Spatiotemporal evolution and driving mechanisms of multiple scales ecological security in Shanxi Province from the perspective of service, risk and health

Ecological Security (ES) refers to the capacity of an ecosystem to maintain its structural integrity and functional stability against internal and external stresses, thereby continuously providing essential ecosystem services for human survival and development. It is a critical component of the national security framework¹. Global ecosystems are currently confronting numerous threats, such as climate change, biodiversity loss, and land degradation. Rapid urbanization and intensive land use further exacerbate the vulnerability of regional ecosystems, posing serious ecological and environmental challenges². As a prerequisite for achieving regional sustainable development, supporting the national “dual carbon” strategy, and facilitating high-quality development in the Yellow River Basin, ecological security assessment has emerged as a central focus in global change research and territorial spatial planning³.

Ecological security research typically encompasses three dimensions: “benefits–state–risks”⁴. Ecosystem services quantify nature’s contributions to humanity from the perspective of “positive benefits”; ecosystem health focuses on diagnosing the integrity of the system’s internal “state”; while landscape ecological risks evaluate the potential impacts of external pressures from the angle of “negative risks”⁵. Domestic ecological security research began in the early 21 st century, with Xiao Duning et al.⁶ pioneering the definition of regional ecological security and proposing the “ecosystem integrity-health-service capacity” evaluation framework. Subsequently, scholars conducted numerous studies at different scales (e.g., watershed⁷, metropolitan area⁸. Evaluation methods have diversified, incorporating models such as XGBoost-MCR⁹, InVEST¹⁰, and the three-dimensional Emergy ecological footprint model¹¹. Recent advancements in remote sensing (RS) and geographic information systems (GIS) have further enhanced evaluation precision, making ecological security assessment a prominent research focus. Scholars have extensively studied models including PSR^12,13, service health risk assessment frameworks^14,15,16, and ecosystem services¹⁷.

Although research on ecological security continues to deepen, its evaluation remains significantly inadequate. Most studies tend to focus on a single dimension: some emphasize the assessment of ecosystem service provisioning capacity¹⁸, others concentrate on evaluating ecological quality and sustainable development¹⁹, some are confined to landscape ecological risk assessment²⁰, while others center on identifying the “source-corridor-choke point” pattern of ecological security²¹. Such single-perspective evaluations struggle to fully capture the multidimensional essence of ecological security as a complex system, failing to simultaneously consider the synergistic and antagonistic relationships among the system’s “supply capacity” (services), “external pressures” (risks), and “internal state” (health). The “Service-Risk-Health” model developed in this study is not a simple superposition of existing multidimensional models but an ecological security assessment methodology tailored to resource-based regions. Compared to the “Supply-Demand-Risk” framework emphasizing spatial mismatches in “service flows,” this framework introduces the “ecosystem health” dimension, strengthening diagnostics of the system’s structural and functional stability to identify cumulative developmental damage at an earlier stage. Compared to the PSR framework and its derivatives, which follow linear causality, this framework emphasizes the network of synergistic, trade-off, and antagonistic relationships among services, risks, and health, better aligning with the complex game scenarios where ecological restoration coexists with development pressures. Compared to resilience-focused “resilience-risk-service” frameworks, this framework centers on “landscape ecological risk,” with its indicator system directly addressing landscape fragmentation and habitat destruction caused by mineral resource extraction.

The deepening advancement of China’s national “carbon peak and carbon neutrality” (“dual carbon”) strategy has endowed regional sustainable development with new policy implications and target constraints. This strategy not only commits to building a clean and low-carbon energy system but also emphasizes enhancing ecosystem carbon sink capacity and promoting green transformation of territorial space. Its core lies in balancing development and security while coordinating emission reduction with carbon sink enhancement²². As a typical resource-based province in the middle reaches of the Yellow River basin, Shanxi serves both as a vital hub for national energy security and a key component of the basin’s ecological barrier²³. The “dual carbon” strategy necessitates a shift in its development model from scale-and-speed-driven to green-and-benefit-oriented²⁴. Long-term coal mining has triggered prominent ecological issues such as land subsidence and vegetation degradation. Although ecological restoration projects have achieved preliminary results in recent years, the systemic evolution mechanism of ecological security remains unclear²⁵. Therefore, systematically revealing the spatiotemporal evolution patterns and driving mechanisms of ecological security in Shanxi Province holds significant practical implications for advancing regional green transformation and high-quality development in the Yellow River Basin.

The study aims to systematically evaluate the spatiotemporal dynamics of ecological security in Shanxi Province from 2000 to 2023 by constructing an innovative three-dimensional evaluation framework integrating “ecosystem services–landscape ecological risks–ecosystem health.”(Fig. 1) The specific objectives are: (1) to quantify the evolution trajectory of ecological security at multiple scales (grid and county); (2) Reveal the driving mechanisms and interactions of natural-societal-landscape factors on ecological security patterns; (3) Trace the migration path of ecological security centers and uncover their coupling with human activities. The findings not only provide an ecological security assessment paradigm applicable to resource-dependent regions but also offer direct scientific basis for zoned ecological management.

