)-{\overline{SD\left(CSA\right)}}_i}{S_{S{D\left(CSA\right)}_i}}(5)$

where $CS{A}_i\left(x\right)$ and $S{D\left(CSA\right)}_i\left(x\right)$ denote the multi-year average CSA and its standard deviation for grid cell x in ecological zone i (g C/m² a¹). ${\overline{CSA}}_i$ and ${\overline{SD\left(CSA\right)}}_i$ represent the mean CSA and mean standard deviation across all grid cells within zone i (g C/m² a¹), while $S_{CS{A}_i}$ and $S_{S{D\left(CSA\right)}_i}$ are the corresponding standard deviations (g C/m² a¹). $Z_{CS{A}_i}\left(x\right)$ and $Z_{S{D\left(CSA\right)}_i}\left(x\right)$ thus represent the standardized scores of the CSA value and its variability for a given grid cell x in ecological zone i. Specifically, Quadrant I ($Z_{CS{A}_i}\left(x\right)$ >0, $Z_{S{D\left(CSA\right)}_i}\left(x\right)$ >0) indicates high CSA with high variability; Quadrant II ($Z_{CS{A}_i}\left(x\right)$ <0, $Z_{S{D\left(CSA\right)}_i}\left(x\right)$ >0) indicates low CSA with high variability; Quadrant III ($Z_{CS{A}_i}\left(x\right)$ <0, $Z_{S{D\left(CSA\right)}_i}\left(x\right)$ <0) represents low CSA with low variability; while Quadrant IV ($Z_{CS{A}_i}\left(x\right)$ >0, $Z_{S{D\left(CSA\right)}_i}\left(x\right)$ <0) represents high CSA with low variability.

2.3.4 Analysis of driving factors

2.3.4.1 Pearson correlation

This study employed the Pearson correlation coefficient to evaluate the temporal response strength of CSA to individual driving factors, with statistical significance assessed using the t-test (Li et al., 2023). The formulas are given in Equations 6, 7:

$R=\frac{{\sum }_{t=1}^n\left(CSA\left(x,t\right)-\overline{CSA\left(x\right)}\right)\left(X\left(x,t\right)-\overline{X\left(x\right)}\right)}{\sqrt{{\sum }_{t=1}^n{\left(CSA\left(x,t\right)-\overline{CSA\left(x\right)}\right)}^2}\sqrt{{\sum }_{t=1}^n{\left(X\left(x,t\right)-\overline{X\left(x\right)}\right)}^2}}(6)$

$T=\frac{R\sqrt{n-2}}{\sqrt{n-R^2}}(7)$

where R represents the Pearson correlation coefficient; t represents the time sequence; n = 24 is the number of years in the time series; $CSA\left(x,t\right)$ and $X\left(x,t\right)$ are the CSA and the driving factor values for grid cell x in year t, respectively; $\overline{CSA\left(x\right)}$ and $\overline{X\left(x\right)}$ are the mean values over time for grid cell x; and T is the t-statistic used to evaluate the significance of the correlation.

2.3.4.2 GeoDetector analysis

The GeoDetector model was employed to quantify the spatial explanatory power of each driving factor on CSA across ecological zones. This method is effective in detecting spatial heterogeneity in geophysical phenomena and identifying the factors that account for it (Chen and Bi, 2022). In this study, CSA was treated as the dependent variable, while twelve potential driving variables were treated as independent factors. Each continuous independent variable was discretized into five categories using the natural breaks (Jenks) classification method to meet the model’s requirements. The explanatory power of each factor was then assessed using the factor detector module in GeoDetector, which calculates the q-statistic. A higher q-value indicates a stronger explanatory capacity of the variable for the spatial heterogeneity of CSA, while a lower q-value implies a weaker influence (Wang et al., 2022).

2.3.4.3 Partial least squares structural equation modeling

Partial Least Squares Structural Equation Modeling (PLS-SEM) is a multivariate statistical approach designed to examine the relationships between latent variables (Hair et al., 2017). It comprises both a measurement model and a structural model, and is particularly suitable for analyzing non-normally distributed data and constructs measured through both formative and reflective indicators. In this study, PLS-SEM was employed to investigate the direct and indirect effects of multiple latent drivers on CSA across the ecological zones of the ZC region (Gu et al., 2024). Based on the regional characteristics of the ZC region, a conceptual model was constructed with the following assumptions: (1) Climate and hydrology, Vegetation and moisture, Human activities, and Energy processes exert direct effects on CSA; (2) Climate and hydrology, Human activities, and Energy processes indirectly influence CSA by affecting Vegetation and moisture; and (3) Climate and hydrology also indirectly affect CSA through their influence on Human activities and Energy processes.

2.3.5 CSA trend analysis

This study employed a combined approach using the Theil–Sen median trend analysis and the Mann–Kendall significance test to assess the temporal dynamics of CSA across ecological zones in the ZC region from 2000 to 2023 (Shao et al., 2024). The rate of change was estimated using the Theil–Sen slope, calculated as the median of all pairwise rates of change in the CSA time series, as expressed in Equation 8:

$Sen=Median\left(\frac{CS{A}_j-CS{A}_i}{j-i}\right)(8)$

Where Sen denotes the trend slope, $CS{A}_i$ and $CS{A}_j$ represent the CSA values for years i and j, respectively, and i, j are elements of the time series with 2000 ≤ i < j ≤ 2023.

To evaluate the statistical significance of the detected trend, the Mann–Kendall test was applied. The test statistic Z_s was computed using Equations 9–11:

$Z_s=\left\{\begin{array}{l}\frac{S-1}{\sqrt{Var\left(S\right)}},S>0\\ 0,S=0\\ \frac{S+1}{\sqrt{Var\left(S\right)}},S<0\end{array}\right.(9)$

$S={\displaystyle \sum _{i=1}^{n-1}}{\displaystyle \sum _{j=i+1}^n}sgn\left(CS{A}_j-CS{A}_i\right)(10)$

$sgn\left(CS{A}_j-CS{A}_i\right)=\left\{\begin{array}{c}+1,CS{A}_j-CS{A}_i>0\\ 0,CS{A}_j-CS{A}_i=0\\ -1,CS{A}_j-CS{A}_i<0\end{array}\right.(11)$

Where $Z_s$ is the standardized test statistic, $S$ is the rank-based test statistic, $Var\left(S\right)$ denotes the variance of S, and n = 24 is the number of years in the time series. A significance level of α = 0.05 was adopted. The trend is considered statistically significant at 90%, 95%, and 99% confidence levels when $\left|Z_s\right|$ exceeds 1.65, 1.96, and 2.58, respectively. In this study, a threshold of $\left|Z_s\right|$ ≥ 1.96 was selected to ensure conservative significance testing.

