Urban agglomerations are typical compound systems integrating nature, society, and economy, facing the interactive impacts of multidimensional driving factors. Their ecological risks are characterized by multiple sources of input, exposure diffusion, and complex disturbance mechanisms (Babí, 2023). This study follows the principles of systematics, quantifiability, and relevance to ecological processes. It constructs an indicator system from the dual perspectives of natural constraints and human activity disturbances, comprising four categories of elements: geographical foundation, landscape pattern, human activities, and natural disasters. This forms an information-closed network of “natural foundation constraints–landscape functional responses–anthropogenic disturbance pressures–disaster risk feedbacks,” which is used to delineate the spatial differentiation characteristics of landscape ecological risk.

In the geographical foundation elements, topographic position index, terrain relief, and soil type are selected to represent the terrain stability and soil foundation conditions of ecosystems. The topographic position index and terrain relief can reflect the potential risks of erosion and geological disasters (Li et al., 2020c), while soil type influences habitat quality and ecosystem regulation capacity through its carbon sequestration ability and resistance to erosion. In the landscape pattern elements, the landscape disturbance index, NDVI, Shannon’s diversity index, and contagion index are chosen to reflect the structural stability of the system. The former indicates the degree of landscape disturbance (commonly calculated using patch density, landscape division index, and landscape separation index), while the latter three respectively reflect the extent of vegetation coverage, landscape heterogeneity, and integrity, all of which are negatively correlated with landscape ecological risk (Yan et al., 2021). The human activity elements cover population aggregation pressure, economic pressure (GDP), air pollution pressure (PM2.5), energy consumption pressure (nighttime light intensity), transportation network pressure, and land use pressure, which collectively depict the ecological stress caused by urbanization. These dynamic factors directly reflect environmental pollution, infrastructure load, and fragmentation of the ecological base (Mentaschi et al., 2022). In the natural disaster elements, geological disasters, rainstorm flooding, and land surface thermal environment are selected to characterize the system’s threshold for collapse. Geological disasters describe the risks of extreme climate events such as landslides and debris flows, rainstorm flooding reflects the risk of flooding caused by extreme precipitation, and the land surface thermal environment reveals the urban heat island effect, with higher land surface temperatures indicating potential ecological risks (Jing et al., 2021).

To eliminate the dimensional effects of different indicators, all indicators are processed using range standardization, with a value range of [0,1] (Table 2). In the weight assignment, the final weights of the indicators are derived by averaging the Analytic Hierarchy Process (AHP) and the Standard Correlation Weight Analysis Method (CRITIC) methods (Zhang et al., 2023b). AHP constructs a judgment matrix based on pairwise comparisons, calculates the weight vector, and passes the consistency test (CR = 0.0516, meeting the standard CR< 0.1); CRITIC objectively assigns weights based on the standard deviation and Pearson correlation coefficient matrix. The combination of these two methods effectively reduces subjective intervention, enhances the scientific and rational nature of indicator weighting, and thereby improves the reliability and explanatory power of the ecological risk assessment.

The study refers to the improved ecosystem health assessment model (Peng et al., 2011) and combines the assessment of ecosystem service functions to evaluate the ecosystem health level of the study area (Equations 4-5).

Where: EHI represents Ecosystem Health Index; PH represents the health of ecosystem structure and organization; ESV represents Ecosystem Service Value, which includes carbon sequestration, habitat quality, water retention, and soil conservation services. These services are calculated using the InVEST 3.14 model, normalized, and then summed to obtain the overall ecosystem service value (Tian et al., 2021). EV, EO, and ER represent the vitality, organization, and resilience of the ecosystem, respectively.

Vitality is the metabolism or primary productivity of the ecosystem, characterized by Net Primary Productivity (NPP) as an indicator of ecosystem vitality. Organization force is the quantity and diversity of interactions among various components of the ecosystem, evaluated using Landscape Heterogeneity (LH), Landscape Connectivity (LC), and Landscape Morphology (IPC) (Equation 6).

Where: SHDI is the Shannon Diversity Index; AMPFD is the Area-Weighted Mean Patch Fractal Dimension; PD is the Patch Density; CONTAG is the Contagion Index; PD1 and PD2 are the patch densities of water bodies and forest landscapes, respectively. COHESION1 and COHESION2 are the contagion indices of water bodies and forest landscapes, respectively.