Overview of the study area

Shanxi Province is located in the eastern Loess Plateau of northern China, spanning latitudes 34°34’N to 40°44’N and longitudes 110°14’E to 114°33’E. It is bounded by Hebei Province to the east, Shaanxi Province to the west, Henan Province to the south, and the Inner Mongolia Autonomous Region to the north (Fig. 2). The province encompasses a total area of approximately 156,700 km²²⁶.

The topography of Shanxi Province is characterized by a distinctive “two-mountains-flanking-a-basin” configuration, formed by the Taihang Mountains in the east and the Lüliang Mountains in the west. A series of basins, including those of Xinding and Taiyuan, are situated in the central region. The elevation generally decreases from the northeastern to the southwestern areas, creating a typical Loess Plateau landscape that is highly fragmented by dense gullies and severely affected by soil erosion. Climatically, the region experiences a temperate continental monsoon climate with four distinct seasons. The mean annual temperature ranges from 9 to 11 °C, and the mean annual precipitation is between 400 and 650 mm.

In 2023, Shanxi Province had a permanent resident population of 34.66 million and a gross domestic product (GDP) of 2.57 trillion yuan. This translates to a per capita GDP of approximately 73,984 yuan. The economic structure was dominated by the secondary industry, which accounted for 51.9% of the total GDP, underpinning the province’s role as a nationally critical energy supply base²⁷.

Shanxi Province is situated within the core ecological conservation zone of the Yellow River Basin. Key tributaries, such as the Fen River, provide a continuous supply of water to the main stream of the Yellow River²⁸. A suite of implemented ecological protection measures has been effective in safeguarding the quality of water entering the river. The region is recognized for its significant ecosystem services and considerable biodiversity, functioning as a vital barrier for regional ecological security. In parallel, the province is actively advancing ecological restoration and green transition demonstration projects, accelerating its shift from a traditional energy base toward a green and low-carbon development model²⁹.

Data sources and preprocessing

Data source

The datasets utilized in this study encompass land use, meteorological conditions, socioeconomic statistics, topographic attributes, and other relevant geospatial information (Table 1).

Data preprocessing

The 30-meter resolution land use raster dataset was obtained from the research group of Professor Huang Xin at Wuhan University. The original classification system, which consisted of nine categories, was consolidated into six primary types: cropland, forest land, grassland, water bodies, construction land, and unused land, using the Reclassify tool in ArcGIS. The data layer for Shanxi Province was then extracted by applying a mask based on the provincial administrative boundary. A unified spatial reference (WGS 1984 geographic coordinate system) and projection (Asia North Albers Equal Area Conic) were established for all datasets. Data with a resolution other than 1 km were resampled to a unified 1 km grid using Kriging interpolation. The required metrics for calculating the landscape ecological risk index were computed using FragStats software. Spatial joins between vector and raster data were performed to link corresponding attributes. ArcGIS10.5 version used in this article. (https://pan.baidu.com/s/10YxtXRUBZGa04F_B1ScoLQ? pwd = 6789).​FragStats4.2 version used in this article. (https://www.umass.edu/landeco/research/fragstats/downloads/fragstats_downloads.html)

This study selected the years 2000, 2005, 2010, 2015, 2020, and 2023 as temporal milestones. Based on data availability and temporal consistency, it delineates.

the typical evolutionary stages of Shanxi Province as a representative resource-based region, systematically revealing the dynamic response process of ecological security under intensive resource exploitation and sustained ecological interventions. 2000–2005 marked the period of intensifying ecological deficits and the initiation of governance efforts. Coal mining reached its peak, with prominent issues such as land subsidence and vegetation degradation. Projects like the Grain-for-Green Program began implementation. 2005–2010 transitioned into the pilot restoration and partial recovery phase. Alongside the advancement of the “Shanxi Ecological Province Construction” pilot initiative, remediation projects in key mining areas and soil erosion zones were progressively launched. 2010–2015 marked a period of heightened tension between development and conservation. Against the backdrop of an energy-dominated economy, systematic ecological restoration projects like the Fen River Basin initiative commenced, with human pressures and ecological remediation efforts intensifying simultaneously. 2015–2020 marked the transition to systematic governance and strategic-driven development. Driven by ecological civilization initiatives and the national strategy for ecological conservation and high-quality development in the Yellow River Basin, integrated efforts advanced ecological restoration in mining areas, green industrial transformation, and comprehensive ecological construction across the region. From 2020 to 2023, the region entered a consolidation phase for green development and ecological security. Under the dual carbon goals and high-quality development requirements, ecological conservation and resource-based economic transformation interacted deeply. While the ecological security pattern stabilized, structural risks persisted.

Research methods

Ecological security is a comprehensive concept encompassing multiple dimensions and processes. To overcome the limitations of traditional single-dimensional assessments, this study constructs a three-dimensional evaluation framework integrating “ecosystem services—landscape ecological risks—ecosystem health.” Theoretically, this framework places all three components on equal and complementary levels, representing the functional outputs, external stresses, and intrinsic states of ecosystems respectively, thereby enabling a systematic assessment of ecological security. The study constructs an ecological security evaluation system from three perspectives—ecosystem services, landscape ecological risks, and ecosystem health³⁰—to assess the ecological security status of Shanxi Province. The specific research framework is illustrated in Fig. 3.