2.3.6 Hurst exponent analysis

The Hurst exponent is widely used to predict the long-term behavior of time series data (Chen et al., 2022; Noorisameleh et al., 2021). In this study, the rescaled range (R/S) method was employed to calculate the Hurst exponent (H) of CSA in each ecological zone of the ZC region, thereby characterizing the persistence of CSA trends. The value of H ranges from 0 to 1: values near 0 indicate strong anti-persistence, meaning that future trends are likely to oppose past ones, while values approaching 1 suggest strong persistence, indicating that future changes are expected to continue in the same direction. The Hurst exponent is categorized into four classes: strong anti-persistence (0 ≤ H < 0.35), weak anti-persistence (0.35 ≤ H ≤ 0.5), weak persistence (0.5 < H ≤ 0.65), and strong persistence (0.65 < H ≤ 1).

3 Results

3.1 Spatiotemporal variations in NEP

As shown in Figure 3, the NEP in the study area exhibited a generally increasing trend with fluctuations from 2000 to 2023. In 2000, NEP levels were relatively low across most of the region, with the majority of areas below 150 g C/m²·a and some exhibiting negative values, indicating a net carbon source. Over time, NEP values progressively increased, reflecting an enhanced carbon sequestration function. By 2023, most areas exceeded 225 g C/m²·a, with some surpassing 375 g C/m²·a, indicating a substantial spatial improvement in carbon sink capacity. Figure 4a shows that areas with the most significant NEP increases were primarily located in the southwestern and southeastern hilly zones, with annual growth rates exceeding 1.5 g C/m²·a and even surpassing 3.0 g C/m²·a in some locations. In contrast, parts of the central and northern regions experienced a decline in NEP, with annual decreases reaching −1.5 g C/m²·a. As shown in Figure 4b, the mean NEP increased from approximately 120 g C/m²·a in 2000 to around 200 g C/m²·a in 2023, displaying a phased upward trend. The close alignment of the median and mean NEP values suggests a relatively symmetric distribution throughout the period.

Figure 3

Figure 3. Spatial distribution of NEP in the ZC region from 2000 to 2023.

Figure 4

Figure 4. Spatiotemporal variation of NEP in the ZC region from 2000 to 2023. (a) Spatial distribution of NEP change rates; (b) Temporal trends of NEP.

3.2 Spatiotemporal dynamics of CSA in different ecological zones

3.2.1 Spatial patterns and temporal trends of CSA

As shown in Figure 5a, CSA exhibited pronounced spatial heterogeneity across the three ecological zones. Zone A displayed predominantly negative CSA values, with some areas falling below −120 g C/m²·a, indicating a relatively weak carbon sequestration capacity. In contrast, Zone B showed consistently positive CSA values, with parts exceeding 120 g C/m²·a, suggesting a significant advantage in carbon sink function across the ZC region. Zone C presented intermediate characteristics, with stronger spatial variability in CSA values. Figure 5b illustrates the interannual trends of CSA from 2000 to 2023. Zone B exhibited a notable declining trend (r = −1.73, R² = 0.43), suggesting that although it remains the dominant carbon sink, its relative advantage has been narrowing. In comparison, Zone C showed a steady upward trajectory (r = 1.49, R² = 0.68), indicating a continuous enhancement in its carbon sequestration advantage. Zone A, while experiencing some fluctuations, showed no statistically significant trend (r = −0.46, R² = 0.14) and remained in a disadvantaged position in terms of carbon sink capacity.

Figure 5

Figure 5. Spatiotemporal patterns of carbon sequestration advantage (CSA) mean values and their standard deviations (Std. Dev.) across ecological zones in the ZC region. (a) Spatial distribution of multi-year mean CSA; (b) Temporal trends of CSA mean values by ecological zone; (c) Spatial distribution of CSA Std. Dev.; (d) Temporal trends of CSA Std. Dev. by ecological zone.

3.2.2 Stability characteristics and regional differences of CSA

The standard deviation of CSA was used to assess the variability of carbon sequestration advantage across ecological zones. As shown in Figure 5c, Zone B exhibited relatively high CSA variability, with some areas exceeding 55 g C/m²·a, indicating lower stability in its carbon sink capacity. In contrast, Zones A and C demonstrated generally lower variability, with most areas showing standard deviations below 25 g C/m²·a, suggesting more stable carbon sequestration performance. From a temporal perspective (Figure 5d), all three zones showed a declining trend in CSA variability over the 2000–2023 period. The decrease was particularly evident in Zone B (r = −2.43, R² = 0.48) and Zone C (r = −1.51, R² = 0.59), reflecting a trend toward greater stability in their carbon sink functions. Zone A experienced a more moderate decline (r = −1.28, R² = 0.45), and its CSA variability remained relatively high.

3.2.3 Matching patterns between CSA mean and stability

As shown in Figure 6, Zone A is predominantly characterized by grid cells in Quadrant III (low CSA, low Std. Dev.), mainly distributed across the central-northern plateau and the central-western agro-pastoral ecotone. This indicates that while the carbon sequestration capacity in this zone is relatively weak, it remains relatively stable over time. Zone B is dominated by grid cells in Quadrant IV (high CSA, low Std. Dev.), particularly concentrated in the central and southeastern hilly areas, reflecting both strong carbon sink capacity and high temporal stability. Notably, some areas along the northern margins and southeastern transition zones fall into Quadrant I (high CSA, high Std. Dev.), suggesting that although these regions exhibit strong sequestration potential, they are subject to considerable fluctuations possibly due to topographic complexity or anthropogenic disturbances. Zone C is primarily composed of Quadrant III grid cells, widely distributed in gently sloping hilly areas, indicating a pattern of low but stable carbon sink function. Of particular interest is the presence of a few Quadrant I grids concentrated along the southern mountainous margins and transitional zones of Zone C, implying that local ecological restoration efforts may have enhanced CSA levels, although long-term stability has yet to be established.

Figure 6

Figure 6. Spatial matching patterns between carbon sequestration advantage (CSA) and its standard deviation across ecological zones. (a) Spatial distribution of CSA and standard deviation quadrants; (b) Quadrant distribution of grid units in Zone A; (c) Quadrant distribution of grid units in Zone B; (d) Quadrant distribution of grid units in Zone C.

