Dynamic Simulation of Land Use Change and Assessment of Ecosystem Services Under Climate Change Scenarios: A Case Study of Shanghai, China

Abstract

Climate change and rapid urbanization exert significant impacts on ecosystem services (ESs). The rational assessment and prediction of ESs are crucial for urban sustainable development. This study analyzes the spatiotemporal changes in land use in Shanghai from 2000 to 2020 and evaluates the key ESs, including water yield, soil retention, carbon storage, and habitat quality. Furthermore, integrated “climate change-land use” scenarios were constructed to systematically simulate the response characteristics of ESs under different climate change and development pathways. The results indicate that Shanghai’s land use from 2000 to 2020 was characterized by continuous expansion of built-up land and a significant reduction in cropland. Ecological land exhibited a low and fragmented coverage. By 2040, the ecological protection (EP) scenario could effectively curb the disorderly expansion of built-up land and maintain the stability of cropland and woodland, whereas the natural development (ND) scenario would exacerbate urban sprawl towards the east and further fragment ecological land. From 2000 to 2020, water yield in Shanghai showed an increasing trend, soil retention initially decreased followed by a gradual recovery, carbon sequestration experienced minor fluctuations, and habitat quality exhibited a continuous decline. By 2040, the EP scenarios will effectively maintain water yield and soil retention functions, steadily enhance carbon sequestration and habitat quality, and mitigate the negative impacts of climate change. In contrast, the ND scenarios show an unstable trend of initial increase followed by decrease. Spatially, the western and northern regions consistently remain high-value ESs zones under both scenarios. In 2040, Shanghai’s ESs will exhibit distinct administrative district disparities, characterized by “peripheral sensitivity and central stability”. This pattern underscores the necessity for implementing zone-specific regulation strategies in future urban planning.

1. Introduction

Ecosystems provide indispensable multidimensional services for human survival and development. These services encompass provisioning functions (e.g., food and water supply), regulating functions (e.g., climate regulation and hydrological cycle), cultural functions (e.g., recreation and spiritual values), and supporting functions (e.g., nutrient cycling) [1,2]. However, with the rapid advancement of urbanization globally, this process has profoundly reshaped natural landscape patterns through large-scale land use changes, subjecting ecosystems to immense pressure [3,4]. The interaction and complementary process of technological change, population growth, and urbanization have led to rapid changes in land use cover through complex feedback, which has a significant impact on ecosystem services (ESs) [5,6]. As the fundamental medium of ecosystems, land resources sustain the provision of diverse ESs through complex processes of energy flow and material cycling. Different land use patterns and development intensities, however, significantly alter the types and quality levels of these services provided [7]. More critically, ecosystem degradation driven by urbanization and climate change form a negative feedback loop. For instance, the intensification of the urban heat island effect leads to an increase in carbon emissions, further amplifying risks from extreme climate events [8,9]. Climate change exerts both direct and indirect impacts on ESs. Direct effects occur through alterations in biophysical parameters such as temperature and precipitation. Indirect effects manifest via biogeochemical pathways, disrupting species distribution and ecological processes. These impacts typically result in the degradation or loss of ES functions [10,11]. Consequently, the global configuration of ESs is undergoing profound transformation under the dual drivers of land use change and climate change [12]. In this context, predicting future land use dynamics and identifying sustainable management schemes for urban ecosystems under various climate and socioeconomic trajectories has emerged as a critical and albeit challenging global imperative.

