Impacts of Future Land Use Change on Ecosystem Service Trade-Offs and Synergies in Water-Abundant Cities: A Case Study of Wuhan, China

Abstract

Under rapid urbanization, water-abundant cities face severe challenges of ecological space compression and ecosystem service (ES) degradation. This study focuses on Wuhan, a representative water-abundant city, integrating the PLUS model, InVEST model, correlation analysis, and geographically weighted regression (GWR) to simulate land use patterns in 2040 under three scenarios: natural development (ND), ecological protection (EP), and urban expansion (UE). We quantitatively assessed the spatiotemporal evolution of carbon storage (CS), water yield (WY), soil conservation (SC), and habitat quality (HQ), along with the trade-offs/synergies among these ES. The results reveal that the continuous expansion of construction land in Wuhan has extensively encroached upon cultivated land and water bodies. Although the woodland area increased, it was insufficient to offset the negative impacts of construction land expansion on ES. Under the ND scenario, ES declined by 1.89% to 5.33%. The EP scenario, by implementing ecological protection measures and restricting construction land expansion, enhanced ES by 1.4% to 10%. Conversely, the UE scenario saw construction land increase by over 60%, triggering a chain reaction of “urban expansion—reduction of cultivated land—encroachment on woodland/water bodies”, leading to a 4.77% to 10.75% decline in ES. Furthermore, this study uncovered complex interrelationships among ES: synergistic relationships generally prevailed among CS, SC, and HQ; trade-offs characterized the relationships between WY and both CS and HQ; and the relationship between WY and SC dynamically shifted between trade-off and synergy depending on land use patterns. Urban expansion (UE) intensified trade-off conflicts among ES, whereas ecological protection (EP) alleviated most trade-offs. However, water body expansion under EP weakened the synergy between CS and HQ due to the inherent characteristics of aquatic ecosystems (high HQ but low carbon sequestration). This research provides a scientific basis for water-abundant cities to coordinate development and ecological protection, informing the formulation of differentiated land use policies to optimize ES synergies.

1. Introduction

Ecosystem services (ES) represent the sum of various benefits that humans obtain directly or indirectly from natural ecosystems and constitute the foundation for sustaining life on Earth and human well-being [1,2,3]. However, land use/land cover change (LULCC), driven by socio-economic development, is reshaping the landscape pattern globally. This spatial reconfiguration fundamentally alters the capacity of ecosystems to provide critical services, and such alterations are often detrimental [4,5,6,7,8,9]. Moreover, unsustainable land use management, pursued solely with economic gain as a singular objective, is a primary driver leading to ES degradation, biodiversity loss, and numerous environmental issues [10,11,12]. Meanwhile, global population growth and rising consumption levels are intensifying the contradiction between ES supply and demand. Without effective land management strategies to buffer and guide development, the immense pressure of demand will lead to continued ecosystem degradation or even collapse [13,14]. Therefore, it is imperative to clarify the mechanisms by which land use change impacts regional ES, as this will help explore sustainable land use patterns.

ES do not exist in isolation; complex, nonlinear interactions often exist between different ES during their changes [15,16]. These interrelationships are generally categorized into two main types: trade-offs and synergies [3,17]. A trade-off refers to a relationship where an increase or improvement in one ES leads to a decrease or degradation in another ES. A synergy, conversely, describes a win–win situation where two or more ES increase or improve simultaneously [17]. In recent years, scholars have employed various methods to clarify the trade-off and synergy relationships between different ES across spatial and temporal dimensions. Multiple statistical analysis techniques, including Pearson correlation [18], Spearman correlation [19], Generalized Additive Models (GAMs) [20], Principal Component Analysis (PCA) [21], and Production-Possibility Frontiers (PPFs) [22], have proven to be effective tools for assessing these inter-ES relationships. However, the trade-off/synergy relationships among ES often vary with geographic location, topography, climate, or the intensity of human activities [23,24,25]. Traditional quantitative statistical methods struggle to reveal the spatial heterogeneity of these ES relationships. Recently, several researchers have adopted the geographically weighted regression (GWR) model to uncover this spatial heterogeneity in ES trade-offs/synergies [26,27,28]. Compared to other models, the GWR model generates local regression coefficients for each spatial unit. This capability visually represents the spatial variation in the correlation strength, providing crucial insights for guiding differentiated land use optimization layouts and formulating zoning-based control policies.

Land use change, as a direct manifestation of human activities, has become a key driver reshaping the supply of ES, directly altering the original structure and functioning of ecosystems [29,30,31,32]. Therefore, it is essential to simulate future land use based on different land use policies and analyze changes in ES trade-offs/synergies. Models such as the CA-Markov model [33], CLUE-S model [34], FLUS model [35], and OS-CA model [36] have proven to be effective tools for simulating future land use change. However, traditional land use simulation models, balancing computational efficiency with real-world complexity, struggle to capture the nonlinear interactions inherent in land use change and exhibit insufficient capability in representing spatial patterns. Recently, the Patch-generating Land Use Simulation (PLUS) model has gained widespread recognition among researchers [37,38]. The PLUS model comprises two core modules: a Land Expansion Analysis Strategy (LEAS) for rule mining and a Cellular Automata model based on Multi-type Random Seeds (CARS). Together, these modules address the challenge faced by traditional models in capturing complex nonlinear transitions [39]. Furthermore, the PLUS model supports customization of transition matrices, neighborhood weights, and constraint factors, enabling flexible formulation of different land use policy scenarios.

Water-abundant cities can be defined as a distinct urban typology characterized by significantly higher than average water coverage, where the aquatic network exerts a dominant influence on the urban ecological structure [40,41,42]. Although limited in number globally, these cities have unique ES functions and complex sustainable development challenges: while urban water bodies contribute to the environmental resilience of densely populated areas through ecological processes such as hydrological regulation, habitat maintenance, and climate buffering, these bodies also face ecological space compression and service function imbalances caused by urban expansion [43]. Taking China as an example, cities like Wuhan and Suzhou exhibit water coverage rates of 15–42%, forming a distinctive water–urban mosaic spatial pattern. However, rapid urbanization has led to continuous shrinkage of aquatic spaces, highlighting the intensified tensions between human development and ecological conservation in such regions [44,45]. Current research on water-abundant cities/regions primarily focuses on historical land use change impacts on ES [41,46]. The ES indicators investigated remain relatively limited, chiefly concentrating on hydrological metrics such as water yield (WY) and water purification (WP) [43,47]. In contrast, there is a notable scarcity of research exploring how future land use change—driven by different land use policies—will affect ES and their trade-off/synergy dynamics in water-rich urban/regional contexts. Critically, the strength and direction of ES interactions may undergo significant alterations over time and under land use change. This is particularly relevant in China, where water-abundant cities/regions often represent epicenters of human–land conflicts experiencing pronounced land use transformations [48]. Consequently, a deeper understanding of the mechanisms through which future land use change affects ES in water-abundant cities is paramount for achieving harmonious urban development alongside water resource conservation.