3.3 Response of CSA to driving factors

3.3.1 Temporal response strength of CSA to driving factors

Pearson correlation analysis was employed to quantify the temporal response strength of CSA to driving factors, and the significance of these relationships was assessed using t-tests. As shown in Figure 7, PRE, NDVI, VT, and SSM exhibit consistently strong positive correlations with CSA across all three ecological zones. Among them, NDVI stands out as one of the most sensitive indicators, with correlation coefficients exceeding 0.8 in many areas. In contrast, PET and SHF generally show moderate negative correlations with CSA. PET demonstrates extensive negatively correlated areas in the northern and central mountainous regions of Zone B, where correlation coefficients often fall below −0.6, suggesting that high evaporative demand may weaken the carbon sequestration advantage. SHF displays significant negative correlations mainly in the southern parts of Zones B and C, indicating that strong thermal fluxes may destabilize carbon sink functions. LUI and PD also show noticeable negative correlations with CSA in specific areas, particularly within agro-pastoral ecotones and urban fringe regions. Significance testing (Figure 8) further confirms these findings, with PRE, NDVI, VT, SSM, LUI, PD, and SHF showing statistically significant zones (p < 0.05) in multiple ecological regions. Notably, PET and SHF exhibit concentrated significant negative correlation zones in Zone B. Although LUI and PD are less strongly correlated than vegetation-related variables, they still display certain spatial clusters of significance. Figure 9 highlights that NDVI, VT, and PRE exhibit the highest average correlation coefficients and the broadest significant spatial coverage across the three zones, identifying them as the core positive drivers of CSA. Conversely, PET and SHF emerge as the primary limiting factors.

Figure 7

Figure 7. Spatial distribution of Pearson correlation coefficients between carbon sequestration advantage (CSA) and driving factors across the Zhangjiakou–Chengde region. Panels (a–l) represent correlations between CSA and precipitation (PRE), potential evapotranspiration (PET), temperature (TMP), standardized precipitation–evapotranspiration index (SPEI), normalized difference vegetation index (NDVI), vegetation transpiration (VT), surface soil moisture (SSM), land use intensity (LUI), population density (PD), nighttime light intensity (NTL), sensible heat flux (SHF), and solar radiation (SR), respectively. Warmer colors indicate stronger positive correlations, while cooler colors indicate stronger negative correlations.

Figure 8

Figure 8. Spatial distribution of significance levels for the correlations between carbon sequestration advantage (CSA) and driving factors. Panels (a–l) correspond to precipitation (PRE), potential evapotranspiration (PET), temperature (TMP), standardized precipitation–evapotranspiration index (SPEI), normalized difference vegetation index (NDVI), vegetation transpiration (VT), surface soil moisture (SSM), land use intensity (LUI), population density (PD), nighttime light intensity (NTL), sensible heat flux (SHF), and solar radiation (SR), respectively. Different color intensities represent statistical significance levels (p < 0.01, p < 0.05, p < 0.1, and p ≥ 0.1).

Figure 9

Figure 9. Radar charts illustrating the response characteristics of carbon sequestration advantage (CSA) to driving factors across ecological zones. Panel (a) shows the mean Pearson correlation coefficients between CSA and driving factors, while panel (b) presents the proportion of areas with statistically significant correlations (p < 0.05). Different symbols and colors represent ecological Zones A, B, and C.

3.3.2 Spatial heterogeneity of CSA response to driving factors

Using the GeoDetector method, this study quantified the explanatory power of each driving factor on the spatial heterogeneity of CSA (Figure 10). In Zone A, CSA variability was predominantly explained by NDVI (q = 0.721) and VT (q = 0.595), indicating that vegetation growth strongly shaped the spatial distribution of CSA. PRE (q = 0.536) and TMP (q = 0.381) also showed considerable influence, while human activity factors (LUI, PD, and NTL) had limited explanatory power. In Zone B, TMP (q = 0.435) and PET (q = 0.399) emerged as the dominant explanatory variables, with SR (q = 0.313) further highlighting the significance of energy input. Notably, the explanatory power of NDVI and VT decreased markedly in this zone (q = 0.136 and 0.094, respectively), suggesting that vegetation-related drivers play a diminished role in shaping CSA spatial patterns in this mountainous region. Zone C exhibited a more complex, multi-factor-driven pattern. NDVI remained the strongest contributor (q = 0.735), followed by VT (q = 0.501) and SSM (q = 0.368), indicating a dominant role of vegetation and moisture conditions in controlling CSA variation. In addition, LUI (q = 0.356) and PET (q = 0.375) also showed relatively high explanatory power, suggesting that both natural and anthropogenic factors jointly influence the spatial heterogeneity of CSA in this transitional ecological zone.

Figure 10

Figure 10. Explanatory power of driving factors for the spatial heterogeneity of carbon sequestration advantage (CSA) in each ecological zone.

3.3.3 Pathway analysis of CSA driving mechanisms

Using PLS-PM, this study clarified the direct and indirect effects of different categories of driving factors on CSA across ecological zones. As shown in Figure 11, vegetation and moisture emerged as the dominant positive drivers of CSA in all ecological zones, with most effects transmitted through direct pathways. In Zone A, CSA was strongly enhanced by vegetation and moisture, whereas climate and hydrology exerted multiple negative effects through both direct and indirect routes. Additionally, energy processes and human activities had significant negative effects, suggesting that CSA in this region is highly sensitive to environmental stressors. In Zone B, although climate and hydrology continued to suppress CSA, energy processes began to show a modest positive effect. The contribution of vegetation and moisture was slightly diminished compared to Zone A. In Zone C, vegetation and moisture remained the strongest positive influence on CSA. While climate and hydrology showed a weaker direct effect, its indirect impact—mediated through vegetation and moisture and energy processes—constituted a substantial suppressive force.

Figure 11

Figure 11. Pathways of interaction between driving factors and carbon sequestration advantage (CSA) in each ecological zone. (a) Zone A; (b) Zone B; (c) Zone C.

Figure 12 further quantifies these relationships. In Zone A, vegetation and moisture contributed the highest total positive effect (0.871), while climate and hydrology had a strong total negative effect (−0.793), primarily driven by indirect pathways (−0.487). In Zone B, climate and hydrology remained the dominant negative driver (−0.648), although it exhibited a small positive indirect effect (0.101). The total effect of energy processes was nearly neutral (0.034), and human activities had a relatively minor influence. In Zone C, the total positive effect of vegetation and moisture (0.746) remained highest, whereas the total negative effect of climate and hydrology (−0.729) was still significant. Notably, more than 60% of this negative effect was mediated indirectly, highlighting the high sensitivity of CSA in this zone to indirect climatic and hydrological stress.

Figure 12

Figure 12. Magnitude of direct and indirect effects of driving factors on carbon sequestration advantage (CSA) in each ecological zone. (a) Zone A; (b) Zone B; (c) Zone C.