Resilience is the ability of an ecosystem to maintain its functions and structures in the face of external risk disturbances (Jiang et al., 2024). The study sets a resilience coefficient based on the difficulty of recovery for different land use types, and represents the ecosystem’s resilience as the weighted sum of the areas of each land use type and their respective ecological recovery coefficients (Equation 7) (Lei et al., 2023).

Where: ER represents the Ecosystem Resilience; A represents the total area of land use; A_i represents the area of the i type of land use. RCi represents the ecological recovery coefficient of the i type of land use. The ecological recovery coefficients for cultivated land, forest, grassland, waterbody, construction land, and bare land are 0.3, 0.8, 0.6, 0.8, 0.2, and 1, respectively (Pan et al., 2021).

Z-score standardization transforms data with different units and scales into a standard normal distribution, effectively eliminating dimensional disparities among indicators and improving their comparability (Wu et al., 2022). In this study, Z-score standardization is used to construct a two-dimensional coordinate system, with LER on the vertical axis and EHI on the horizontal axis. Based on this framework, the study area is divided into four ecological zones: Strict Ecological Protection Zone (High risk, High health): Ecosystems with high functions but under significant external stress. These areas require enhanced ecological protection and risk control; Ecological Control Zone (High risk, Low health): Fragile ecosystems with high-risk exposure, which are priority intervention and key control areas; Ecological Conservation Zone (Low risk, High health): Areas with stable ecosystem functions and low risk exposure. These zones should be preserved to maintain ecological stability; Ecological Enhancement Zone (Low risk, Low health): Degraded ecosystems with low external pressure but high potential for restoration, representing key areas for ecological improvement. This zoning approach captures the interaction between ecological conditions and risk disturbances, highlights spatial heterogeneity, and provides a scientific basis for differentiated ecological management and targeted conservation policies. The Z-score standardization calculation formula is (Equations 8-10).

where, xi represents the value of LER or EHI for each grid cell; x¯ represents the average value of LER or EHI for the entire study area; s represents the standard deviation of LER or EHI for the entire study area; n represents the total number of grid cells.

The Coupled Model Intercomparison Project Phase 6 (CMIP6) recommends the use of Shared Socioeconomic Pathways (SSPs) and Representative Concentration Pathways (RCPs) to simulate the combined effects of socio-economic and climatic changes on land use patterns (Chen et al., 2020a; Riahi et al., 2017; van Vuuren et al., 2011). This study adopted three SSP–RCP scenarios—SSP126, SSP245, and SSP585—to explore potential trajectories of socioeconomic development and climate change and to analyze the spatiotemporal evolution of ecosystems under multiple futures. SSP126 represents a low-emission sustainable development pathway, assuming slow population growth, coordinated economic–environmental development, and limited climate change. SSP245 represents a medium-emission pathway, assuming moderate population growth, economic expansion, and climate change, consistent with current trends. SSP585 represents a high-emission pathway, assuming rapid population growth, intensified industrialization, high energy dependence, and strong warming with precipitation variability. The SSP–RCP combinations incorporate four drivers: population, GDP, precipitation, and temperature. GDP projections are derived from global SSP datasets (Murakami et al., 2021). Population data are obtained from China’s kilometer-scale grid dataset (Chen et al., 2020b). Climate variables (precipitation and temperature) are based on future kilometer-scale climate grids for China under SSP–RCP pathways (Chen and Ning, 2024b). The analysis was conducted at a high-resolution grid scale. Land use policies and management practices were assumed constant, while short-term shocks from sudden socioeconomic or extreme climate events were excluded. Based on these assumptions, multiple future development trajectories were constructed to support land use simulations and ecosystem response assessments (Table 3).