Ecological Security Index

Where: ESI denotes the Ecological Security Index, ESsI represents the Ecosystem Services Index, ERI signifies the Landscape Ecological Risk Index, and EHI indicates the Ecological Health Index.

Entropy weighting method

To avoid excessive influence of subjective factors on the evaluation weights, this paper employs the entropy weight method to calculate the weights among various indicators³¹. The specific calculation formula is as follows:

Where: ({W}_{j}) represents the weight of the jth indicator, ({P}_{ij}) denotes the indicator weight of the jth indicator, and ({e}_{j}) signifies the entropy value of the jth indicator.

Ecosystem services Index (ESsI)

Building upon prior research³², this study employs six indicators—Food Production Services³³(FP), Soil Conservation Service (SC), Water Yield (WY), Habitat Quality (HQ), Carbon Storage Services (CS), and Recreational Service³⁴(RS)—to conduct a quantitative assessment of ecosystem service functions in Shanxi Province. The calculation formula is as follows:

Where: (FP) denotes food production, (SC) represents soil conservation, (WY) indicates water yield, (HQ) signifies habitat quality, (CS) stands for carbon storage, (RS) refers to recreational services, and a, b, c, d, e, and f denote the weighting coefficients corresponding to each indicator (Table A 1).

Ecological risk index (ERI)

The ecological risk index is a consistent evaluation method for measuring the adverse consequences faced by ecosystems under external pressures³⁵. Its specific calculation formula is as follows (Table 2):

Ecosystem health index (EHI)

The ecosystem health index serves as an effective method for evaluating the safety status of ecosystems. This study employs the BPSR model to assess ecosystem health³³, with the formula defined as follows:

Where: B, P, S, and R represent the composite values of the foundation layer, pressure layer, state layer, and response layer, respectively.

To eliminate differences between each indicator, they were normalized. The weights corresponding to each indicator were calculated using the entropy weight method, and finally the ecological health index was computed (Table A 2).

Uncertainty analysis in ecological security assessment

To quantify the uncertainty in ecological security assessment results, this study employs a Monte Carlo simulation method that systematically considers uncertainty sources from data, parameters, and model structure³⁶. For ecosystem service and ecosystem health weights, random perturbations were applied using a Dirichlet distribution (concentration parameter α = 0.1). A total of 1000 independent simulations were conducted, with all intermediate indicators and the final Ecological Security Index (ESI) recalculated for each iteration. The robustness and spatial patterns of uncertainty in the evaluation results were systematically assessed by calculating the standard deviation, coefficient of variation, and 95% confidence intervals for the ESI of each grid cell.

Where: CV denotes the coefficient of variation, S denotes the sample standard deviation, and x̅ denotes the sample mean.

Center of gravity-standard deviation ellipse

Center of Gravity-Standard Deviation Ellipse Analysis simultaneously determines the orientation and distribution of a set of discrete points by calculating the standard deviation between the mean center and other points in the dataset³⁷. The formula is as follows:

In the formula: ({SDE}_{x}) and ({SDE}_{y}) represent the center coordinates of the standard deviation ellipse, where x and y denote the spatially weighted average coordinates; θ is the azimuth angle of the standard deviation ellipse; x̅ and y̅ denote the coordinate differences between the mean center and the coordinates ({x}_{i}) and ({y}_{i}).

Geographic detector

The core concept of the Geographic Detector is to decompose a geographic phenomenon into multiple sub-factors, systematically analyze the relationships between these sub-factors and environmental variables, and thereby determine the strength of influence each factor exerts on the geographic phenomenon along with its spatial distribution characteristics. It represents a novel statistical method for detecting spatial differentiation and revealing the underlying driving factors²⁸.

(1) Factor detection.

The spatial differentiation degree of a single factor X to the dependent variable Y is usually represented by q value. The calculation formula is:

Where: q is the strength of explanatory power; Lis the classification number of dependent variables and independent variables; h is the number of driving factor layers; N and N_h are the number of samples of global and each classification h, respectively; σ² is the variance of ESI.

(2) Interaction detection.

Interaction detection can identify the interaction of different driving factors on the dependent variable, that is, whether the evaluation factors X1 and X2 will increase or decrease the explanatory power of the dependent variable Y, or the influence of these factors on Y is independent of each other. There are five results of the interaction between the two independent variables on the dependent variable.(Table 3).

Spatiotemporal variations in ESs, ER, and EH

From 2000 to 2023, the ecological security of Shanxi Province revealed pronounced spatiotemporal heterogeneity across its three constituent dimensions: ecosystem services (ESs), ecological risk (ER), and ecosystem health (EH) (Fig. 4).