This study selects Wuhan City, the capital of Hubei Province, as the research area. As a representative water-abundant city in China, Wuhan exhibits water coverage exceeding 15%, forming a distinctive water–urban mosaic spatial pattern. Situated at the confluence of the Yangtze and Han Rivers, Wuhan is a key implementation area for national ecological strategies such as the “Beautiful Rivers and Lakes Protection and Construction Action Plan” and the “Yangtze River Conservation Initiative”. However, rapid population growth and urban expansion have led to shrinking ecological spaces and declining habitat quality in Wuhan. This study selected carbon storage (CS), water yield (WY), soil conservation (SC), and habitat quality (HQ) as the primary ES indicators. By integrating land use multi-scenario simulation (PLUS model), ecosystem service assessment (InVEST model), and trade-off/synergy analysis (Spearman correlation coefficient, GWR model), this study aims to achieve the following objectives: (1) Based on historical evolution trends and policy orientations, three typical scenarios—natural development (ND), ecological protection (EP), and urban expansion (UE)—were constructed to simulate the spatial pattern of land use in Wuhan by 2040, revealing the characteristics and spatial differentiation of land use changes under different policy directions. (2) The InVEST model was used to assess the spatiotemporal differentiation characteristics of key ecosystem services such as CS, WY, SC, and HQ under different scenarios, revealing their evolutionary trends and comparing service capacities across various scenarios. (3) By integrating correlation analysis and the GWR model, trade-offs and synergies among different environmental services were identified, clarifying their spatial heterogeneity and revealing the driving mechanisms of land use change on these relationships, thereby providing a basis for optimizing land use policies.

Therefore, this study takes Wuhan as an example to explore the ecological conservation issues in water-abundant cities during rapid urbanization. With a surface water coverage exceeding 15%, Wuhan features a dense river network and well-developed water systems, making it both a strategic area for national ecological conservation and a typical case reflecting the pronounced contradiction between rapid urbanization and water ecological protection. To address the gaps in existing research, this study constructs a coupled framework integrating the PLUS model (multi-scenario land use simulation), the InVEST model (ES assessment), and geographically weighted regression (GWR) to capture the nonlinear responses of ES to land use changes and their spatial heterogeneity. Based on this framework, three future scenarios—natural development (ND), ecological protection (EP), and urban expansion (UE)—are established to simulate divergent pathways of land use evolution and their cascading effects on ES bundles, thereby providing scientifically grounded and policy-forward insights. The findings are expected to offer scientific evidence and practical implications for water-abundant cities, contributing to optimized land use planning, mitigation of ecosystem service trade-offs, and the achievement of sustainable urban development.

2. Data and Methods

2.1. Study Area

Wuhan City (Figure 1) is situated in central China’s hinterland, eastern Hubei Province (113°41′–115°05′ E, 29°58′–31°22′ N). It lies at the confluence of the Yangtze and Han Rivers, positioned within the transition zone between the eastern fringe of the Jianghan Plain and the hilly southern foothills of the Dabie Mountains. The city covers a total land area of 8569.19 km². Wuhan City is endowed with extremely abundant water resources. It features 165 rivers longer than 5 km, 166 lakes listed in the protection directory, and 262 reservoirs. Since the beginning of the 21st century, Wuhan City has undergone rapid urbanization. By the end of 2020, the city’s permanent resident population reached 12.32 million, with an urbanization rate as high as 84.31%. The built-up area has expanded significantly due to the concentration of population. It grew from 210 km² in 2000 to 885.11 km² in 2020, which is an increase of approximately 4.21 times. The rapid expansion of non-ecological land use has led to a decrease in ecological land and increased ecological space fragmentation.

2.2. Data Source

The data used in this study are summarized in

Table 1. Data sources.

Category	Date Name	Year	Data Type	Source
Land Use Data	Land USE/Land Cover [49]	2010, 2020	Grid/30 m	https://doi.org/10.5281/zenodo.4417809 (accessed on 7 August 2024)
Natural environment data	DEM		Grid/30 m	https://www.gscloud.cn/
	Precipitation	2010, 2020	Grid/1 km	https://data.tpdc.ac.cn/
	Evapotranspiration	2010, 2020	Grid/1 km
	Temperature	2010, 2020	Grid/1 km
	Soil texture		Grid/250 m	https://www.isric.org/
	Soil depth [50]		Grid/100 m	http://globalchange.bnu.edu.cn/research/cdtb.jsp (accessed on 3 January 2020)
Socio-economic data	GDP	2020	Grid/1 km	https://www.resdc.cn/
	POP	2020	Grid/1 km	https://hub.worldpop.org/
Distance Accessibility Data	Railway, Road, River	2020	Shapefile	https://openstreetmap.org/

. Land use data were sourced from the publicly available China Land Cover Dataset (CLCD), comprising 30 m-resolution raster data for 2010 and 2020 [49]. The original CLCD categorizes land cover into 9 classes. For this study, we reclassified the land use types into 6 categories aligned with China’s prevailing land resource classification system and research objectives: cultivated land, woodland, grassland, water, construction land, and unused land. All data preprocessing was performed using ArcGIS Pro 3.1.6, including projection transformation, mask extraction, boundary clipping, and resampling, to ensure that each dataset was in the same coordinate system and had consistent row and column numbers.

2.3. Simulation of Future Land Use Patterns Under Different Scenarios

2.3.1. Scenario Settings

Based on the current development status of Wuhan City, and fully considering policy planning and future land use requirements, the socio-economic factors and natural environmental factors driving land use changes were incorporated into the PLUS model (Figure S1). Three land use scenarios—natural development (ND), ecological protection (EP), and urban expansion (UE)—were established to simulate land use in 2040. The transition rules for each scenario are shown in

Table 2. Transition matrix for different scenarios.