Among simulation models, the Markov model effectively predicts changes in land use quantity and spatial distribution (Tan et al., 2019), while system dynamics (SD) models are particularly suitable for capturing system complexity, nonlinear interactions, and temporal evolution, offering strong support for scenario analysis and decision-making (Wang et al., 2022b).The System Dynamics (SD) model is used to capture the nonlinear relationships between socio-economic development and land use change by representing causal feedback mechanisms among multiple driving factors (Song et al., 2024). It includes four key subsystems: economic, population, climate, and land use (Figure 3). The economic subsystem simulates the influence of infrastructure investment on the expansion of construction land, which in turn reduces ecological land. The population subsystem reflects changes in urban-rural migration and associated consumption demands. The climate subsystem captures the direct effects of temperature and precipitation changes on habitat quality and ecosystem stability. The land use subsystem integrates feedback from population and economic growth, driving the conversion among different land use. Based on the dynamics of land use transformation, a system dynamics model was developed using Vensim PLE x64. The simulation spans from 2000 to 2050, with a one-year time step. The period from 2000 to 2020 is used for historical simulation and model calibration. Simulation results show that the average relative error between simulated and actual land use values is less than 2%. The SD model demonstrates high accuracy in reproducing land use change and is suitable for forecasting future land use demand. Ultimately, the model is driven by population, climate, and GDP scenario data for the period 2021–2050 to derive future land use demands under different scenarios.

Compared with traditional models such as CA–Markov, CLUE-S, and FLUS, the Patch-generating Land Use Simulation (PLUS) model—enhanced with a Random Forest algorithm—demonstrates superior explanatory power and spatial accuracy in modeling land use drivers, making it highly applicable to ecological zoning in complex regions (Liang et al., 2021). The Patch-generating Land Use Simulation (PLUS) model consists of two integrated sub-modules. The first is the rule-mining module based on the Land Expansion Analysis Strategy (LEAS), which identifies the relationships between land use changes and driving factors. This module uses land use data from two time points and applies the Random Forest algorithm to extract the influence of each driving factor on land use transitions, thereby estimating the probability of expansion for different land uses. The second module is the Cellular Automata simulation module, driven by the Cellular Automata Random Seeds (CARS) algorithm. It integrates pixel-based neighborhood weights, land use transition matrices, and patch generation rules to simulate future land use patterns. This approach enables the spontaneous emergence and expansion of land use patches under spatiotemporal dynamics. This study incorporates land use demand predictions from the SD model and applies four categories of driving factors in the PLUS model: Natural environmental factors: elevation (DEM), slope, soil type, NDVI; Socio-economic factors: nighttime light (NTL), population (POP), GDP; Transportation factors: distances to residential areas, government seats, highways (national, provincial, expressway, county), railways, and subway/light rail lines; Climate factors: evapotranspiration, annual average precipitation, and temperature. These variables are used to spatially allocate land use demand under different future scenarios. Model accuracy was evaluated using the Kappa coefficient, Overall accuracy, and the Figure of Merit (FoM), while Moran’s I index was applied to assess the robustness of spatial predictions.

Throughout 2000, 2010, and 2020, forest remained the dominant land use, covering 63.79%, 62.99%, and 62.28% of the study area, respectively, followed by cultivated land and construction land (Figures 4a–c). Cultivated land, forest, and grassland showed gradual declines, with cultivated land experiencing the largest reduction (7.45%). Construction land expanded markedly (60.20%), primarily around core urban areas, reflecting rapid urbanization and infrastructure development. Grassland area decreased by 3.20%, indicating moderate disturbance to natural ecosystems. Although the overall changes in grassland and bare land were relatively limited, the proportion of bare land increased significantly by 42.72%, likely due to intensified construction activities and expanded engineering land use. From 2000 to 2020, substantial land use conversions occurred within the study area, with cultivated land experiencing the most pronounced transformation. The cumulative area converted from cultivated land reached 1,029.37 km², primarily into construction land (555.65 km²) and forest (394.97 km²). Forest also exhibited considerable conversion, with a cumulative loss of 890.88 km², mainly transformed into construction land (466.22 km²) and cultivated land (362.87 km²). Other land use categories underwent relatively minor conversions (less than 100 km²), exerting limited influence on the overall spatial pattern.

To further illustrate the dynamics of land use change, land use transition chord diagrams were generated for the periods 2000–2010 and 2010–2020 (Figures 4d–f). The results indicate that the total converted area of cultivated land in both phases remained nearly constant (712.36 km² and 711.22 km², respectively), suggesting that the intensity of cultivated land conversion has consistently remained at a high level over the past two decades. From 2000 to 2010, cultivated land was primarily converted to construction land (accounting for 47.80% of the total converted area), reflecting how rapid urbanization drove industrial development and sustained demand for construction land. From 2010 to 2020, the area of cultivated land converted to construction land decreased (19.29 km²), while the area converted to forest increased (39.42 km²). This shift may have been influenced by ecological civilization policies, under which some cultivated land was incorporated into ecological restoration and afforestation projects. At the same time, regional development entered a phase of structural optimization and stock adjustment, slowing the expansion of construction land. Overall, while the land use pattern characterized by “equal emphasis on development and ecology” has somewhat alleviated the pressure of cultivated land conversion into construction land, the overall trend of cultivated land reduction driven by urbanization remains significant. Consequently, the issue of land resource scarcity is becoming increasingly prominent.