Regarding ecosystem services, the province showed a steady increasing trend overall. The mean Ecosystem Services Index (ESsI) increased from 0.5892 in 2000 to 0.6139 in 2023, representing a growth of 4.19%. Spatially (Fig. 5), Areas of high ecosystem service provision persistently and stably clustered in major mountain ranges, such as the Taihang and Lüliang Mountains, which are characterized by robust ecological foundations. These regions support dense vegetation cover and exhibit strong capacities for ecological.

regulation and support functions. Zones of medium service levels were primarily distributed around major river and lake systems, including the Yellow River and Fen River, forming crucial ecological corridors that connect high-value areas. Notably, traditional low-service zones, represented by the Yuncheng and Linfen basins, showed a marked contraction trend during the study period, while medium- and high-service zones expanded correspondingly. These spatiotemporal dynamics suggest that a series of ecological conservation measures—such as the Grain for Green Program and soil-water conservation projects—have yielded positive outcomes in enhancing the regional capacity for ecosystem service provision.

In terms of ecological risk, the province exhibited a fluctuating but overall upward trend. The mean Ecological Risk Index (ERI) increased from 0.6673 in 2000 to a peak of 0.6923 in 2005. Although it subsequently declined, the index reached 0.6791 by 2023, reflecting a net increase of 1.76% over the entire period. Spatially (Fig. 6), high-risk areas were predominantly concentrated in plains, basins, and certain water bodies subject to intensive human activity and severe landscape fragmentation. These regions experience strong pressure from urbanization and industrial/agricultural production. In contrast, low-risk zones showed a strong spatial correspondence with areas of high ecosystem service provision, primarily located along the eastern and western mountain ranges. This pattern underscores the critical role of natural ecosystems as barriers against external risk disturbances.

Regarding ecosystem health, the province showed a concerning trajectory, with its mean Ecosystem Health Index (EHI) exhibiting a pronounced “M-shaped” downward trend. The index decreased from 0.5042 in 2000 to 0.4023 in 2023, marking a substantial overall decline of 20.21%. This significant deterioration reveals that, despite the enhancement of some ecosystem service functions, the structural integrity and functional stability of the ecosystems themselves have been subjected to immense pressure. Spatially (Fig. 7), areas of high ecosystem health remained confined to regions with robust ecological foundations, such as the Taihang and Lüliang Mountains. In contrast, zones of medium-to-low health were widely distributed across major basins, including Yuncheng and Linfen, as well as parts of the plains. These areas are typically characterized by high population density, intensive economic activity, and pronounced environmental stress, demonstrating a clear spatial correspondence between the degradation of ecosystem health and the intensity of anthropogenic pressures.

Reliability analysis and model validation of shanxi province’s ecological security evaluation results

During the study period, the provincial mean Ecological Security Index (ESI) showed minimal change, increasing only marginally from 0.4556 in 2000 to 0.4574 in 2023. In contrast, the mean Remote Sensing-based Ecological Index (RSEI) demonstrated a substantial improvement, rising from 0.2913 to 0.4925 over the same period. These complementary trends, particularly the pronounced enhancement captured by the RSEI, collectively suggest an overall improvement in the ecological environment of Shanxi Province.

To assess the consistency between the two ecological indices, a correlation analysis was performed between the Ecological Security Index (ESI) and the Remote Sensing-based Ecological Index (RSEI) (Fig. 8). From 2000 to 2023, ESI and RSEI.

showed a statistically significant positive correlation, with the coefficient of determination (R²) ranging from 0.463 to 0.570 (p < 0.001). This result supports the validity of the ecological security evaluation. Spatially, the patterns of ESI and RSEI were generally consistent. with low values concentrated in plains and basins, and high values clustered in mountainous regions(Fig.A 1). It is worth noting that in some areas, ESI values were relatively lower compared to RSEI. This can be attributed to the fact that ESI incorporates socioeconomic and environmental factors, whereas RSEI is derived solely from remote sensing-based ecological indicators.

To quantify the uncertainty associated with the ecological security assessment, a Monte Carlo simulation was applied (Table 4). The results demonstrated high reliability of the ecological security assessments in Shanxi Province during the study period, with coefficients of variation (CV) ranging from 0.081 to 0.089—all below the common reliability threshold of 0.1. Temporally, the CV showed a gradually increasing trend, rising from 0.081 in 2000 to 0.089 in 2023—a 9.9% increase—which reflects a slight elevation in assessment uncertainty over time. This trend may be attributed to the continuous intensification of human activities and the growing complexity of landscape patterns. Spatially, correlation analysis between the mean ESI values and the CV revealed a statistically significant negative relationship (r = −0.713, p < 0.01). This indicates that regions with higher ecological security generally exhibit lower uncertainty in assessment results, whereas areas with lower ecological security are associated with higher uncertainty. Zones of high ecological security are typically characterized by stable ecosystem structures and more predictable responses to external disturbances, leading to highly reliable assessment outcomes. In contrast, regions with low ecological security are often influenced by multiple interacting factors, such as rapid urbanization and intensive industrial or agricultural activities, resulting in more complex ecological responses and consequently greater uncertainty in assessments(Fig. 9).