	NDS						EPS						UES
	1	2	3	4	5	6	1	2	3	4	5	6	1	2	3	4	5	6
1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	0	0	1	0
2	0	1	0	0	1	0	0	1	0	0	0	0	1	1	0	0	1	0
3	1	1	1	1	1	1	0	1	1	1	0	0	1	1	1	1	1	1
4	1	1	1	1	1	1	0	0	0	1	0	0	1	0	0	1	1	0
5	0	0	0	0	1	0	0	0	0	0	1	0	0	0	0	0	1	0
6	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1

1: cultivated land; 2: woodland; 3: grassland; 4: water; 5: construction land; 6: unused land.

.

Natural development (ND): This scenario assumes that future development patterns, speeds, and spatial preferences will be highly similar to those of the historical period (2010–2020), with no major new policy interventions or external disruptions altering the existing development trajectory. All land use transition rules are set based on historical trends.

Ecological protection (EP): This scenario prioritizes ecological and environmental protection. It is assumed that the government will implement stricter policies and measures to strengthen the protection of ecological land. In accordance with national ecological strategies such as the “Yangtze River Conservation Initiative” and the “Beautiful Rivers and Lakes Action Plan,” the regional characteristics of Yangtze River protection and lake governance are emphasized. Regarding transition rules, water bodies and woodlands are strictly protected, allowed to gain area but prohibited from loss, while also restricting the transition of other land use types into construction land.

Urban expansion (UE): Assuming that economic development and urban expansion are the primary objectives, land supply policies are relatively lenient to meet the rapidly growing demand for construction and population growth, thereby weakening ecological protection constraints. The National New-Type Urbanization Plan proposes to “orderly urbanization of rural migrants,” implicitly acknowledging the demand for construction land resulting from population concentration in cities. Under this scenario, all land use types are permitted to be transitioned into construction land.

2.3.2. PLUS Model

The Patch-generating Land Use Simulation (PLUS) model, developed by the HPSCIL Lab at the China University of Geosciences (Wuhan), is a land use/land cover change (LULCC) simulation tool based on Cellular Automata (CA). It offers advantages in patch-level dynamic simulation and multi-scenario flexibility [39]. The model is particularly suited to water-abundant cities characterized by intricate river–lake networks and fragmented ecological spaces. It effectively captures the complex, nonlinear transitional relationships between water bodies, built-up areas, and other ecological spaces. This advantage is essential for simulating urban expansion and ecological conservation processes within densely networked river–lake regions. The model primarily consists of the Land Expansion Analysis Strategy (LEAS) and the Multi-type CA with Random Patch Seeds (CARS). The LEAS module extracts historical land expansion patches and utilizes the Random Forest method to analyze the nonlinear relationships between expansion and various driving factors, generating growth probabilities for different land use types (Figure S2). The driving factors identified in this study include precipitation, temperature, elevation, slope, population density, GDP, distance to roads, distance to rivers, and distance to water bodies (Figure S1). Furthermore, the Kappa coefficient and overall accuracy, obtained by comparing the prediction results with real data, were used to assess the accuracy of the PLUS model.

2.4. Ecosystem Services Evaluations

2.4.1. Carbon Storage

Calculation was performed through the Carbon Storage and Sequestration module within the InVEST model. This module divides ecosystem carbon storage into four primary carbon pools: aboveground biomass, belowground biomass, soil organic carbon, and dead organic matter. Each land use/land cover (LULC) type is assigned corresponding carbon density values (t/ha) covering these four carbon pools. The input data include land use and the corresponding carbon density values:

$${\mathit{CS}}_{i }= C_{\mathit{\text{above-i}}} + C_{\mathit{\text{below-i}}} + C_{\mathit{\text{soil-i}}} + C_{\mathit{\text{dead-i}}}$$

(1)

$${\mathit{CS}}_{\mathit{total}} =\sum _{i=1}^6{\mathit{CS}}_i \times { \mathit{Area}}_i$$

(2)

In the equation, CS_i represents the carbon storage (t/ha) of land use type i. C_above-i, C_below-i, C_soil-i, and C_dead-i represent the carbon density (t/ha) of aboveground biomass, belowground biomass, soil organic carbon, and dead organic matter, respectively, for land use type i. Area_i represents the area of land use type i, and CS_total represents the total carbon storage. The carbon density data were primarily referenced from previous studies (Table S1) [51,52].

2.4.2. Water Yield

The calculation is performed by the water yield module within the InVEST model. This module is based on the water balance principle and the Budyko curve, assuming that the water yield for each grid cell is the residual portion of precipitation minus actual evapotranspiration. The input data required include precipitation, evapotranspiration, plant available water content, depth to bedrock, land use, and their corresponding biophysical coefficients:

$$W{Y}_{x,i} = \left(1-{\displaystyle \frac{\mathrm{AE}{\mathrm{T}}_{\mathrm{x},\mathrm{i}}}{{\mathrm{P}}_{\mathrm{x}}}}\right) \text{×}{ P}_x$$

(3)

In the equation, WY_x,i denotes the water yield (mm) of land use type i within grid x. AET_x,i denotes the actual evapotranspiration (mm) of land use type i in grid x. P_x denotes the precipitation (mm) within grid x. The biophysical table is shown in Table S2 [43].

2.4.3. Soil Conservation

Soil conservation refers to the capacity of ecosystems to mitigate soil loss induced by precipitation. It is typically quantified using the Revised Universal Soil Loss Equation (RUSLE), which calculates the difference between the potential soil erosion under bare soil conditions and the actual soil erosion. This difference indicates the soil conservation capacity of the ecosystem:

$$\mathit{RKL}{\mathrm{S}}_x = R_{x }\text{×}{ \mathrm{K}}_{x }\text{×} L{S}_x$$

(4)

$$\mathrm{USL}{E}_{x }={ R}_x \text{×}{ K}_x \text{×} L{S}_x \text{×}{ C}_{x }\text{×} P_x$$

(5)

$$\mathrm{S}{\mathrm{C}}_x=\mathit{RKL}{\mathrm{S}}_{x }-\mathit{USL}{E}_x$$

(6)

In the equation, RKLS_x denotes the potential soil erosion (t/ha) in grid x. USLE_x denotes the actual soil erosion (t/ha) in grid x. SC_x denotes the soil conservation (t/ha) in grid x. R_x is the rainfall erosivity factor (MJ mm/ha/h). K_x is the soil erodibility factor (t h/ha/MJ/mm). LS_x is the slope length and steepness factor (dimensionless). C is the vegetation cover and management factor (dimensionless), reflecting the mitigating effect of surface vegetation and its residues on soil erosion under specific vegetation types or land cover conditions, ranging from 0 (complete erosion suppression) to 1 (no suppression). P is the support practice factor (dimensionless). It reflects the erosion reduction effect of different management practices, ranging from 0 (complete control) to 1 (no practices). The parameters for C and P are shown in Table S3 [53].