3.4 Projected trends of CSA in different ecological zones

Using the Mann-Kendall test to determine historical trends and the Hurst exponent to assess trend persistence, the future development trajectories of CSA across the three ecological zones were classified. As shown in Table 2, Zone C demonstrates the highest potential for improvement, with 34.46% of the area categorized as weak improvement and 0.83% as strong improvement, together accounting for more than 35% of the region. This indicates a clear enhancement trend in carbon sequestration advantage. In contrast, Zone B faces the greatest risk of degradation, with 34.42% of its area classified as weak degradation or strong degradation, the latter being the highest proportion among all zones at 8.07%. Zone A exhibits the most pronounced variability, with 34.93% of its area showing signs of improvement, while nearly 40% is projected to undergo degradation, suggesting considerable uncertainty in future CSA dynamics. Spatial patterns, illustrated in Figure 13, further clarify these trends. Zone A presents a highly fragmented and mixed spatial pattern, where the western grassland-cropland mosaic is interspersed with weak improvement and anti-strong improvement zones, and the eastern hilly region shows extensive and clustered degradation patches. In Zone B, both weak and strong degradation areas are concentrated in the northern and southeastern mountain belts, reflecting a declining carbon sequestration advantage—particularly at the forest margins of higher elevations, where degradation appears to be expanding. Nevertheless, some southern and central parts exhibit a coexistence of strong improvement and anti-strong improvement, highlighting internal spatial heterogeneity. Zone C is dominated by weak and strong improvement areas in the central hilly zone and southern low mountain basins, forming relatively contiguous distributions that signal a solid foundation for ecological recovery. However, the western boundary region presents alternating patches of anti-improvement and weak degradation, indicating localized instability in CSA trends.

Table 2

Table 2. Future development trends of carbon sequestration advantage (CSA) and area proportions in each ecological zone.

Figure 13

Figure 13. Spatial distribution of future CSA trends in each ecological zone. (a) MK trend test; (b) Hurst index; (c) Development types.

4 Discussion

4.1 CSA index and carbon sequestration patterns

Compared with traditional approaches that evaluate carbon sink capacity using absolute values of NPP or NEP (Goulden et al., 2011), the CSA index provides a more insightful measure by assessing the relative carbon sequestration strengths across ecological zones. While NEP and NPP describe ecosystem-level carbon exchange processes, CSA operates at a higher analytical level by diagnosing relative spatial dominance and vulnerability among ecological zones, thereby providing information that absolute carbon metrics alone cannot offer for differentiated regional carbon governance. This relative framework offers enhanced explanatory power in regions with pronounced spatial heterogeneity. CSA not only reflects the absolute capacity of a region to sequester carbon but also captures its relative advantage compared to other ecological zones, thereby overcoming the limitations of single-region carbon assessments in cross-regional analyses. The CSA analytical framework is particularly applicable to regions characterized by strong ecological heterogeneity, distinct functional zoning, and competing land-use pressures. Its effectiveness relies on the availability of long-term carbon flux data and ecologically meaningful zoning schemes that allow for robust inter-regional comparison. While the framework is transferable to other regions, CSA should be interpreted as a flexible, comparative diagnostic indicator rather than a fixed-threshold metric, and its application should be adapted to local ecological baselines, dominant disturbance regimes, and governance objectives. Within these conditions, the CSA framework provides a stable and operationally scalable tool for identifying relative carbon sequestration advantages and vulnerabilities across diverse ecological contexts. This method has been widely adopted to identify spatial dominance in ecosystem services (Burkhard et al., 2012), evaluate the relative carbon sequestration performance of various land-use types during urban expansion (Wu and Wang, 2023), assess trade-offs between farmland use intensity and carbon budget efficiency (Bai et al., 2022), and gauge the effectiveness of ecological restoration efforts on relative carbon outcomes (Liao et al., 2020). At the regional scale, differences in ecosystem types, resource endowments, and disturbance regimes make it essential to assess carbon sequestration advantages at the ecological zone level. On one hand, this facilitates the identification of high-advantage and vulnerable areas, supporting the design of targeted governance strategies (Bhutia et al., 2024); on the other hand, it aligns with the practical needs of implementing ecological engineering in a regionally differentiated manner, promoting effective alignment between ecological policies and underlying biogeographic conditions (Zhang L. et al., 2024).

The ecological zoning adopted in this study is based on the National Ecosystem Assessment and Ecological Security Database of China, a standard framework widely used in China’s ecological conservation and governance efforts. Therefore, conducting CSA evaluations by ecological zone is both scientifically justified and practically relevant. The agro-pastoral ecotone of the Bashang Plateau (Zone A) is largely characterized by negative CSA values, indicating weak and unstable carbon sink performance. This area shows a typical carbon source pattern under anthropogenic pressure, primarily resulting from grassland degradation and high evapotranspiration (Zhu et al., 2023). The northern Hebei and the Yanshan Mountains (Zone B) has maintained consistently high CSA values with a generally concentrated spatial pattern and strong overall carbon sink performance. However, some areas exhibit high variability, suggesting that although the region benefits from dense forest cover and favorable climate conditions, local terrain variation or anthropogenic disturbances introduce uncertainty in carbon flux stability. Previous studies also confirm that, despite the region’s strong carbon sequestration capacity, certain areas exhibit pronounced spatiotemporal fluctuations (Guo et al., 2022). The upper basin of the Yongding River (Zone C) displays moderate CSA values with a steadily increasing trend over time, indicating a continued enhancement of carbon sequestration. The spatial aggregation of CSA in this region corresponds with reported ecological restoration efforts and outcomes (Fu et al., 2022; Yang et al., 2019). Overall, the three ecological zones differ significantly in natural endowments, vegetation types, and anthropogenic pressure, resulting in distinct spatiotemporal patterns of carbon sequestration.

4.2 Integrating multiple methods to identify driving mechanisms

This study employed a multi-method framework combining Pearson correlation analysis, GeoDetector, and PLS-SEM to comprehensively assess the drivers of CSA from three dimensions: temporal response, spatial heterogeneity, and latent causal pathways. Pearson correlation focused on quantifying the temporal sensitivity of CSA to individual variables (Azam and Khan, 2016), GeoDetector revealed the spatial explanatory power of each factor, making it suitable for uncovering local pattern formation mechanisms (Zhao et al., 2021), while PLS-SEM elucidated both direct and indirect effects among latent variables, capturing the hierarchical and synergistic nature of ecological drivers (Peeling et al., 2022). By integrating temporal response strength (Pearson), spatial differentiation (GeoDetector), and causal linkages (PLS-SEM), this synergistic approach offers a more complete understanding of the CSA driving mechanisms than any single statistical technique, enabling more targeted ecological regulation strategies.