Landscape Ecological Risk (LER) was classified into five levels using the natural breaks method (Table 4). By analyzing changes in the area of each risk level, the study examined the spatiotemporal evolution of LER under multi-factorial drivers and identified its spatial distribution patterns. Results show that LER was predominantly characterized by medium and medium-low risk levels. These two categories accounted for 76.97%, 74.40%, and 63.48% of the total area in 2000, 2010, and 2020, respectively, reflecting a consistent downward trend (a total decrease of 13.49%) and indicating a gradual increase in ecological risk. From 2000 to 2010 and from 2010 to 2020, the areas of low, medium-low, and medium risk declined by 784.24 km² and 3,266.91 km², respectively. Compared with the first period, the second period saw area reductions that were 3.18, 2.39, and 10.69 times larger for low, medium-low, and medium risk levels, respectively—indicating that medium risk areas experienced the most pronounced decline. Meanwhile, medium-high and high risk areas expanded continuously. Between 2010 and 2020, medium-high and high risk zones increased by 2,247.41 km² and 1,019.50 km², respectively—7.0 and 2.2 times the increase observed from 2000 to 2010. This highlights medium-high risk as the dominant expanding category, with ecological risk spreading across a wider area and ecosystem stability facing increasing threats. These findings underscore the need for enhanced ecological protection and restoration efforts in medium and high risk zones to improve regional ecological resilience.

In terms of spatial distribution (Figure 5), the mean LER values in 2000, 2010, and 2020 were 0.423, 0.431, and 0.453, respectively, showing a spatial pattern of “high in the center, low at the edges.” High risk zones were mainly concentrated in central urban areas (such as Furong District), typically in clustered or strip-shaped formations. These zones align with dense construction land and major transportation infrastructure, indicating strong impacts from human activities and economic development. By 2020, high risk zones had expanded outward, further intensifying the imbalance between urban growth and ecological protection. Medium-high and medium risk areas were mainly distributed in transitional zones between urban cores and peripheral counties, such as Yuhua District. The LER value in Yuhua increased from 0.525 in 2000 to 0.611 in 2020. During this period, industrial restructuring and infrastructure expansion led to rapid population growth and increased construction density, exacerbating ecosystem disturbance. Some medium risk zones consequently shifted to medium-high risk due to economic spillover effects. Medium-low and low-risk areas were largely found in outer counties such as Liuyang, Chaling, and Yanling, which possess strong ecological foundations. For example, Yanling County maintains 85.67% forest coverage, with well-preserved natural landforms and stable landscape structures, and is subject to minimal anthropogenic interference. This further confirms the critical role of ecological land (especially forest) in maintaining landscape stability and mitigating ecological risk.

The Ecosystem Health Index (EHI) was classified into five levels using the natural breaks method to identify its spatial distribution characteristics (Figure 6). Overall, ecosystem health exhibited a spatial pattern of “low in the center, high at the edges.” High and medium-high EHI values were mainly distributed in peripheral areas such as Liuyang, Yanling, and Chaling, where forest and grassland dominate, and ecosystem stability remains strong. Medium and medium-low levels were concentrated in the transitional zones between urban centers and outlying counties, accounting for 52.40%, 57.64%, and 63.68% of the area in 2000, 2010, and 2020, respectively. These areas are primarily composed of cultivated land, forming the ecological matrix of the region. However, they have become increasingly fragmented under urban expansion, leading to a continuous decline in ecosystem health. Low EHI values were mainly concentrated in the core urban districts, such as Furong District, where construction land dominates. Under the combined pressures of rapid economic growth and population concentration, these areas exhibited significantly lower EHI levels than surrounding regions. The mean EHI values in the study area declined from 0.555 in 2000 to 0.543 in 2010 and 0.518 in 2020. The 2.16% decrease from 2000 to 2010 may be attributed to the implementation of the CZXUA integration strategy in 1997, which accelerated economic development and construction land expansion, compressing ecological space despite limited ecological degradation at that stage. The subsequent 4.60% decline between 2010 and 2020 may relate to the accelerated urban integration process, the accumulation of prior ecological risks, and delayed ecological restoration measures.