The classification of ecological security (ES) levels is fundamental to its assessment, yet standardized criteria for defining these levels remain lacking. To address this, the present study established a five-tier classification system for the Ecological Security Index (ESI) tailored to the specific context of Shanxi Province. The classification intervals are as follows: Low (ESI ≤ 0.35), Medium-Low (0.35 < ESI ≤ 0.40), Medium (0.40 < ESI ≤ 0.45), Medium-High (0.45 < ESI ≤ 0.50), and High (ESI > 0.50).

Spatiotemporal variations in ecological security across multiple scales in Shanxi Province

Spatial-temporal variations in ecological security at the grid scale in Shanxi Province

During the study period, the ecological security of Shanxi Province was predominantly classified at the medium, medium-high, and high levels. Spatially, this translated into a pattern of higher security levels in the eastern and western regions and a lower level in the central part of the province.

Temporally, the provincial mean Ecological Security Index (ESI) exhibited an “inverted-N” fluctuating trend but demonstrated an overall upward trajectory from 2000 to 2023. The index initially decreased from 0.4556 in 2000 to 0.4448 in 2005, then rose steadily to 0.4584 in 2020, before a slight decline to 0.4574 in 2023. In terms of security level composition, the high-security category consistently accounted for the largest proportion of the province, exceeding 28% annually. This was followed by the medium-security category, which maintained a share of over 20% each year. In contrast, the low-security category constituted the smallest proportion, remaining below 8% throughout the study period.

Spatially, ecological security demonstrated a distinct altitudinal gradient, with security levels generally decreasing as elevation declined (Fig. 10). Areas of high ecological security were primarily distributed in mountainous regions, including the Taihang and Lüliang Mountains. Moderate security levels were mainly observed along major rivers and lakes, such as the Yellow River and Fen River, as well as in transitional zones between mountains and plains or basins—characterized by valley topography and relatively low elevations. In contrast, low-security areas were predominantly concentrated in basin regions, exemplified by the Yuncheng and Linfen Basins, which feature flat terrain and low-lying elevations.

During the study period, transitions between ecological security levels in Shanxi Province predominantly occurred among the low-to-medium, medium, and medium-to-high categories (Fig. 11).

Between 2000 and 2005, the most significant shift in ecological security levels was the transition from medium to medium-low security, accounting for 32.92% of the total changes. This transition was predominantly observed between the Xin-Ding Basin and the Changzhi Basin. The second most frequent transition was from medium-high to medium security (22.91%), concentrated mainly in the Xinzhou City area.

During 2005–2010, the largest transition was from medium-high to high security (19.33%), occurring primarily between the Yunzhong and Lüliang Mountains. This was followed by a shift from medium-high to medium security (17.57%), mainly identified in the Fen River upstream and PiGuang River regions.

From 2010 to 2015, the most notable transition was from medium-low to medium security (31.10%), widely distributed across several basins, including the Yuncheng, Taiyuan, and Jincheng Basins. The second most common transition was from medium to medium-high security (17.19%), largely located in ecotones between mountainous areas and plain basins.

Between 2015 and 2020, the dominant transition remained from medium-low to medium security (20.37%), primarily occurring between plain basins and river-lake water areas. This was followed by a shift from medium to medium-high security (18.00%), distributed mainly along the Yellow River corridor.

In the most recent period (2020–2023), the most prominent transition was from medium-high to high security (30.27%), concentrated predominantly in the Taihang Mountains. The second largest transition was from medium to medium-high security (20.44%), again observed primarily in transitional zones between mountains and plains.

Spatiotemporal Variations in Ecological Security at the County Level in Shanxi Province

At the county level, the spatial distribution of ecological security grades in Shanxi Province exhibited distinct regional characteristics, following a general “high in the southeast, low in the northwest” pattern (Fig. 12), which remained largely stable throughout the study period.

Twelve counties, including Qinyuan, Anze, Qinshui, and Yuanqu, consistently maintained high ecological security grades. These areas are characterized by dense vegetation and high forest coverage, strong water conservation capacity due to proximity to rivers and lakes, rich biodiversity, and minimal anthropogenic disturbance—collectively contributing to their superior ecological security.

Sixteen counties, such as Tianzhen, Youyu, Shenchi, and Wuzhai, persistently exhibited a medium ecological security level. This group primarily represents two scenarios: one where initially poor ecological conditions have been improved through conservation measures like the Grain for Green Program and afforestation; and another where a relatively high proportion of artificially green areas has moderately elevated the ecological security status.

In contrast, Pingcheng District consistently recorded a low ecological security level. As a major coal mining area, extensive coal extraction has significantly damaged the local ecological environment, resulting in its persistently low rating.

From a dynamic perspective, the distribution of high- and low-security zones remained relatively stable; however, certain counties exhibited noticeable fluctuations in their ecological security levels. For instance, several counties in the northwestern region—such as Pi Guan and Baode—underwent transitions between medium and medium-high security levels. These fluctuations signify a degree of ecological vulnerability but also indicate a latent potential for recovery, despite their overall security levels remaining lower than those observed in the southeastern part of the province.