Within the ZC region, the influence mechanisms of drivers varied significantly across ecological zones, closely tied to the respective ecological contexts. In Zone A, CSA was primarily driven by positive correlations with NDVI, VT, and PRE, while PET and land use intensity (LUI) showed strong negative associations. This suggests that in areas characterized by limited hydrothermal resources and intensified land use, carbon sinks are subject to compounded pressures. Similar patterns have been observed in grassland-dominated ecosystems, where the combined impact of human disturbance and climate stress has been shown to reduce carbon sequestration capacity (Chu et al., 2020; Zhou et al., 2019). In Zone B, CSA was most strongly associated with temperature (TMP) and solar radiation (SR), reflecting the negative impacts of rising temperatures and increased radiative energy on forest carbon balances. Under continued warming, the carbon uptake efficiency of temperate deciduous forests may decline, increasing their vulnerability (Cheng et al., 2023). Zone C exhibited a compound-driven pattern, with CSA strongly and positively influenced by NDVI and VT, while LUI and soil surface moisture (SSM) acted as coordinating factors that mediate the pathways between anthropogenic and hydrological variables. These interactions reflect a disturbance–regulation–recovery process. The dominance of indirect effects revealed by the PLS-SEM analysis indicates that climatic and energy-related stressors do not regulate carbon sequestration advantage solely through direct pathways. Instead, these drivers primarily exert their influence by modifying vegetation condition and moisture processes, which act as key mediators in vegetation–energy–climate interactions. This mediation highlights the buffering role of vegetation, whereby changes in vegetation structure and physiological activity can either amplify or dampen the impacts of climate and energy fluxes on carbon sequestration. Compared with traditional direct-effect models, which tend to attribute CSA variation directly to individual drivers, the mediated pathways identified here reveal hidden regulatory mechanisms and ecosystem resilience processes that are critical for understanding carbon dynamics under complex environmental conditions. Multiple studies have confirmed that post-disturbance functional reconstruction in ecosystems often depends on the coupled regulation of multiple drivers—a dynamic that is broadly representative and widely applicable in restoring ecosystem carbon fluxes (Maleki et al., 2019; Zhou et al., 2021).

4.3 Evolutionary trends and governance implications

The combined application of the Theil-Sen trend estimator and the Hurst exponent enables simultaneous assessment of both the direction and persistence of ecological change, allowing for the classification of ecosystem evolution patterns such as steady enhancement, phased fluctuation, or long-term degradation (Zhang H. et al., 2024). Compared to single-method trend analyses, this integrated approach offers enhanced sensitivity in detecting temporal persistence and is particularly well-suited for ecosystems under multiple stressors. In this study, it was applied to assess the evolving CSA trends across the ZC region. The policy implications derived from CSA patterns should be interpreted in light of the causal pathways identified by the PLS-SEM analysis. In regions where CSA is primarily constrained by indirect climatic stress mediated through vegetation degradation, governance strategies should prioritize vegetation restoration and disturbance reduction to stabilize vegetation–climate interactions and enhance ecosystem resilience. Conversely, in zones where energy processes and human activities exert stronger direct pressures, adaptive management should focus on optimizing land-use structure and regulating energy fluxes to mitigate their direct impacts on carbon sequestration. By explicitly aligning governance strategies with dominant regulatory pathways, policy interventions can move beyond descriptive spatial prioritization toward process-oriented, resilience-based ecosystem management.

Zone A, situated in the transitional belt between the plateau and the plain, suffers from limited hydrothermal resources and considerable land-use pressure. With a Hurst exponent near 0.5, the CSA evolution in this area exhibits high uncertainty. Ecological protection should take precedence here, focusing on grazing exclusion in degraded grasslands, livestock-grassland balance, and the implementation of ecological compensation and restoration programs such as grassland subsidies and grazing bans to enhance carbon sink stability through policy–ecology synergies (Li et al., 2019; Su et al., 2025) one B, characterized by mountainous forest ecosystems, currently maintains high CSA levels but is experiencing increased volatility in Hurst values, especially in forest edge zones where degradation is becoming more pronounced. Governance in this region should prioritize optimizing forest stand structures, enhancing climate resilience, and conserving biodiversity. These actions should be aligned with large-scale forest-focused programs such as the Natural Forest Protection Project and the Beijing–Tianjin Sandstorm Source Control Project to form an integrated response mechanism (Li et al., 2018; Zhang et al., 2020) Zone C, comprising hilly basins and transitional ecosystems, has shown clear signs of ecological recovery in recent years. It exhibits continuous CSA growth and a high Hurst index, indicating strong sustainability in carbon sequestration. Continued implementation of land consolidation, soil and water conservation, and restorative agricultural practices is recommended to further enhance ecosystem resilience.

Tailoring governance strategies to the ecological context of each zone not only improves the long-term stability and adaptability of carbon sink systems but also provides actionable pathways for achieving regional ecological security and climate mitigation goals. More importantly, this zone-based diagnostic approach to carbon dynamics offers transferable insights for other regions with ecological heterogeneity, advancing the global shift from generalized assessments to precision-targeted ecological regulation and improving the effectiveness of ecosystem management.

5 Conclusion

Based on NEP estimation and the National Ecosystem Assessment and Ecological Security Database of China, this study developed a relative carbon sequestration advantage (CSA) indicator for ecological zones and analyzed the spatiotemporal patterns, driving mechanisms, and future trajectories of carbon sinks in the ZC region from 2000 to 2023. The main conclusions are as follows:

1. CSA exhibited clear spatial heterogeneity across ecological zones. Zone A showed low and highly variable CSA, indicating elevated degradation risk; Zone B consistently maintained high CSA and represented the core regional carbon sink; Zone C demonstrated substantial potential for CSA enhancement under ecological restoration.

2. Vegetation- and moisture-related factors were the dominant positive drivers of CSA. Climate and hydrological variables mainly constrained CSA through indirect pathways, while human activities affected CSA primarily by altering vegetation conditions and energy-related processes.

3. Trend analysis based on the Theil–Sen estimator and the Hurst index revealed significant but heterogeneous CSA dynamics. Zone A showed high uncertainty and low stability, Zone B faced increasing degradation risks despite strong carbon capacity, and Zone C exhibited a stable and sustainable positive trajectory, highlighting the importance of differentiated and adaptive carbon sequestration management.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Author contributions

HH: Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing. Y-XK: Writing – original draft. YW: Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing. Z-CX: Writing – review and editing. WW: Conceptualization, Writing – original draft.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This work was supported by the National Natural Science Foundation of China (42474012), the Joint Open Fund of Water Cycle Field Station of the Heihe River Basin, CGS (WCSHR-2024-03), the Youth Fund Project for Hebei Minzu Normal University (QN2024002), the Hebei Key Laboratory of Mountain Geological Environment (HBKLMGE202403), the Higher Education Science and Technology Research Project for Hebei Province (grant numbers: BJK2023105), the Major Competitive Project of Hebei Minzu Normal University (STZD2021001), the Science and Technology + Joint Plan of Jiangxi Province (2023KDG01008).