Climate and land use interactively influence ESs, rather than acting independently. Climatic conditions shape land use patterns by regulating vegetation dynamics, while land cover changes, in turn, influence local climate characteristics. This bidirectional coupling necessitates a systemic perspective for assessment of climate and land use change simultaneously [13]. This facilitates a more rational assessment of regional ESs and their dynamics [14]. Models such as the cellular automata (CA), Markov chain, future land use simulation (FLUS) model, and patch-general land use simulation (PLUS) model, are commonly employed for land use and land cover change (LUCC) analysis and simulation. Among these, the PLUS model, which integrates both natural and socioeconomic factors, demonstrates high accuracy in simulating future land use changes [7,15,16]. The introduction of Coupled Model Intercomparison Project Phase 6 (CMIP6) has advanced research coupling climate change and socioeconomic scenarios. This enables the exploration of potential future climate–socioeconomic interactions at a macro scale [7,17,18,19]. Umwali, Chen, Ma, Guo, Mbigi, Zhang, Umugwaneza, Gasirabo and Umuhoza [7] integrated SSP-RCP scenarios to project future land use and ESs changes, and their results indicated that climate change exerts a more significant impact on ecosystems than land use change. Wang et al. [20] estimated terrestrial ecosystem carbon sequestration change in the Yangtze River Economic Belt under SSP-RCPs scenarios, revealing that the Grain-for-Green Program and urban expansion were key factors driving TECS change in the Yangtze River Delta. Zhou, Qu, Wang, Wu and Shi [13] evaluated ES bundles under SSP-RCP and local scenarios, noting that increasing ES provision is not the primary goal of local spatial management; instead, resolving critical trade-offs between services should be prioritized. Although existing research, by integrating historical climate and land use data, has made significant progress in deciphering the mechanisms of ES evolution and provided theoretical support for current ecological security maintenance, its guidance for adaptive management in response to future climate change remains limited [7,15,21]. Concurrently, climate change can induce nonlinear fluctuations in ES provision by altering the frequency and intensity of extreme events [10,22]. This underscores the urgency of conducting multi-scenario predictive studies. Such research needs to couple climate change and land use models to construct integrated ‘climate-land use’ scenarios. This will enable the systematic simulation of ES response characteristics under different development pathways, thereby providing a quantitative basis for formulating climate-resilient ecological management strategies.

Over recent decades of rapid urbanization, China has transformed from an agricultural nation into a megacity-dominant nation where the majority of its population resides in cities [3]. This continuous population agglomeration in urban areas has triggered significant eco-environmental problems, including widespread water, soil, and air pollution, degradation of soil, and biodiversity loss [23,24]. The National New-Type Urbanization Plan released in March 2014 aims to promote urbanization as a key driver of sustained economic growth while steering it towards more sustainable outcomes. This involves a shift from land-driven development to prioritizing people’s well-being, and from pursuing quantitative expansion to emphasizing the quality of urban development. Relevant research also suggests that future urbanization should strive to avoid the severe environmental degradation witnessed in many Chinese cities. Prioritizing the speed of transformation over people’s well-being may pose enduring challenges and adversely impact urban sustainability [25]. As China’s largest economic center and the core city of the Yangtze River Delta urban agglomeration, Shanghai serves not only as a vital engine for national economic development but also as a demonstration zone for coordinating urbanization with ecological conservation. Being a typical megacity in a large river estuary, Shanghai has undergone rapid urban expansion over the past decades. Rapid industrialization and population concentration have exerted immense pressure on the regional ecosystem, particularly through the extensive encroachment upon ecological spaces along the river and coastal zones. Concurrently, as one of the world’s largest port cities, Shanghai is highly vulnerable to climate change-induced disasters such as typhoons, meteorological hazards, and flooding due to its unique geographical location, dense population, and concentrated assets [26,27]. Achieving a balance between economic development and ecological protection constitutes a major challenge for current urban sustainable development.

Therefore, this study selected Shanghai as the research area. A novel framework was developed, integrating the PLUS model and the Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) with SSP-RCP scenarios from CMIP6, to assess the combined impacts of climate change and land use change on ecosystem services (ESs). This framework enables a comprehensive evaluation of the integrated effects of climate and land use change on urban ESs. Specifically, based on land use data, meteorological data, soil data, and other relevant datasets from 2000 to 2020, this study analyzes the spatiotemporal changes in Shanghai’s land use during this period and projects future land use for 2040 under different scenarios. Subsequently, four key ESs in the study area are assessed: water yield, soil retention, carbon storage, and habitat quality. Finally, by coupling two potential land use scenarios for 2040 (natural development, ND, and ecological protection, EP) with three climate change scenarios (SSP1-1.9, SSP2-4.5, SSP5-8.5), the impacts of climate and LUCC on ESs are quantitatively explored under six distinct future scenario combinations. This approach allows for the projection of ES development trends in Shanghai by 2040 and the assessment of ecological risks and pressures across different regions. The findings are intended to provide effective urban governance recommendations for addressing climate change, promoting adaptation, mitigating urban ecological risks and pressures, and facilitating city-oriented sustainable development transitions.