2.4.4. Habitat Quality

Habitat quality reflects the capacity of an ecosystem to provide suitable conditions for the survival and reproduction of individual organisms or populations, and it is highly sensitive to anthropogenic disturbance. This study utilizes the habitat quality module within the InVEST model for assessment. This module employs a spatially explicit “threat-distance-sensitivity” framework, simplifying complex ecological processes into quantifiable parameters. The assessment results are expressed as a continuous value between 0 and 1. Input data include land use, a habitat type sensitivity table, threat sources, and their corresponding model parameters:

$$Q_{x,i} = H_i \text{×} (1 - {\displaystyle \frac{D_{x,i}^z}{D_{x,i}^z+K^z}})$$

(7)

$$D_{x,i }=\sum _{r=1}^R\sum _{y=1}^{Y_r}\left(\frac{w_r}{{\sum }_{r=1}^R{w}_r}\right) {\mathrm{R}}_y{ \text{×} i}_{r,x,y}{ \text{×} \beta }_x{ \text{×} S}_{i,r}$$

(8)

In the equation, Q_x,i represents the habitat quality of land use type i in grid x. D_x,i is the habitat degradation index. H_i is the habitat suitability of land use type i. K is the half-saturation constant, generally set to half of the maximum habitat degradation value. In this study, K was set to 0.5. z is the normalizing constant, with a default value of 2.5. R is the total number of threat factors; Y_r is the total number of grids for threat factor r. W_r is the weight of threat factor r. R_y denotes the intensity value of threat factor r at its source grid y. i_r,x,y denotes the distance–decay influence function representing the impact of threat factor r from its source grid y to the target grid x. β_x is the accessibility level of grid x. S_i,r is the sensitivity of land use type i to threat factor r. The sensitivity of habitat types to threat factors and the parameters of threat factors are shown in Tables S4 and S5.

2.5. Spearman Correlation

ES are often not independent but exhibit highly complex and tightly interwoven relationships. This study employs Spearman’s correlation coefficient to quantitatively analyze the trade-offs/synergies among four ES under different land use scenarios. First, a 1 km fishnet grid covering the study area was created in ArcGIS Pro, and the average values of the four ES within each grid cell were calculated. Subsequently, Spearman’s correlation coefficients between the ES were computed using the Corrplot package in R 4.4.3. The coefficient ranges from −1 to 1, with positive values indicating synergies and negative values indicating trade-offs:

$$\rho = {\displaystyle \frac{{\sum }_{i=1}^n\left(x_i-\overline{x}\right)\left(y_i-\overline{y}\right)}{\sqrt{{{\sum }_{i=1}^n\left({\mathrm{x}}_i-\overline{x}\right)}^2{{\sum }_{i=1}^n\left(y_i-\overline{y}\right)}^2}}}$$

(9)

In the equation, ρ is the correlation coefficient; n is the total number of grid cells in the study area; x_i and y_i are the values of two different ES corresponding to the grid i; $\overline{x}$ and $\overline{y}$ are the mean values of the two ES.

2.6. Geographically Weighted Regression

The trade-offs and synergies between ES often vary geographically, exhibiting distinct spatial characteristics. However, Spearman’s correlation coefficient can only reflect their global, overall relationship. Therefore, this study employs a geographically weighted regression (GWR) model at the township level to capture the spatial variability of ES trade-offs/synergies. GWR generates unique regression coefficients by constructing local regression models for each geographic unit. This approach enables the precise identification of localized spatial variation patterns in ES relationships, avoiding misleading interpretations that may arise from a “global average” Similar to the Spearman correlation coefficient, the sign of the coefficient indicates whether the relationship between ES is trade-off (negative values) or synergy (positive values):

$$y_i = {\beta }_0 (u_i, v_i) + \sum _{k=1}^m{\beta }_k (u_i, v_i) x_{i,k} + {\epsilon }_i$$

(10)

In the equation, y_i is the dependent variable value at observation point i. (u_i, v_i) are the spatial coordinates of observation point i. β₀ is the intercept coefficient. β_k is the regression coefficient. m represents the total number of independent variables. x_i,k is the k-th independent variable at observation point i. ε_i is the random error term.

3. Results

3.1. Spatial–Temporal Changes of Land Use in Wuhan City

Based on CLCD land use data (Figure 2;

Table 3. The area of land use types under different scenarios from 2010 to 2040 (km²).

Land Use Types	2010	2020	NDS	EPS	UES
Cultivated land	5930.02	5550.83	4900.05	4867.78	4996.75
Woodland	517.6	657.27	872.52	1014.15	635.91
Grassland	3.96	0.90	0.60	2.87	0.38
Water	1264.41	1183.27	1048.09	1502.09	1046.24
Construction land	858.87	1182.47	1753.55	1186.71	1895.71
Unused land	0.26	0.38	0.32	1.53	0.14

), cultivated land constitutes the largest proportion (69.15%) of all land use types in Wuhan City, followed by water bodies (14.74%). Cultivated land is distributed in concentric rings around the central urban built-up area. Water bodies are more dispersed, including the main stems of the Yangtze River and the Han River in the central region, as well as large lakes in the southern and eastern areas, forming a distinctive river–lake network. Construction land accounts for 10.01%, concentrated in the central region and radiating outward. Woodland accounts for 6.03%, primarily distributed in the mountainous and hilly areas of the north. From 2010 to 2020, cultivated land decreased by 379.18 km², while water bodies decreased by 81.14 km² (Figure 3; Table 3). Woodland increased by 139.67 km², and construction land increased by 323.6 km². These changes are primarily driven by two competing processes: ecological restoration and urban expansion. The Grain-for-Green Program accounted for 28.44% of the lost cultivated land area, enhancing regional soil conservation and carbon storage capabilities. In contrast, 50.32% of the lost cultivated land area was converted to construction land, reflecting the pressure of rapid urbanization radiating from urban centers. Meanwhile, 91.78% of the lost water area was attributed to agricultural reclamation, indicating that although urbanization is more visibly prominent, agricultural expansion remains an underlying driver of aquatic ecosystem loss. Overall, Wuhan is characterized by a conflicting scenario where ecological restoration coexists with urban expansion, while water resources face dual pressures from both urban expansion and agricultural activities.