Acknowledgements

The authors are grateful to the editor and reviewers for their careful and valuable suggestions.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Azam, M., and Khan, A. Q. (2016). Testing the environmental Kuznets curve hypothesis: a comparative empirical study for low, lower middle, upper middle and high income countries. Renew. Sustain. Energy Rev. 63, 556–567. doi:10.1016/j.rser.2016.05.052

CrossRef Full Text | Google Scholar

Bai, S. T., Yang, J. C., Zhang, Y. B., Yan, F. Q., Yu, L. X., and Zhang, S. W. (2022). Evaluating ecosystem services and trade-offs based on land-use simulation: a case study in the farming-pastoral ecotone of northern China. Land 11 (7), 1115. doi:10.3390/land11071115

CrossRef Full Text | Google Scholar

Bhutia, K. D. O., Mishra, M., Guria, R., Baraj, B., Naik, A. K., da Silva, R. M., et al. (2024). Evaluation of large-scale deforestation susceptibility mapping in the mountainous region of the himalayas: a case study of the khangchendzonga biosphere reserve, India. Remote Sens. Applications-Society Environ. 36, 101285. doi:10.1016/j.rsase.2024.101285

CrossRef Full Text | Google Scholar

Burkhard, B., Kroll, F., Nedkov, S., and Müller, F. (2012). Mapping ecosystem service supply, demand and budgets. Ecol. Indic. 21, 17–29. doi:10.1016/j.ecolind.2011.06.019

CrossRef Full Text | Google Scholar

Chen, C., and Bi, L. (2022). Study on spatio-temporal changes and driving factors of carbon emissions at the building operation stage- A case study of China. Build. Sci. 219, 109147. doi:10.1016/j.buildenv.2022.109147

CrossRef Full Text | Google Scholar

Chen, J., Shao, Z., Huang, X., Zhuang, Q., Dang, C., Cai, B., et al. (2022). Assessing the impact of drought-land cover change on global vegetation greenness and productivity. Sci. Total Environ. 852, 158499. doi:10.1016/j.scitotenv.2022.158499

PubMed Abstract | CrossRef Full Text | Google Scholar

Cheng, R., Zhang, J., Wang, X., and Zhang, Z. (2023). Growth suitability evaluation of Larix principis-rupprechtii mayr based on potential NPP under different climate scenarios. Sustainability 15 (1), 331. doi:10.3390/su15010331

CrossRef Full Text | Google Scholar

Chu, X., Zhan, J., Wang, C., Hameeda, S., and Wang, X. (2020). Households' willingness to accept improved ecosystem services and influencing factors: application of contingent valuation method in bashang Plateau, Hebei Province, China. J. Environ. Manage. 255, 109925. doi:10.1016/j.jenvman.2019.109925

PubMed Abstract | CrossRef Full Text | Google Scholar

Eigenbrod, F., Armsworth, P. R., Anderson, B. J., Heinemeyer, A., Gillings, S., Roy, D. B., et al. (2010). The impact of proxy-based methods on mapping the distribution of ecosystem services. J. Appl. Ecol. 47 (2), 377–385. doi:10.1111/j.1365-2664.2010.01777.x

CrossRef Full Text | Google Scholar

Fu, H., Zhao, S., and Liao, C. (2022). Spatial governance of beijing-tianjin-hebei urban agglomeration towards low-carbon transition. China Agric. Econ. Rev. 14 (4), 774–798. doi:10.1108/caer-04-2022-0069

CrossRef Full Text | Google Scholar

Goulden, M. L., McMillan, A. M. S., Winston, G. C., Rocha, A. V., Manies, K. L., Harden, J. W., et al. (2011). Patterns of NPP, GPP, respiration, and NEP during boreal forest succession. Glob. Change Biol. 17 (2), 855–871. doi:10.1111/j.1365-2486.2010.02274.x

CrossRef Full Text | Google Scholar

Grassi, G., House, J., Dentener, F., Federici, S., den Elzen, M., and Penman, J. (2017). The key role of forests in meeting climate targets requires science for credible mitigation. Nat. Clim. Change 7 (3), 220–226. doi:10.1038/nclimate3227

CrossRef Full Text | Google Scholar

Gu, Z., Chen, X., Ruan, W., Zheng, M., Gen, K., Li, X., et al. (2024). Quantifying the direct and indirect effects of terrain, climate and human activity on the spatial pattern of kNDVI-based vegetation growth: a case study from the minjiang river basin, southeast China. Ecol. Inf. 80, 102493. doi:10.1016/j.ecoinf.2024.102493

CrossRef Full Text | Google Scholar

Guo, W., Teng, Y., Yan, Y., Zhao, C., Zhang, W., and Ji, X. (2022). Simulation of land use and carbon storage evolution in multi-scenario: a case study in beijing-tianjin-hebei urban agglomeration, China. Sustainability 14 (20), 13436. doi:10.3390/su142013436

CrossRef Full Text | Google Scholar

Hair, J., Hollingsworth, C. L., Randolph, A. B., and Chong, A. Y. L. (2017). An updated and expanded assessment of PLS-SEM in information systems research. Industrial Manag. and Data Syst. 117 (3), 442–458. doi:10.1108/imds-04-2016-0130

CrossRef Full Text | Google Scholar

He, B., Chen, C., Lin, S., Yuan, W., Chen, H., Chen, D., et al. (2022a). Worldwide impacts of atmospheric vapor pressure deficit on the interannual variability of terrestrial carbon sinks. Natl. Sci. Rev. 9 (4), 1–8. doi:10.1093/nsr/nwab150

PubMed Abstract | CrossRef Full Text | Google Scholar

He, N., Zhou, Y., Wang, L., Li, Q., Zuo, Q., and Liu, J. (2022b). Spatiotemporal differentiation and the coupling analysis of ecosystem service value with land use change in Hubei Province, China. Ecol. Indic. 145, 109693. doi:10.1016/j.ecolind.2022.109693

CrossRef Full Text | Google Scholar

Heinsch, F. A., Zhao, M., Running, S. W., Kimball, J. S., Nemani, R. R., Davis, K. J., et al. (2006). Evaluation of remote sensing based terrestrial productivity from MODIS using regional tower eddy flux network observations. IEEE Trans. Geosci. Electron. 44 (7), 1908–1925. doi:10.1109/TGRS.2005.853936

CrossRef Full Text | Google Scholar

Li, H., Zhang, X., Zhang, Y., Yang, X., and Zhang, K. (2018). Risk assessment of forest landscape degradation using Bayesian network modeling in the miyun reservoir catchment (China) with emphasis on the Beijing-Tianjin sandstorm source control program. Land Degrad. Dev. 29 (11), 3876–3885. doi:10.1002/ldr.3133

CrossRef Full Text | Google Scholar

Li, Y., Fan, J., and Yu, H. (2019). Grazing exclusion, a choice between biomass growth and species diversity maintenance in BeijingTianjin sand source control Project. Sustainability 11 (7), 1941. doi:10.3390/su11071941

CrossRef Full Text | Google Scholar

Li, Q., Li, W., Wang, S., and Wang, J. (2022). Assessing heterogeneity of trade-offs/synergies and values among ecosystem services in Beijing-Tianjin-Hebei urban agglomeration. Ecol. Indic. 140, 109026. doi:10.1016/j.ecolind.2022.109026

CrossRef Full Text | Google Scholar

Li, Y., Ji, C., Wang, P., and Huang, L. (2023). Proactive intervention of green infrastructure on flood regulation and mitigation service based on landscape pattern. J. Clean. Prod. 419, 138152. doi:10.1016/j.jclepro.2023.138152