2. Methodology

2.1. Study Area

Shanghai, situated in eastern China at the mouth of the Yangtze River and facing the Pacific Ocean, features predominantly low and flat terrain. It gently slopes from west to east, with an average elevation of approximately 4 m, constituting a typical estuarine alluvial plain [28]. The primary land use types encompass built-up land, water area, and agricultural land. Experiencing a subtropical monsoon climate, Shanghai exhibits four distinct seasons with synchronous occurrences of rainfall and warm temperatures. Its climate is markedly moderated by maritime influences. The urban area records an average annual temperature of 17.1 °C and receives approximately 1250 mm of annual precipitation. Shanghai covers an area of 6340.5 km² with a permanent resident population of 24.89 million. Shanghai administers 16 districts (Figure S1), with its central urban area comprising Huangpu, Xuhui, Changning, Jing’an, Putuo, Hongkou, and Yangpu districts, along with portions of the Pudong New Area, Minhang, Baoshan, and Jiading districts, while its suburban districts comprise the Qingpu, Songjiang, Fengxian, Jinshan, and Chongming districts [29]. Together with the neighboring provinces of Zhejiang, Jiangsu, and Anhui, it constitutes the Yangtze River Delta—one of China’s most economically dynamic, open, and innovation-driven regions (Figure 1). As China’s paramount economic center and a pivotal international financial hub, propelled by intense population pressure and urbanization demands, Shanghai has emerged as one of the nation’s most prominent cities in terms of land expansion [28,30].

2.2. Data Sources and Data Pre-Processing

This study utilized multiple data sources. One of these was land cover data: land use data for Shanghai in 2000, 2010, and 2020 were obtained from the China Annual Land Cover Dataset (1985–2023) at a 30 m resolution. This dataset was developed primarily by utilizing Landsat imagery from the Google Earth Engine platform, integrating stable samples from the Chinese Land Use/Cover Dataset (CLUD), and incorporating manually interpreted samples from satellite time-series data, Google Earth, and Google Maps. Multi-temporal indices were constructed using available Landsat data and then input into a random forest classifier to generate the land use classification results [31].

Driving factor data were also used: referencing the relevant literature and considering regional characteristics, 12 driving factors of land use encompassing natural and socioeconomic elements were selected for land use change simulation. These included DEM, slope, aspect, temperature, precipitation, soil type, distance to water bodies, distance to highways, distance to railways, distance to residential areas, GDP spatial distribution, and population spatial distribution (

Table 1. Data type and source.

Data Type	Data Name	Resolution/m	Data Sources	Year
Land use data	LUCC	30	The 30 m annual land cover datasets and its dynamics in China from 1985 to 2023 (https://doi.org/10.5281/zenodo.12779975 (accessed on 2 September 2025))	2000, 2010, 2020
Driving factors	DEM	30	Geospatial Data Cloud (https://www.gscloud.cn/ (accessed on 2 September 2025))	2024
	Average annual precipitation	1000	National Tibetan Plateau data center (https://data.tpdc.ac.cn/ (accessed on 2 September 2025))	2020, 2040
	Annual average temperature	1000	National Tibetan Plateau data center (https://data.tpdc.ac.cn/ (accessed on 2 September 2025))	2020, 2040
	Soil type	1000	Resource and Environmental Science Data Platform (https://www.resdc.cn/ (accessed on 2 September 2025))	1995
	Slope	30	Derived from DEM data calculations	2024
	Aspect of slope	30	Derived from DEM data calculations	2024
	Spatial distribution of GDP	1000	Resource and Environmental Science Data Platform (https://www.resdc.cn/ (accessed on 2 September 2025))	2020
	Spatial distribution of population density	1000	Resource and Environmental Science Data Platform (https://www.resdc.cn/ (accessed on 2 September 2025))	2020
	Distance to water system, traffic road (highway, railway) and residential area	—	National Catalogue Service For Geographic Information (https://www.webmap.cn/ (accessed on 2 September 2025))	2021
Auxiliary data	Evapotranspiration data	1000	National Tibetan Plateau data center (https://data.tpdc.ac.cn/ (accessed on 2 September 2025))	2040
	Available moisture content of vegetation	1000	Harmonized World soils Database version 2.0 (https://gaez.fao.org/pages/hwsd (accessed on 2 September 2025))	2020
	Administrative boundary	—	National Platform for Common GeoSpatial Information Services (https://www.tianditu.gov.cn/ (accessed on 2 September 2025))	2024
	Nature reserve data	—	China Nature Reserves Specimen Resources Sharing Platform (http://www.papc.cn/html/ (accessed on 2 September 2025))	2024

).