This study validated the accuracy of the PLUS model using the Kappa coefficient. The validation results demonstrated high accuracy, with a Kappa coefficient of 0.881 and an overall accuracy of 0.939. Based on the simulation results, significant changes in the land use pattern of Wuhan City are projected between 2020 and 2040 (Figure 2 and Figure 3; Table 3). Under the ND scenario, which continues the historical development trajectory, cultivated land and water bodies are projected to decrease by 650.74 km² and 135.17 km², respectively. Conversely, woodland and construction land are expected to increase by 215.24 km² and 571.07 km², respectively. Woodland, primarily located in the northern mountainous areas, is projected to expand without conflicting with construction land expansion. Water bodies, however, are more widely distributed and face direct threats from both urban expansion and agricultural activities. Under the EP scenario, which implements strict ecological conservation policies, woodland and water bodies are projected to increase by 356.87 km² and 318.82 km², respectively. Cultivated land is expected to decrease by 683.04 km², while construction land area remained virtually unchanged, maintaining the status quo. Under the UE scenario, where urban expansion is prioritized, construction land is projected to increase by 60% (713.23 km²). A significant proportion (93.4%) of this new construction land is converted from cultivated land. Simultaneously, forest land and water bodies are projected to decrease substantially. While some areas are directly converted to construction land, the remainder is converted to cultivated land. This large-scale urban expansion not only directly consumes cultivated land but also indirectly forces agricultural activities to expand into ecological spaces such as forest land and water bodies, intensifying the encroachment of agricultural space upon ecological space.

3.2. Spatial–Temporal Variation of ES in Wuhan City

3.2.1. Carbon Storage

Areas with high CS are primarily distributed in the northern forested regions, while low-CS areas are concentrated in the central urban built-up zones (Figure 4a). Furthermore, these low-CS areas exhibit a pattern of outward diffusion from the center, corresponding with urban expansion. Between 2010 and 2020, Wuhan’s total CS showed a downward trend, decreasing from 8.38 × 10⁷ t to 8.30 × 10⁷ t, with the mean value also dropping from 97.68 t/ha to 96.79 t/ha (Figure 4b). Projections for 2020–2040 indicate that under the ND and UE scenarios, continued built-up area expansion will lead to a further decline in total CS, reducing it to 8.14 × 10⁷ t and 7.9 × 10⁷ t, respectively. The conversion of vegetation-covered land use types to impervious surfaces not only directly destroys carbon pools but also results in the permanent loss of carbon sink functions. Conversely, under the EP scenario, total CS is projected to increase to 8.42 × 10⁷ t, benefiting from the expansion of forest areas and restrictions on construction land development.

3.2.2. Water Yield

Precipitation exerts a significant influence on the spatial distribution of WY. In 2010, high-WY areas shifted eastward, with low-WY zones concentrated in the northern and southwestern regions (Figure 4c). By 2020, however, low-WY areas had shifted to the eastern and northern parts, while high-WY areas migrated westward. In terms of quantitative changes, both total WY and mean values remained relatively stable between 2010 and 2020 (Figure 4d). Projections for 2020–2040 indicate that under the ND and UE scenarios, reduced water areas and continued construction land expansion will increase surface runoff, driving total WY up to 64.64 × 10⁸ m³ and 66.16 × 10⁸ m³, respectively, with mean values rising to 753.85 mm and 771.58 mm. This increase is not a reflection of improved water resource sustainability, but rather a hydrological response triggered by surface hardening. Impervious surfaces prevent infiltration, drastically reducing groundwater recharge and amplifying rapid, polluting surface runoff, which exacerbates flood risks and water pollution. Conversely, under the EP scenario, the combined effects of expanded forest coverage (which reduces runoff and increases evapotranspiration) and restored regulatory and storage functions of aquatic ecosystems will decrease total WY and mean values to 59.4 × 10⁸ m³ and 692.73 mm, respectively.

3.2.3. Soil Conservation

The spatial distribution pattern of soil SC remained relatively stable across different years (Figure 4e). High-SC areas were consistently distributed in the northern mountainous and hilly regions, while low-SC areas were concentrated in the alluvial plains along river valleys. From 2010 to 2020, total SC decreased from 13.81 × 10⁷ t to 13.36 × 10⁷ t (Figure 4f). Projections for 2020–2040 indicate that under both ND and UE scenarios, surface sealing from construction land expansion disrupts sediment transport pathways, reducing actual soil erosion and consequently increasing total SC to 13.83 × 10⁷ t and 13.66 × 10⁷ t, respectively. Conversely, under the EP scenario, SC rises to 13.88 × 10⁷ t due to ecological protection initiatives.

3.2.4. Habitat Quality

The spatial distribution of HQ closely corresponds to the urban center–fringe–rural structure (Figure 4g), exhibiting a distinct gradient: low in urban centers, medium in transitional fringes, and high in rural areas. From 2010 to 2020, construction land expansion reduced the mean HQ from 0.48 to 0.46, while forest expansion increased high-HQ areas (0.7) to 1.82 × 10³ km² (Figure 4h). Projections for 2020–2040 indicate that under the ND scenario, mean HQ declines to 0.44 while high-HQ areas rise to 1.83 × 10³ km². Under the UE scenario, ecological space contracts significantly, reducing mean HQ to 0.41 and high-HQ areas to 1.58 × 10³ km². Conversely, the EP scenario increases mean HQ to 0.51 and expands high-HQ areas to 2.50 × 10³ km².

3.3. Trade-Offs and Synergies Among ES

3.3.1. Correlation Analysis

As shown in Figure 5, the Spearman correlation coefficients among ES in Wuhan City reveal distinct interrelationships. Specifically, CS-HQ, CS-SC, and SC-HQ exhibit synergistic relationships, while CS-WY and WY-HQ demonstrate trade-offs. The correlation direction of WY-SC varies across different land use patterns. The coexistence of ecological restoration and urban expansion has resulted in complex dynamics in ES trade-offs and synergies. By 2020, the trade-off between WY and HQ had weakened, whereas the CS-WY trade-off intensified. Synergies between CS-HQ and SC-HQ strengthened, while the CS-SC synergy diminished. Notably, the WY-SC relationship transitioned to synergy.