CrossRef Full Text | Google Scholar

Liao, Q., Wang, Z., and Huang, C. (2020). Green infrastructure offset the negative ecological effects of urbanization and storing water in the three gorges Reservoir area, China. Int. J. Environ. Res. Public Health. 17 (21), 8077. doi:10.3390/ijerph17218077

PubMed Abstract | CrossRef Full Text | Google Scholar

Lim, B. S., Joo, S. J., Seok, J. E., and Lee, C. S. (2024). An assessment of the carbon budget of the passively restored willow forests along the miho river, central South Korea. Climate 12 (11), 182. doi:10.3390/cli12110182

CrossRef Full Text | Google Scholar

Liu, Z., Deng, Z., Davis, S., and Ciais, P. (2023). Monitoring global carbon emissions in 2022. Nat. Rev. Earth Environ. 4 (4), 205–206. doi:10.1038/s43017-023-00406-z

PubMed Abstract | CrossRef Full Text | Google Scholar

Luyssaert, S., Inglima, I., Jung, M., Richardson, A. D., Reichstein, M., Papale, D., et al. (2007). CO₂ balance of boreal, temperate, and tropical forests derived from a global database. Glob. Change Biol. 13 (12), 2509–2537. doi:10.1111/j.1365-2486.2007.01439.x

CrossRef Full Text | Google Scholar

Maleki, S., Koupaei, S. S., Soffianian, A., Saatchi, S., Pourmanafi, S., and Rahdari, V. (2019). Human and climate effects on the hamoun wetlands. Weather Clim. Soc. 11 (3), 609–622. doi:10.1175/wcas-d-18-0070.1

CrossRef Full Text | Google Scholar

Meetei, K. B., Tsopoe, M., Chandra, G., Mukhopadhyay, D., and Giri, K. (2025). Ecosystem productivity and carbon dynamics in keibul lamjao national park, Manipur, India: a gray relational analysis perspective. Environ. Monit. Assess. 197 (2), 140. doi:10.1007/s10661-024-13608-9

PubMed Abstract | CrossRef Full Text | Google Scholar

Noorisameleh, Z., Gough, W. A., and Mirza, M. M. Q. (2021). Persistence and spatial-temporal variability of drought severity in Iran. Environ. Sci. Pollut. Res. 28 (35), 48808–48822. doi:10.1007/s11356-021-14100-4

PubMed Abstract | CrossRef Full Text | Google Scholar

Pan, Y., Birdsey, R. A., Fang, J., Houghton, R., Kauppi, P. E., Kurz, W. A., et al. (2011). A large and persistent carbon sink in the world's forests. Sci 333 (6045), 988–993. doi:10.1126/science.1201609

PubMed Abstract | CrossRef Full Text | Google Scholar

Peeling, J. A., Singh, A., and Judge, J. (2022). A structural equation modeling approach to disentangling regional-scale landscape dynamics in Ghana. Front. Env. Sci. 9, 729266. doi:10.3389/fenvs.2021.729266

CrossRef Full Text | Google Scholar

Pei, Z., Ouyang, H., Zhou, C., and Xu, X. (2009). Carbon balance in an alpine steppe in the Qinghai-Tibet Plateau. J. Integr. Plant Biol. 51 (5), 521–526. doi:10.1111/j.1744-7909.2009.00813.x

PubMed Abstract | CrossRef Full Text | Google Scholar

Peng, J., Tang, H., Su, C., Jiang, H., Dong, J., and Xu, D. (2023). Regarding reference state to identify priority areas for ecological restoration in a karst region. J. Environ. Manage. 348, 119214. doi:10.1016/j.jenvman.2023.119214

PubMed Abstract | CrossRef Full Text | Google Scholar

Qi, S., Zhang, H., and Zhang, M. (2024). Evolutionary characteristics of carbon sources/sinks in Chinese terrestrial ecosystems regarding to temporal effects and geographical partitioning. Ecol. Indic. 160, 111923. doi:10.1016/j.ecolind.2024.111923

CrossRef Full Text | Google Scholar

Raich, J. W., Rastetter, E. B., Melillo, J. M., Kicklighter, D. W., Steudler, P. A., Peterson, B. J., et al. (1991). Potential net primary productivity in South America: application of a global model. Ecol. Appl. 1 (4), 399–429. doi:10.2307/1941899

PubMed Abstract | CrossRef Full Text | Google Scholar

Running, S. W., Nemani, R. R., Heinsch, F. A., Zhao, M. S., Reeves, M., and Hashimoto, H. (2004). A continuous satellite-derived measure of global terrestrial primary production. Bioscience 54 (6), 547–560. doi:10.1641/0006-3568(2004)054[0547:acsmog]2.0.co;2

CrossRef Full Text | Google Scholar

Seto, K. C., Güneralp, B., and Hutyra, L. R. (2012). Global forecasts of urban expansion to 2030 and direct impacts on biodiversity and carbon pools. Proc. Natl. Acad. Sci. U. S. A. 109 (40), 16083–16088. doi:10.1073/pnas.1211658109

PubMed Abstract | CrossRef Full Text | Google Scholar

Shao, Y., Zhu, Q., Feng, Z., Sun, L., Yang, X., Li, X., et al. (2024). Temporal and spatial assessment of carbon flux dynamics: evaluating emissions and sequestration in the three northern protection forest project areas supported by google Earth engine. Remote Sens. 16 (5), 777. doi:10.3390/rs16050777

CrossRef Full Text | Google Scholar

Shen, X., Shen, M., Wu, C., Peñuelas, J., Ciais, P., Zhang, J., et al. (2024). Critical role of water conditions in the responses of autumn phenology of marsh wetlands to climate change on the Tibetan Plateau. Glob. Change Biol. 30 (1), e17097. doi:10.1111/gcb.17097

PubMed Abstract | CrossRef Full Text | Google Scholar

Su, J., Xu, F., Wang, Y., Zhao, Y., and Bai, Y. (2025). Collaborative restoration between vegetation and soil attributes in degraded grassland of different grassland types. Chin. Sci. Bulletin-Chinese 70 (11), 1519–1536. doi:10.1360/tb-2024-0442

CrossRef Full Text | Google Scholar

Turner, D. P., Ritts, W. D., Cohen, W. B., Gower, S. T., Running, S. W., Zhao, M. S., et al. (2006). Evaluation of MODIS NPP and GPP products across multiple biomes. Remote sens. Environ. 102 (3-4), 282–292. doi:10.1016/j.rse.2006.02.017

CrossRef Full Text | Google Scholar

Wang, H., Liu, L., Yin, L., Shen, J., and Li, S. (2021). Exploring the complex relationships and drivers of ecosystem services across different geomorphological types in the beijing-tianjin-hebei region, China (2000-2018). Ecol. Indic. 121, 107116. doi:10.1016/j.ecolind.2020.107116