From the perspective of land use, EHI generally exhibited a declining trend between 2000 and 2020 (Figure 7). The EHI of forest, water body, and grassland decreased by 6.29%, 4.21%, and 1.97%, respectively, with forest showing the most significant decline. This may be attributed to the combined effects of natural disasters (wildfires, pests, extreme weather, and soil degradation) and anthropogenic pressures (illegal logging and construction encroachment). The EHI of cultivated land increased initially—possibly due to the positive effects of ecological protection measures such as nature reserves, ecological redlines, and afforestation programs—but later declined as the long-term effectiveness of these measures diminished. The EHI of bare land first declined due to urban infrastructure expansion, then improved as a result of vegetation recovery through ecological restoration initiatives. Construction land consistently exhibited a sharp decline in EHI, with a total decrease of 22.15%. This was driven by the continuous expansion of built-up areas, the reduction of ecological land, and the resulting degradation of ecosystem services such as water retention, soil conservation, and habitat quality. Consequently, the stability and resilience of the ecosystem structure weakened, leading to a significant drop in overall ecosystem health.

Based on the standardized Z-scores of LER and EHI, ecological zones were delineated using zero as the classification threshold (Figures 8a–c). The study area was categorized into four functional zones: Ecological Conservation Zone (Low LER—High EHI), Ecological Control Zone (High LER—Low EHI), Strict Ecological Protection Zone (High LER—High EHI), and Ecological Enhancement Zone (Low LER—Low EHI) (Table 5).

In terms of overall spatial structure, the Ecological Conservation Zone accounted for 41.28%, 41.37%, and 41.38% of the total area in 2000, 2010, and 2020, respectively. This zone is primarily located in peripheral counties with rich ecological resources (e.g., Yanling and Chaling), dominated by forest and grassland. These areas experience minimal human disturbance, have high vegetation coverage, stable ecosystem structures, and provide significant ecological services. The overlapping areas and proportions with statutory protected areas were 2,641.97 km² (22.06%), 2,669.47 km² (22.63%), and 2,608.64 km² (22.10%), respectively. This indicates that the overlap between ecological conservation red lines and ecological conservation zones remained consistently high with minimal fluctuation, thereby supporting the stability of ecosystems and ecological service functions.

The Ecological Control Zone is the second largest, comprising 32.14%, 32.41%, and 32.36% of the area in 2000, 2010, and 2020, respectively. It is mainly distributed in the central and northern parts of the core urban areas and is characterized by construction and cultivated land. Under the pressure of urban expansion and population concentration, this zone experiences reduced ecological space, increased landscape fragmentation, and weakened ecosystem function, contributing to heightened ecological security risks. The overlapping areas and proportions with legally designated protected zones were 306.34 km² (3.34%), 292.99 km² (3.17%), and 295.30 km² (3.20%), respectively. The fluctuating proportions reflect urban expansion and land use changes within the ecological control zone, highlighting the need to strengthen coordination between ecological conservation and land use management.

The Strict Ecological Protection Zone showed a decreasing trend, accounting for 17.99%, 17.48%, and 17.44% of the area in 2000, 2010, and 2020, respectively. It is mainly located in the transitional zones surrounding the urban core and is dominated by cultivated land. The ecosystem in this region faces degradation risks, requiring particular attention to ecological fragility and fragmentation issues. The overlapping areas with legally designated protected areas were 330.39 km² (6.44%), 258.39 km² (5.18%), and 312.19 km² (6.28%), respectively. The notable fluctuations indicate that the extent of legally protected areas has been influenced by urbanization processes and development policies during specific periods. Accordingly, efforts should focus on restoring and safeguarding ecological functions within strictly protected areas.

The Ecological Enhancement Zone has the smallest proportion, covering 8.72%, 8.52%, and 8.69% of the total area in 2000, 2010, and 2020, respectively. It is mainly distributed in areas such as You County, dominated by a mix of forest and cultivated land. This zone lies between Ecological Conservation and Control zones and serves as a key regulatory area with potential for ecological recovery and functional restoration. The overlapping areas with legally designated protected zones were 159.04 km² (6.49%), 172.03 km² (6.89%), and 176.76 km² (7.03%), respectively. The continuous increase in these proportions demonstrates that the implementation of ecological conservation policies has strengthened ecological restoration potential and functions in this region.