Across the three future scenarios, the correlation directions among ES pairs remained consistent except for WY-SC. Specifically, under the ND and UE scenarios, the trade-off effects between CS-WY and WY-HQ intensified significantly due to the large-scale expansion of construction land. Conversely, in the EP scenario, implementation of ecological protection measures effectively mitigated trade-offs between these ES pairs, reducing their trade-off intensity to the lowest level among all scenarios. Additionally, synergistic effects of CS-SC and SC-HQ in the EP scenario were substantially stronger than those in the ND and UE scenarios. The correlation direction of WY-SC differed across scenarios: impervious surfaces from construction land reduce soil erosion but increase surface runoff. Consequently, WY-SC exhibited a trade-off in the EP scenario (where construction land expansion was restricted), but showed synergy under ND and UE scenarios. Notably, compared to other scenarios, the CS-HQ synergy was weakest in the EP scenario. This is primarily because aquatic ecosystems exhibit less efficient carbon sequestration (CS). Therefore, in the EP scenario, the policy-driven aquatic expansion improved habitat quality but diluted regional carbon density, weakening the CS-HQ synergy effect. In contrast, under the ND and UE scenarios, the reduction in water area amplified the contribution of forest land to the CS-HQ synergy effect.

3.3.2. Spatiotemporal Differences in ES Trade-Offs and Synergies

Given that the Spearman coefficient only reflects the overall correlation between ES and struggles to reveal their spatial heterogeneity characteristics, this study employs a GWR model to illustrate the spatially varying patterns of ES correlations at the township scale (Figure 6). In 2010, most townships in CS-HQ exhibited spatial synergies, with only sporadic townships in the south and east showing spatial trade-offs. A high proportion of townships demonstrated spatial synergies for the CS-SC (75%) and SC-HQ (64%). These synergistic relationships were predominantly concentrated in rural townships within the study area. In contrast, townships exhibiting spatial trade-offs were primarily distributed along the transition zone from the urban center to the fringe. For WY-SC, the proportion of townships showing trade-offs (64%) was significantly higher than that for townships showing synergies (36%). Synergies were mainly found in the central urban area and adjacent fringe areas, while trade-offs prevailed elsewhere. The CS-WY and WY-HQ pairings were dominated by spatial trade-offs, with areas of high trade-off intensity significantly clustered around the urban center. By 2020, the intensity of spatial relationships between the various ES pairs had changed significantly (Figure 7). Specifically, the synergy of CS-HQ strengthened markedly, with 73% of townships showing only increased synergy intensity.

A total of 18% showed a decrease, and an additional 7% transitioned from trade-off to synergy. Trade-offs for CS-WY and CS-SC generally weakened (85% and 54% of townships, respectively). The trade-off for WY-HQ intensified significantly (92% of townships). Conflict for WY-SC overall intensified, with 55% of townships experiencing a decrease in synergy intensity, 23% seeing an increase in trade-off intensity, and 4% shifting from synergy to trade-off. SC-HQ remained dominated by trade-offs, with 36% of townships showing increased trade-off intensity, and only 12% showing increased synergy intensity. Spatially, the overall pattern of ES relationships remained relatively stable between 2010 and 2020; the increased synergy intensity of CS-HQ led most originally trade-off townships to become synergistic, causing the spatial trade-off zone to contract and persist mainly near the southeastern administrative boundary, while the intensified trade-off intensity of WY-HQ prompted the northward expansion of the high trade-off area.

Under future land use scenarios, the spatial distribution patterns of ES trade-offs and synergies remain generally consistent with 2020, but their relationship intensities show significant differentiation. Under the ND scenario, 82% of townships experienced increased synergy intensity for CS-HQ. For other ES pairs, the proportion of townships with increased trade-off intensity significantly exceeded historical 2010–2020 levels: CS-WY (55%), CS-SC (13%), WY-HQ (95%), WY-SC (26%), and SC-HQ (46%). The EP scenario featured coexisting enhanced synergies and residual trade-offs. CS-SC and WY-SC reached peak scenario values for townships with increased synergy intensity (31% and 33%, respectively), while SC-HQ showed a higher proportion of townships with reduced trade-off intensity (31%) than other scenarios. Additionally, 3% of townships achieved a transition from trade-off to synergy for CS-HQ. Conversely, the proportion of townships with increased CS-HQ synergy intensity (52%) was relatively weaker than in other scenarios. Conflict intensified further among ES pairs under the UE scenario. Townships with increased WY-HQ trade-off intensity reached 95%, driving the outward diffusion of high-trade-off zones from the urban center. The proportions of townships with decreased CS-SC, WY-SC, and SC-HQ synergy intensity (19%, 47%, and 28%) and increased trade-off intensity (13%, 22%, and 49%) were all higher than under other scenarios.

4. Discussion

4.1. Impact of Land Use Change on ES

Land use change significantly impacts ecological processes, including energy exchange, water cycling, soil erosion and sedimentation, and biogeochemical cycles within ecosystems [3,31]. These changes ultimately affect the capacity of ecosystems to provide services. This study identifies woodland and water bodies as the primary components of ecological land in the research area. Both provide significantly higher levels of multiple ES compared to other land use types (Table S6). In contrast, construction land exhibits a higher level only in WY. Surface sealing associated with construction land effectively prevents soil exposure, reduces soil and water loss, and inhibits water infiltration, thereby significantly increasing surface runoff while diminishing rainfall erosion [54]. However, surface sealing replaces original vegetation or water cover, directly removing the biomass carbon pool and leading to a decline in carbon storage (CS). Furthermore, the patchy or corridor-like expansion of construction land fragments previously continuous natural habitat patches, reducing landscape connectivity—a key spatial mechanism for the decline in habitat quality (HQ) [55]. Woodlands, due to their high vegetation coverage and deep root systems, play a vital role in ecosystems as stable carbon sinks, resource reservoirs, and water conservators. They hold extremely high value for carbon sequestration, oxygen release, soil conservation, and biodiversity protection [56]. Nevertheless, high vegetation coverage also implies high water demand and high evapotranspiration, resulting in lower water yield (WY) [57,58]. In the study area, although water bodies accounted for only 15% of the total area, they contributed over 64% of the high-habitat-quality zones. Due to their unique physical morphology and connectivity, water bodies provide critical habitats for aquatic and water-dependent species and significantly enhance regional landscape connectivity. However, compared to ecosystems with vegetation cover, water bodies themselves have limited carbon sequestration capacity, resulting in relatively lower carbon storage (CS) per unit area [59,60].