CrossRef Full Text | Google Scholar

Wang, Z., Li, X., and Zhao, H. (2022). Identifying spatial influence of urban elements on road-deposited sediment and the associated phosphorus by coupling Geodetector and bayesian networks. J. Environ. Manage. 315, 115170. doi:10.1016/j.jenvman.2022.115170

PubMed Abstract | CrossRef Full Text | Google Scholar

Wu, F., and Wang, Z. (2023). Assessing the impact of urban land expansion on ecosystem carbon storage: a case study of the changzhutan metropolitan area, China. Ecol. Indic. 154, 110688. doi:10.1016/j.ecolind.2023.110688

CrossRef Full Text | Google Scholar

Xu, K., Chi, Y., Wang, J., Ge, R., and Wang, X. (2021). Analysis of the spatial characteristics and driving forces determining ecosystem quality of the beijing-tianjin-hebei region. Environ. Sci. Pollut. Res. 28 (10), 12555–12565. doi:10.1007/s11356-020-11146-8

PubMed Abstract | CrossRef Full Text | Google Scholar

Yang, Y., Zheng, H., Xu, W., Zhang, L., and Ouyang, Z. (2019). Temporal changes in multiple ecosystem services and their bundles responding to urbanization and ecological restoration in the beijing-tianjin-hebei metropolitan area. Sustainability 11 (7), 2079. doi:10.3390/su11072079

CrossRef Full Text | Google Scholar

Yang, Y., Shi, Y., Sun, W., Chang, J., Zhu, J., Chen, L., et al. (2022). Terrestrial carbon sinks in China and around the world and their contribution to carbon neutrality. Sci. China Life Sci. 65 (5), 861–895. doi:10.1007/s11427-021-2045-5

PubMed Abstract | CrossRef Full Text | Google Scholar

Yang, F., Yang, X., and Li, X. (2024). China's diverse energy transition pathways toward carbon neutrality by 2060. Sustain. Prod. Consum. 47, 236–250. doi:10.1016/j.spc.2024.04.007

CrossRef Full Text | Google Scholar

Zeng, J., Zhou, T., Wang, Q., Xu, Y., Lin, Q., Zhang, Y., et al. (2023). Spatial patterns of China's carbon sinks estimated from the fusion of remote sensing and field-observed net primary productivity and heterotrophic respiration. Ecol. Inf. 76, 102152. doi:10.1016/j.ecoinf.2023.102152

CrossRef Full Text | Google Scholar

Zhang, T., Lan, J., Yu, J., Liu, Z., and Yao, S. (2020). Assessment of forest restoration projects in different regions using multicriteria decision analysis methods. J. For. Res. 25 (1), 12–20. doi:10.1080/13416979.2019.1704360

CrossRef Full Text | Google Scholar

Zhang, H., He, Z., Zhang, L., Cong, R., and Wei, W. (2024a). Spatial-temporal changes and driving factor analysis of net ecosystem productivity in Heilongjiang province from 2010 to 2020. Land 13 (8), 1316. doi:10.3390/land13081316

CrossRef Full Text | Google Scholar

Zhang, L., Shi, L., and Yang, F. (2024b). Assessing ecosystem service dynamics in China's coastal shelterbelt: implications for ecosystem restoration. Environ. Impact Assess. Rev. 106, 107515. doi:10.1016/j.eiar.2024.107515

CrossRef Full Text | Google Scholar

Zhang, J., Zhao, M., Ma, R., Feng, M., Ding, C., Wu, L., et al. (2025). Different impacts of temperature and precipitation on vegetation net primary production of global marsh wetlands. Glob. Planet. Change 252, 104893. doi:10.1016/j.gloplacha.2025.104893

CrossRef Full Text | Google Scholar

Zhao, M., and Running, S. W. (2010). Drought-induced reduction in global terrestrial net primary production from 2000 through 2009. Sci 329 (5994), 940–943. doi:10.1126/science.1192666

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhao, J., Ma, J., and Zhu, Y. (2019). Evaluating impacts of climate change on net ecosystem productivity (NEP) of global different forest types based on an individual tree-based model FORCCHN and remote sensing. Glob. Planet. Change 182, 103010. doi:10.1016/j.gloplacha.2019.103010

CrossRef Full Text | Google Scholar

Zhao, W., Hu, Z., Guo, Q., Wu, G., Chen, R., and Li, S. (2020). Contributions of climatic factors to interannual variability of the vegetation index in northern China grasslands. J. Clim. 33 (1), 175–183. doi:10.1175/jcli-d-18-0587.1

CrossRef Full Text | Google Scholar

Zhao, X., Liu, J., and Bu, Y. (2021). Quantitative analysis of spatial heterogeneity and driving forces of the thermal environment in urban Built-up areas: a case study in Xi'an, China. Sustainability 13 (4), 1870. doi:10.3390/su13041870

CrossRef Full Text | Google Scholar

Zheng, H., Li, Y., Robinson, B., Liu, G., Ma, D., Wang, F., et al. (2016). Using ecosystem service trade-offs to inform water conservation policies and management practices. Front. Ecol. Environ. 14 (10), 527–532. doi:10.1002/fee.1432

CrossRef Full Text | Google Scholar

Zheng, H., Yang, X., Song, C., Zhang, W., Sun, W., and Wang, G. (2024). Distinct environmental controls on above- and below-ground net primary productivity in northern China's grasslands. Ecol. Indic. 167, 112717. doi:10.1016/j.ecolind.2024.112717

CrossRef Full Text | Google Scholar

Zhou, Q., Zhao, X., Wu, D., Tang, R., Du, X., Wang, H., et al. (2019). Impact of urbanization and climate on vegetation coverage in the beijing-tianjin-hebei region of China. Remote Sens. 11 (20), 2452. doi:10.3390/rs11202452

CrossRef Full Text | Google Scholar

Zhou, Y., Fu, D., Lu, C., Xu, X., and Tang, Q. (2021). Positive effects of ecological restoration policies on the vegetation dynamics in a typical ecologically vulnerable area of China. Ecol. Eng. 159, 106087. doi:10.1016/j.ecoleng.2020.106087

CrossRef Full Text | Google Scholar

Zhu, L., Shi, M., Fan, D., Tu, K., and Sun, W. (2023). Analysis of changes in vegetation carbon storage and net primary productivity as influenced by land-cover change in Inner Mongolia, China. Sustainability 15 (6), 4735. doi:10.3390/su15064735

CrossRef Full Text | Google Scholar

Zou, C., Li, H., Chen, D., Fan, J., Liu, Z., Xu, X., et al. (2022). Spatial-temporal changes of carbon source/sink in terrestrial vegetation ecosystem and response to meteorological factors in yangtze river Delta region (China). Sustainability 14 (16), 10051. doi:10.3390/su141610051

CrossRef Full Text | Google Scholar