Overall, the ecological security pattern of Shanxi Province demonstrated relative stability during the study period, albeit with pronounced regional heterogeneity in security levels. The high-security zones in the southeastern region function as critical ecological barriers, underpinning environmental protection efforts across the province. In contrast, the low-security zones in the northwestern region necessitate targeted ecological investments to enhance regional ecological security and foster sustainable development.

Analysis of shifts in Shanxi Province’s ecological security focus

To investigate the spatiotemporal shifts in the ecological security gravity center of Shanxi Province, this study employed the standard deviation ellipse method to analyze its migration trajectory (Fig. 13). The ecological security center was calculated for six time points: 2000, 2005, 2010, 2015, 2020, and 2023. The results show that the center was predominantly located in the northwestern part of Qixian County. An analysis of its migration trajectory revealed the following patterns: from 2000 to 2005, it shifted 0.31 km northeast; between 2005 and 2010, it moved 1.65 km southwest; from 2010 to 2015, it migrated 0.71 km southeast; between 2015 and 2020, it shifted 1.85 km northwest; and finally, from 2020 to 2023, it moved 1.19 km southeast.

Throughout the study period, the ecological security gravity center of Shanxi Province shifted a total of 0.6454 km in a southwestern direction. Although the net displacement was limited, notable local fluctuations were observed in specific periods. These trajectory changes not only reflect the spatiotemporal dynamics of the ecological security center but also offer valuable insights for identifying key regions influencing ecological security and potential risk hotspots.

Ecological security driving mechanism in Shanxi Province

To investigate the driving mechanisms underlying the evolution of ecological security in Shanxi Province, this study applied the geographical detector method based on the combined evaluation outcomes of ecosystem services (ESsI), ecological risk (ERI), and ecosystem health (EHI). A set of 12 driving factors was selected to construct an indicator system reflecting the spatial heterogeneity of ecological security, encompassing 6 natural environmental factors, 4 socioeconomic factors, and 2 landscape pattern factors(Table 5).

These factors were chosen to comprehensively represent the province’s geographic, socioeconomic, and landscape context. Prior to analysis, the variance inflation factor (VIF) and condition number were used to assess multicollinearity among the explanatory variables. All VIF values were below 5, and all condition numbers were below 30, indicating the absence of severe multicollinearity (Table 6). Based on this, the Geodetector model was employed to identify and quantify the influence of each driving factor on the spatial differentiation of ecological security across the province.

During the study period, ecological security drivers in Shanxi Province exhibited a distinct pattern characterized by “alternating dual cores and a stable tripolar state” (Fig. 14).

Analysis of the driving factors across the six study time points (2000, 2005, 2010, 2015, 2020, 2023) reveals a consistent group of dominant drivers shaping the spatial pattern of ecological security in Shanxi Province. The normalized difference vegetation index (NDVI), the proportion of construction land, and population density consistently emerged as the top three most influential factors, albeit with alternating primary dominance (see Table X for complete rankings). This dynamic, wherein the explanatory power (q-statistic) of NDVI and construction land proportion alternated in claiming the highest position, concretely manifests the “alternating dual-core” characteristic of the driving mechanism. Underlying this dual-core oscillation, factors such as slope, elevation, and GDP perennially formed a secondary tier of stable influence, thereby constituting the “stable tripolar” component of the overall pattern.

The remaining factors, including precipitation, distance to roads/rivers, temperature, and landscape metrics (Shannon’s evenness and diversity), consistently demonstrated relatively lower explanatory power throughout the study period.

Across the study period, NDVI emerged as the dominant explanatory factor in 2000, 2005, 2015, and 2023, underscoring vegetation cover as the cornerstone of the ecological baseline. In 2010 and 2020, however, the proportion of construction land surpassed NDVI as the primary driver, reflecting the phase-dominant impact of.

rapid urbanization on ecological security. This alternating dominance between the “ecological baseline” (represented by NDVI) and “land development intensity” (represented by construction land) illustrates the “dual-core” dynamic in the driving mechanism. Throughout these fluctuations, NDVI, population density, and the proportion of construction land consistently remained the top three drivers, collectively forming a stable “tripolar” structure that underpins the spatial pattern of ecological security in the province.

To elucidate how interactions among driving factors influence ecological security, an interaction detector was applied to identify the combined effects (Fig. 15). The results demonstrate that the spatial heterogeneity of ecological security in Shanxi Province arises from the nonlinear and synergistic coupling of multiple drivers.

The interaction between any two factors resulted in either two-factor enhancement or nonlinear enhancement, with the combined explanatory power exceeding the sum of their individual contributions. Within this interaction network, NDVI, the proportion of construction land, and population density functioned as core interaction hubs. The interaction between NDVI and construction land was particularly pronounced, revealing a profound conflict between the natural ecological baseline and anthropogenic development pressure. Meanwhile, the strong interactions of population density with both NDVI and construction land reflect the cumulative pressures exerted by human activities, manifested through direct land conversion and indirect impacts on vegetation ecology. Furthermore, significant interactions were observed between NDVI and GDP, as well as between topographic and anthropogenic factors, highlighting the complex coupling mechanisms involving economic development, topographic constraints, and human activities.