The implementation of different land use policies will lead to variations in the rate and direction of ecosystem service changes. From 2010 to 2020, construction land increased by 37.6%, while cultivated land and water bodies decreased by 6.39% and 6.42%, respectively. A striking 98.26% of the newly added construction land was converted from cultivated land and woodland, resulting in a 0.91% decrease in total CS and a 2.97% decline in the mean HQ (Figure 4; Table S7). Although the total woodland area increased by 26.98% during this period, its concentration primarily in the northern mountainous areas failed to offset the encroachment effect of construction land expansion on ecological spaces surrounding the city, leading to a spatial mismatch between ecological restoration and urbanization. Similarly, under the ND scenario, which continued the historical trend, total CS and mean HQ decreased by 1.89% and 5.33%, respectively, while total WY increased by 4.29% compared to 2020 levels. These results align with previous studies [41,61]. In contrast, the EP scenario strictly protected ecological land and restricted the disorderly expansion of construction land. This policy achieved substantial increases in woodland (up 54.3%) and water bodies (up 26.94%). Compared to 2020, total CS increased by 1.4%, mean HQ rose by 10%, and total WY decreased by 4.16%. The decrease in WY was primarily due to enhanced infiltration and evapotranspiration within woodland and aquatic ecosystems. This reduction essentially reflects strengthened hydrological regulation functions, which promote water quality improvement and reduce flood risks [62,63]. Under the UE scenario, construction land surged by 60%, causing severe degradation of ecosystem services. Compared to 2020, total CS and mean HQ plummeted by 4.77% and 10.75%, respectively. Notably, total SC increased under both the ND and UE scenarios by 3.54% and 2.33%. This "increase" primarily stemmed from the interruption of sediment transport pathways due to construction land expansion, leading to a reduction in actual erosion. However, this type of "improvement" is accompanied by the fragmentation of ecological spaces and significant alterations to natural processes. From the perspective of overall ecosystem function and sustainability, this outcome is undesirable. In comparison, the EP scenario achieved a 3.91% increase in total SC through genuine ecological restoration measures (such as increasing vegetation cover and improving soil structure), representing a more ecologically sustainable pathway.

4.2. Impact of Land Use Change on ES Trade-Offs and Synergies

In Wuhan City, WY exhibits clear trade-offs with CS and HQ, while its relationship with SC shows dynamic trade-offs or synergies depending on the specific land use pattern. Synergistic effects were observed among the other ES pairs. The interrelationships between ES are primarily influenced by topography and land use. Multiple studies indicate that low temperatures in high-altitude areas may reduce vegetation cover, thereby lowering HQ, while steep slopes in mountainous regions increase the risk of soil erosion, reducing SC [64,65]. However, in the mountainous and hilly areas of Wuhan City, the topographic relief acts as a barrier against anthropogenic disturbance. Consequently, multiple ES exhibit high values in these regions. Furthermore, woodland, as the dominant land use type in mountainous areas, can simultaneously provide multiple ES, including CS, HQ, and SC [66]. The integrated ecological functions of woodland result in significant synergies among CS, SC, and HQ. Notably, compared to woodland, water bodies exhibit similarly high habitat quality but relatively weaker carbon sequestration capacity [59]. According to the GWR model analysis, a trade-off between CS and HQ emerges in some lake-concentrated areas of southern Wuhan. Woodland reduces runoff due to its high evapotranspiration characteristics, while construction land increases runoff through surface sealing, but at the cost of removing the biomass carbon pool and fragmenting habitat patches [67]. Consequently, WY shows clear trade-offs with both CS and HQ. A possible explanation for the dynamic trade-offs/synergies observed between WY and SC is that construction land expansion significantly increases surface runoff while simultaneously blocking sediment transport pathways. This can lead to a statistically “synergistic” increase in both WY and SC under specific scenarios, particularly during periods of substantial construction land expansion.

This study found that land use policies significantly influence the intensity of trade-offs and synergies among ecosystem services. Under the EP scenario, the intensity of trade-offs between all ES pairs was alleviated, consistent with some previous research findings [54,68]. Ecological restoration measures promoted the dominance of woodland and water bodies. The expansion of water bodies restored the flood regulation and storage functions of lake groups, alleviating urban waterlogging. Increased vegetation coverage helped enhance carbon sequestration potential and reduce soil and water loss caused by rainfall [44,67]. However, restricting construction land expansion and reducing anthropogenic disturbance, while mitigating water pollution risks to some extent, also reduced surface runoff. Furthermore, due to the weak carbon sequestration capacity of water bodies, their expansion, while improving HQ, diluted the synergy between CS and HQ [59]. The EP scenario achieved controllable trade-offs through the implementation of ecological projects, but it also introduced a substitution trade-off between CS and HQ due to the low carbon sequestration characteristics of water bodies. Under the UE scenario, urban expansion extensively encroached on cultivated land. More critically, large-scale urban expansion not only directly converted cultivated land but also indirectly forced agricultural activities to expand into surrounding ecological spaces like woodland and water bodies, forming a chain reaction of “urban expansion—reduction of cultivated land—encroachment on ecological land”. This ultimately led to intensified trade-off conflicts and a significant decline in key ecosystem services. The ND scenario, continuing historical trends, incorporated ecological protection projects alongside urban development. However, due to the spatial separation between woodland and built-up areas, and the squeezing of water body space by agricultural activities, these protective measures failed to counterbalance the negative impacts of urban expansion on ecosystem services. Overall, the ND scenario exhibited an intensification of trade-offs under historical inertia.

4.3. Policy Implications

Based on the findings of this study, which simulates the evolution of ecosystem services and their trade-off/synergy relationships under different land use scenarios, the ecological challenges and opportunities faced by Wuhan during rapid urbanization are revealed.