From a temporal perspective (Figures 8d–f), notable spatial transitions occurred among the ecological zones. The Ecological Control Zone experienced the largest transition to the Strict Ecological Protection Zone, with a cumulative transfer of approximately 19,875 km². However, the rate of this transition slowed between 2010 and 2020, showing a 22.64% decrease compared to the 2000–2010 period. These transitions were concentrated at the edges of urban cores, where ecological risk control has been insufficient and EHI continues to decline, indicating limited effectiveness of ecological restoration. The second largest transition occurred from the Ecological Control Zone to the Ecological Enhancement Zone, totaling approximately 18,175 km². From 2010 to 2020, the transition volume decreased by 69.70% compared to the previous decade, primarily in non-core marginal areas. This suggests partial relief of landscape risks, but slow progress in ecological health restoration. The Ecological Conservation Zone showed a high transition frequency. Between 2010 and 2020, the volume of outflow transitions exceeded the inflow during 2000–2010 by 56.89%, indicating increasing instability in regional ecosystems. This trend underscores the urgent need to reinforce ecological redline protections and implement targeted restoration strategies to preserve ecological integrity.

Using land use data from 2000 to 2020 as a baseline, this study projected the area and proportion of each land use for the years 2030, 2040, and 2050 under three scenarios: SSP126, SSP245, and SSP585 (Table 6). In all scenarios, forest remains the dominant land use, accounting for approximately 60% of the total area, followed by cultivated land at around 28%. Construction land continues to expand across all scenarios, with average annual increases of 0.59% (SSP126), 0.30% (SSP245), and 0.40% (SSP585). Among them, the SSP585 scenario produces the largest area of construction land by 2050 (2,664.94 km²), reflecting the significant land pressure imposed by a high-emission, high-urbanization development trajectory. Although construction land grows most rapidly under the SSP126 scenario, its final projected extent is smaller than that under SSP585. This outcome may be attributed to the SSP126 pathway’s emphasis on ecological protection and low emissions, aiming to accommodate urban growth while safeguarding natural ecosystems. Cultivated land, grassland, and water body show a consistent declining trend across scenarios. This is primarily due to the encroachment of construction land, coupled with indirect effects of climate change and land degradation on grassland and aquatic ecosystems. Among all land uses, the decline in cultivated land is the most pronounced, with average annual decreases of 0.57% (SSP126), 0.10% (SSP245), and 0.25% (SSP585) projected for 2030, 2040, and 2050, respectively.

Spatial analysis of land use transitions from 2020 to 2050 (Figure 9) reveals that most land conversion under all scenarios involves transformation into construction land, with forest being the most affected. The expansion of construction land is concentrated on the urban periphery, typically along transportation corridors and infrastructure belts, forming an outward expansion pattern that reflects the spatial coupling of urban growth and transportation development. Compared to 2020, the net increase in construction land by 2050 is 850.98 km² (SSP126), 607.40 km² (SSP245), and 1,031.21 km² (SSP585). Concurrently, the forest area converted to other uses is 366.13 km² (SSP126), 438.95 km² (SSP245), and 671.45 km² (SSP585). These figures indicate that urban expansion poses a substantial threat to ecosystem integrity and reflects the intensifying impact of socio-economic development on natural landscapes. Specifically, under SSP126, cultivated land is converted to both forest and construction land, suggesting a balance between urban expansion and ecological restoration. Forest remains relatively stable within core urban areas but is more frequently converted in peripheral counties, primarily to construction and cultivated land. Under SSP245 and SSP585, cultivated land is predominantly converted to construction land, indicating that medium to high intensity urbanization is accelerating the loss of agricultural land. Forest is also increasingly converted to construction land, especially in the hilly eastern and southern regions, suggesting that urban construction pressure is encroaching upon ecologically sensitive zones and intensifying land use conflicts.