Recommendations for ecological security development in shanxi province

Based on a systematic analysis of the spatiotemporal evolution and driving mechanisms of ecological security, this study identifies “natural vegetation baseline (NDVI)” and “human development intensity (proportion of construction land)” as the core factors regulating the ecological security pattern in Shanxi Province. To effectively enhance regional ecological security safeguards and support the ecological conservation and high-quality development strategy for the Yellow River Basin, the following targeted ecological governance strategies are proposed through zoning and classification.

Consolidate ecological barrier zones, enhance service provision and ensure healthy stability

For high-value ecological security zones such as the Taihang Mountains and Lüliang Mountains, their core function is to provide stable ecosystem services and maintain system health. Countermeasures should focus on “consolidation and enhancement”.

Implement targeted forest quality enhancement projects. While increasing vegetation coverage (NDVI), prioritize optimizing forest stand structure and restoring native tree species. This will strengthen water conservation, biodiversity protection, and the stability of high-carbon-storage ecosystems, thereby fundamentally consolidating the ecological foundation.

Strictly control the conversion of ecological spaces. Delineate and strictly enforce ecological protection red lines, restrict mineral resource exploration and mining, and limit large-scale linear engineering projects to prevent landscape fragmentation from damaging ecosystem integrity.

Refine diversified ecological compensation mechanisms. In key counties such as Qinyuan and Anze, pioneer provincial-level pilot programs for calculating and trading “incremental forest carbon sinks” and “water conservation capacity.” Link ecological compensation funds to these ecological function indicators, providing targeted support for green industries like forest-based economies and eco-tourism to foster a virtuous cycle between ecological conservation and improved livelihoods.

Regulating the urban-rural transition zone to alleviate human-induced stress and risk concentration

For low-value ecological security zones and high-risk areas such as central basins and urban expansion zones, the core task is “reduction and restoration”:

Strengthen land use regulation. Set urban development boundaries using the proportion of construction land as a key control indicator to curb sprawling expansion. Optimize the internal structure of construction land, revitalize existing resources, and promote intensive and efficient utilization.

Build an integrated blue-green ecological network linking urban and rural areas. In urban clusters like Taiyuan and Linfen, systematically plan and construct ecological corridors based on rivers, green belts, and parks to mitigate habitat fragmentation, enhance landscape connectivity, and strengthen urban ecosystem resilience.

Promote ecological transformation of industries. In traditional industrial and mining areas, strictly enforce environmental standards, promote clean production technologies, and implement coordinated soil and water remediation in key pollution zones to reduce pressure on ecosystem health at its source.

Implement targeted ecological restoration to promote recovery and stability in vulnerable areas

For counties in the northwest and certain areas with significant fluctuations in ecological security levels, countermeasures should focus on “restoration and adaptation”:

Implement differentiated ecological restoration projects. In areas with restoration potential, such as Pi Guan and Baode, prioritize comprehensive soil erosion control and degraded grassland restoration projects. In coal mining hubs like Pingcheng District, give priority to comprehensive remediation and land reclamation of subsidence areas and open-pit mine pits.

Develop adaptive ecological industries. Encourage the growth of under-forest economies, eco-tourism, and specialty economic orchards within ecological restoration zones. Integrate ecological governance with regional economic development to foster sustainable endogenous momentum.

Improve the integrated monitoring, early warning, and decision-making management system

To achieve intelligent and forward-looking ecological security governance, it is recommended to establish a multi-scale dynamic monitoring platform that integrates remote sensing, geographic information systems, and ground monitoring networks. This platform should enable real-time tracking of core indicators for ecosystem services, risks, and health, while utilizing the trajectory of ecological security shifts as a key indicator for regional ecological stability early warning.

Establish an ecological security early warning and response mechanism. Set tiered warning thresholds based on key drivers identified by geographic detectors (e.g., NDVI, proportion of construction land, population density) to provide scientific basis for local governments to implement preemptive control measures.

Promote the deep integration of ecological security assessment outcomes with planning systems. Systematically incorporate the ecological security patterns, zoning conclusions, and driving mechanisms identified in this study into territorial spatial planning, ecological environment protection, and socioeconomic development plans at provincial, municipal, and county levels, achieving a closed-loop “assessment-planning-management” system.

Research contributions and innovations

This study developed a three-dimensional assessment framework integrating “ecosystem services—landscape ecological risks—ecosystem health” to systematically diagnose the spatiotemporal evolution and driving mechanisms of ecological security in Shanxi Province from 2000 to 2023.

This study breaks through the limitations of traditional single-dimensional ecological security assessments. Unlike previous research focusing solely on ecosystem services¹⁸, landscape ecological risks²⁰, or the PSR model¹², this framework integrates the “positive benefits” (services), “negative pressures” (risks), and “intrinsic state” (health) of ecosystems for a comprehensive evaluation. This design does not merely superimpose existing multidimensional models but constructs an assessment tool tailored to the unique context of resource-based regions, where fragile ecological foundations coexist with intense development pressures. This tool reveals complex internal trade-offs within the system. The identified asynchronous evolution trend of “service enhancement, risk intensification, and health degradation” validates the diagnostic value of this framework, capturing systemic characteristics that single-dimensional assessments struggle to detect.