Drawing from simulation results under EP, ND, and UE scenarios, and considering Wuhan’s distinctive spatial patterns—”exceptionally abundant water resources under dual pressure from urban and agricultural encroachment,” “mountainous forest ecological barriers,” and “central high-intensity development zones”—the following targeted policy recommendations are proposed to address future uncertainties: For core water bodies and riparian zones, referring to the planning principles of the EP scenario, clearly delineate and strictly enforce protection boundaries for all 166 protected lakes along with their surrounding ecological buffer zones. Rigorously restrict urban expansion and agricultural activities from encroaching upon lakes, rivers, and wetlands, prioritizing the ecological conversion direction of water bodies. Enhance ecological connectivity between isolated lakes and rivers to establish a continuous, integrated ecological network. Additionally, addressing the potential trade-off between CS and HQ, as demonstrated in the EP scenario, requires ecological restoration measures that extend beyond the water bodies themselves. Increasing riparian vegetation coverage can enhance carbon sequestration capacity in waterfront areas while preserving aquatic habitat quality. In mountainous forest ecological barrier areas, referring to the ND and EP scenario planning, implement forest closure for natural regeneration and natural forest protection policies, prohibiting or restricting the conversion of woodlands to construction or cultivated land, minimizing human disturbance to habitats, and strengthening ecosystem service provision, to prevent the degradation observed in the UE scenario. Furthermore, addressing the trade-off between WY and other ES requires leveraging the inherent characteristics of woodlands to function alongside water bodies as “urban sponges,” achieving water purification, flood risk reduction, and urban heat island mitigation. Within central urban built-up and expansion areas, abandon the extensive expansion path typified by the “UE scenario.” Control the total volume and pace of new construction land, particularly its sprawl into vital ecological spaces. Strictly enforce land spatial zoning management by clearly demarcating ecological conservation redlines, permanent basic farmland protection redlines, and urban development boundaries. To counter ecosystem degradation and intense trade-offs triggered by impervious surfaces, promote high-density, mixed-use, intensive development and increase the proportion of internal blue-green spaces. In plain agricultural production areas, strictly protect high-quality cultivated land, prohibiting its disorderly conversion to construction land while simultaneously rigorously controlling non-ecological conversion of farmland to water bodies or woodlands. Promote water-saving, fertilizer-saving, and pesticide-saving green agricultural models to reduce agricultural non-point source pollution threats to water bodies and enhance the ecosystem services provided by the farmland itself.

The significantly different outcomes under the three scenarios highlight the importance of adaptive management. Land use planning should not remain static; instead, it requires regular monitoring of ecosystem services and their interactions, utilizing the latest data to adjust policies and ensure their continued effectiveness amid evolving socio-economic and environmental conditions. Furthermore, the Wuhan-based study provides a replicable model for ecosystem service management in similar water-abundant cities. In such cities, where water bodies dominate as the primary type of ecological space, they often occupy a disadvantaged position in urban development. The dense river–lake networks also tend to lead to the fragmentation of ecological spaces. Therefore, in planning practices, it is essential to treat water bodies and their riparian zones as core ecological areas, enhancing their connectivity and functional integrity through rigid protection red lines and the construction of ecological corridors. In summary, Wuhan’s experience demonstrates that achieving sustainable development in water-abundant cities involves more than just protecting individual ecological elements. Future land use management policies must move beyond single-objective thinking and deeply recognize the complex interactions among different ecosystem services and across different spatial units.

4.4. Limitations and Prospects

This study integrates multi-source data and models to assess the impact of land use change on ES and their trade-offs/synergies. The findings can provide references for the implementation of land use management strategies and ecological restoration projects in Wuhan City. However, several limitations remain that need to be addressed in future research. First, this study did not incorporate the potential impact of climate change on ecosystem services. As a typical humid region in the middle reaches of the Yangtze River, precipitation and temperature changes in Wuhan may significantly alter the spatial distribution of WY and CS. Because the models were not coupled with climate scenarios, the current results may overestimate the sustainability of the EP scenario or underestimate the ecological risks of the UE scenario. Second, the mechanisms underlying the trade-offs and synergies between ESs are complex; although they can be attributed to land use change and internal ES interactions, their quantitative explanation remains insufficient. Additionally, parameters for the InVEST model were referenced from previous empirical studies and lacked sufficient localized validation, which may affect the accuracy of the quantitative results. Therefore, subsequent research needs to overcome these limitations to bring the findings closer to the actual situation.

5. Conclusions

This study utilized the PLUS model to simulate Wuhan’s land use patterns in 2040 under different development scenarios and quantitatively assessed the spatiotemporal evolution and trade-off/synergy effects of ecosystem services (ES), including carbon storage (CS), water yield (WY), soil conservation (SC), and habitat quality (HQ) based on the InVEST model. The main conclusions are as follows:

(1)

Rapid urbanization in Wuhan has led to significant changes in land use patterns. Historical trends from 2010 to 2020 indicate a significant expansion of construction land, primarily encroaching upon cultivated land and water bodies. Although the woodland area increased, spatial mismatch prevented it from effectively offsetting the negative impacts of urban expansion on ecological spaces. Under the urban expansion (UE) scenario, construction land not only directly converted cultivated land but also triggered a “urban expansion—reduction of cultivated land—encroachment on woodland/water bodies” chain reaction, intensifying the compression and fragmentation of ecological space. Conversely, under the ecological protection (EP) scenario, strict ecological protection policies effectively promoted the restoration of ecological spaces.

(2)

Land use change dominates the evolution of ecosystem services, with significant differences observed across scenarios. Urban expansion caused the degradation of multiple ES. However, water yield (WY) increased significantly due to surface sealing, which actually exacerbated water pollution risks. Ecological protection measures, on the other hand, favored the improvement in multiple ES. While WY decreased significantly due to enhanced infiltration and evapotranspiration, this essentially improved the region’s water conservation and regulation functions.

(3)

Synergistic relationships were generally observed among CS, SC, and HQ, whereas trade-offs predominantly characterized the relationships between WY and both CS and HQ. The relationship between WY and SC exhibited dynamic trade-offs or synergies depending on the specific land use pattern. As land use transforms, the trade-off/synergy relationships between ES also undergo corresponding changes, manifesting as intensification, weakening, or mutual conversion. Notably, urban expansion significantly intensified trade-off conflicts among ES, leading to the overall degradation of key ES. Ecological protection measures effectively alleviated the intensity of trade-offs between most ES pairs and enhanced synergies. It is important to note that under the EP scenario, the synergy between CS and HQ was weakened due to the characteristics of aquatic ecosystems (high habitat quality but low carbon sequestration capacity). This highlights the unique challenge faced by water-abundant cities in managing ES synergies. In planning practice, zoning and classification management policies should be formulated based on simulation results for different scenarios. Urban core areas should prioritize the restoration of blue-green spaces, water-dense zones should optimize vegetation configuration, and mountainous forest areas should strengthen closed-forest protection to achieve synergistic enhancement of ecosystem services.