Based on land use simulation results for 2030, 2040, and 2050 under multiple scenarios, the LER and the EHI were recalculated to construct the spatial distribution of future ecological zoning (Figure 10). This study reveals the evolution characteristics of ecological spatial structures under different SSP–RCP scenarios. Compared to 2020, the Ecological Control Zone exhibits an overall expansion trend across all scenarios, with an average area increase of 4.90%. In contrast, the Ecological Enhancement Zone, Strict Ecological Protection Zone, and Ecological Conservation Zone show decreasing trends, with average reductions of 2.63%, 1.23%, and 1.04%, respectively.

Further analysis of the phased changes under each scenario reveals the following: The ecological control zone exhibits varying area changes across different scenarios. Under SSP126, it first contracts by 60.25 km² (2030–2040) and then expands by 32.50 km² (2040–2050). Under SSP245, it continuously contracts, decreasing by 38.50 km². Under SSP585, it shows a phased expansion pattern—first shrinking by 7.75 km², then increasing by 26.00 km². The relatively stable trend under SSP126 and SSP245 is linked to stricter ecological policies that constrain urban expansion. In contrast, the SSP585 scenario reflects higher urban development pressure and weaker ecological controls, leading to an increase in ecological risk and the expansion of the control zone. As the largest and most ecologically stable zone (>40%), the Ecological Conservation Zone serves as a natural ecological buffer dominated by forest. Under SSP126, it increases by 167.50 km² from 2030 to 2040, followed by a slight contraction of 8.75 km² from 2040 to 2050. In the SSP245 and SSP585 scenarios, the Ecological Conservation Zone continues to expand, with total increases of 175.50 km² and 166.50 km², respectively, suggesting that moderate to high development paths can still accommodate the preservation of core ecological areas under certain land-use controls. A strict Ecological Protection Zone between conservation areas and urban development is highly sensitive to ecological disturbance. The Strict Ecological Protection Zone decreases in all scenarios: by 0.21% under SSP126, 0.04% under SSP245, and 0.34% under SSP585, relative to 2020. The sharpest decline occurs in SSP585, where accelerated urban boundary expansion causes spatial compression, ecological degradation, and rising risk levels in already vulnerable zones. The Ecological Enhancement Zone is mainly found in fragmented patches along the edges of the conservation zone. In all scenarios, its area decreases steadily: by 0.09% (SSP126), 0.14% (SSP245), and 0.18% (SSP585). Due to its low ecosystem health status, the forest is more susceptible to conversion into construction land, resulting in limited ecological restoration potential and continuous spatial loss.

To further reveal spatial distribution differences in ecological zoning across multiple scenarios, this study analyzes the center of gravity distribution, migration trajectories, and hot and cold spot regions for the ecological control zone and ecological conservation zone (Figure 11). The center of gravity for the ecological control zone is located in Yutang District, Xiangtan City. Under SSP126 and SSP585 scenarios, the migration direction shifts from northwest to southeast, whereas under SSP245 it reverses from southeast to northwest, with migration speeds increasing across all scenarios. The standard deviation ellipse rotation angle across scenarios is approximately 150°, with the major axis about 1.4 times the minor axis, indicating pronounced northwest-to-southeast diffusion associated with urban economic gradient zones. Hotspot areas are concentrated in Changsha, Liuyang, and Chaling, with more extensive regional expansion under SSP585 than under SSP126. Cold spots are mainly located in You County, with regional changes under SSP585 showing weaker contraction compared to SSP126. These results suggest that the expansion of construction land and industrial agglomeration significantly affects the ecological conservation zone. The center of gravity for ecological conservation zones is located in Liling City. Under the SSP126 scenario, its migration direction is from northwest to southeast. Under the SSP245 and SSP585 scenarios, the migration direction shifts to northwest to southwest, with the latter showing a faster migration rate. The standard deviation ellipse rotation angle across scenarios is approximately 160°, with the major axis about 1.6 times the minor axis. This reflects pronounced northwest-to-southeast diffusion, consistent with regional mountain and river distributions. Hotspot areas are relatively dispersed across scenarios, though You, Yanling, and Chaling show persistent expansion. Cold spots are most evident in Liuyang and contract with increasing emission intensity. These findings indicate that ecological conservation zones are shaped by the combined influence of natural geography, human activities, and ecological policies. Therefore, analyzing the spatial response mechanisms of ecological zoning under different development pathways is essential for designing ecological zoning governance and protection policies, and for promoting more coordinated human–land relationships in the region.