This study reveals the unique patterns of ecological security evolution in resource-based regions. It finds that Shanxi Province exhibits a “dual-core alternation, tri-polar steady state” pattern in its ecological security drivers, where NDVI (ecological foundation) and proportion of construction land (development intensity) alternately dominate, forming a tri-polar structure with population density that explains the strongest variation. This finding resonates with Liu et al.⁴² research in northeastern Qinghai-Tibet Plateau, suggesting that alternating dominance of “natural-human” factors may serve as a universal early warning signal for identifying regional developmental phase transitions. The micro-scale oscillation trajectory of the ecological security center in northwestern Qixian, coupled with the spatiotemporal coupling of the Taiyuan metropolitan area’s expansion, further validates the sensitivity of the ecological security center as an indicator for identifying hotspots of human disturbance. These findings provide new empirical evidence for understanding the intrinsic patterns of ecological security evolution in resource-based regions.

Based on a systematic diagnosis of Shanxi Province’s 24-year ecological security evolution, this study identifies high-value ecological barrier zones, low-value risk accumulation zones, and fluctuating vulnerability zones. It proposes targeted ecological governance strategies tailored to each zone and category. These strategies directly support the implementation of the“Ten Major Actions for Carbon Peaking”⁴¹ and the “Ecological Protection and High-Quality Development Plan for the Yellow River Basin”⁴² providing actionable pathways for resource-based regions to balance development and conservation under the dual carbon goals.

Research limitations and future directions

Limited by its 1000 m resolution, ESI remains inadequate in capturing micro-scale local details. Although CESI showed a slight decline in 2023, the Remote Sensing Ecological Index (RSEI) continued to rise, exhibiting a significant positive correlation (R² > 0.46, p < 0.001). However, RSEI alone is insufficient to comprehensively characterize ecological security evolution. Future work will integrate hyperspectral data and deep learning algorithms to precisely identify land-use transitions. Accuracy validation through field sampling and synchronized drone observations will enable the integration of “centimeter-level” ground truth with “kilometer-level” remote sensing products, further enhancing assessment reliability.

This study selected six time points between 2000 and 2023 for analysis, depicting the phased evolution of ecological security. However, this discrete time-point design fails to fully capture the continuous dynamic process of ecosystem responses to disturbances and restoration measures, particularly the lagged and cumulative effects of ecological restoration projects. For instance, the ecological benefits of the Grain-for-Green Program typically take 5–10 years to fully manifest, and the five-year intervals used in this study may struggle to accurately depict this gradual process. Future research could explore incorporating annual continuous data combined with time-series analysis methods to more precisely track the evolutionary trajectory of ecological security.

Although this study incorporated 12 driving factors across three major categories—natural, social, and landscape—it has not fully accounted for the long-term lag effects of climate change and major engineering projects. Against the backdrop of global change, the impact of shifts in the frequency and intensity of extreme climate events on regional ecological security cannot be overlooked. The long-term effects of human interventions such as inter-basin water transfers and large-scale ecological restoration projects also warrant in-depth exploration. Future research could integrate CMIP6 climate scenarios with land-use simulation models like PLUS and FLUS to conduct multi-scenario dynamic simulations of ecological security. This would enable predictions of ecological security evolution under different climate change and policy intervention pathways, providing scientific basis for developing forward-looking regulatory strategies.

The study centers on ecological security, constructing an integrated evaluation model that combines ESs, ER, and EH. The reliability of the evaluation results is validated using the Remote Sensing Ecological Index (RSEI), systematically characterizing the spatiotemporal evolution patterns and driving mechanisms of ecological security in Shanxi Province from 2000 to 2023. The study identifies an asynchronous evolution trend across ecological security dimensions within the region: “service enhancement, risk intensification, and health degradation.” This indicates that while ecosystem service functions have improved rapidly in some areas due to intensive human restoration efforts, structural losses caused by high-intensity development continue to erode the baseline health of the system. This reveals the long-term and complex nature of ecological restoration in resource-based regions. The ecological security pattern exhibits stable spatial differentiation with “higher levels in the southeast and lower levels in the northwest,” while temporal fluctuations follow an inverted “N” shape with overall improvement. The ecological security center exhibits a cyclical micro-migration in the northwest of Qixian County. Its trajectory, spatiotemporally coupled with the expansion of the central urban cluster, serves as a dynamic indicator of regional ecological stability fluctuations and human-induced stress hotspots. Analysis of driving mechanisms reveals a unique “dual-core alternation, tri-polar steady state” pattern. NDVI (representing ecological baseline and resilience) and the proportion of construction land (representing development intensity) alternated as the primary dual-core drivers. Together with population density, they formed a stable tripolar structure with the strongest explanatory power. Any interaction between any two factors exhibited either dual-factor or nonlinear amplification, empirically demonstrating the complex coupled effects of natural elements and human activities on the ecological security pattern.