Climate data, including historical and future climate datasets, were sourced from the National Tibetan Plateau Data Center (https://data.tpdc.ac.cn/ (accessed on 2 September 2025)). Historical temperature and precipitation data came from the China 1 km Monthly Mean Temperature Dataset (1921–2024) and China 1 km Monthly Precipitation Dataset (1901–2024). Future climate scenario datasets included the China 1 km Monthly Mean Temperature Dataset (2021–2100) and China 1 km Monthly Precipitation Dataset (2021–2100) under multiple scenarios and models. These were generated through the Delta spatial downscaling method over China, based on global >100 km climate model data from the Intergovernmental Panel on Climate Change (IPCC) Coupled Model Intercomparison Project Phase 6 (CMIP6) and high-resolution climate data from WorldClim, using the latest IPCC SSP scenarios (SSP1-1.9, SSP2-4.5, SSP5-8.5).

Prior to land use simulation, all datasets were preprocessed in ArcGIS 10.8. Data were unified to the same projection coordinate system and resampled to a 30 m spatial resolution to ensure model compatibility.

2.3. Analysis Methods

2.3.1. PLUS Model

PLUS is a model that simulates future land use/cover change based on existing land use patterns. It integrates a rule-mining framework derived from the land expansion analysis strategy (LEAS) model and a cellular automata model based on multiple random seeds (CARS) [32]. The model initially overlays land use data from two distinct periods, extracting changed pixels from the later dataset to quantify transition areas for each land use type. It then employs the Random Forest (RF) algorithm to explore relationships between land use types and various driving factors, deriving transition rules that represent the development potential of each land use category. Constrained by this development potential, PLUS automatically generates simulated patches [33].

PLUS excels in patch-level simulation capability. It dynamically integrates spatial factors with geographical units to simulate land use changes, producing high-precision spatial distributions [34,35]. Compared to traditional models (e.g., CLUE-S, CA-Markov), PLUS generates land use patches with realistic landscape patterns through its CARS mechanism [15,36]. This is particularly critical for highly urbanized regions like Shanghai, where conventional models yield overly regularized and unrealistic results. PLUS employs a dual-scale simulation framework that organically combines macro-scale land demand forecasting and micro-scale spatial allocation. This structure aligns with Shanghai’s multi-tiered planning management requirements. The model demonstrates robust analytical power in deciphering driving factors, enabling the in-depth assessment of natural and socioeconomic influences on land use transitions—essential for understanding Shanghai’s development dynamics under complex drivers.

PLUS v1.4 was employed to the simulation and prediction of land use change in this study. This study initially utilized the LULC data conversion module to transform the land use data of Shanghai for the years 2010 and 2020 into unsigned character format. Subsequently, the land expansion patterns for these two years were extracted using the land expansion extraction module. The LEAS module was employed to perform random forest classification on the expansion of various land types and their driving factors, ultimately generating development probability maps for each land type region. Within the CARS module, by inputting Shanghai’s 2020 land use type data, development potential, land demand, transition matrix, and neighborhood weights, the final output was the land use distribution data for Shanghai in 2040 under different development scenarios.

Using Shanghai’s 2000 and 2010 historical land use data and 12 driving factors, PLUS simulated 2020 land use patterns. Validation of actual 2020 data yielded a Kappa coefficient of 0.75 and an FoM (Figure of Merit) index of 0.11, confirming a high simulation accuracy sufficient for Shanghai’s future land use projections. Consequently, this study adopts the same methodology to simulate 2040 land use patterns under different climate change scenarios.

2.3.2. ESs Assessment

This study used InVEST 3.15.0 to assess ecosystem services in Shanghai. InVEST, developed by the Natural Capital Project in the United States, evaluates relationships between land use patterns and ESs. This model quantifies ESs based on specific land use types and presents results spatially and graphically. By assessing changes in ESs, it provides decision-making support for regional ecosystem management [37]. This research employs the InVEST model, integrating region-specific geographical characteristics to simulate and quantify four key ESs: water yield, soil retention, carbon sequestration and habitat quality.

1.

Water yield

Due to climate change, Shanghai experiences frequent extreme heavy rainfall events, intensifying urban flooding challenges [38]. Natural ecosystems (e.g., wetlands, forests) mitigate surface runoff and delay flood peaks through precipitation interception and infiltration enhancement. The InVEST model applies the SCS-CN (Soil Conservation Service Curve Number) method to simulate precipitation–runoff relationships, evaluating impacts of different land use scenarios on urban flood control. The Annual Water Yield module calculates regional water yield capacity using annual average precipitation, annual potential evapotranspiration, plant available water content, vegetation types and soil depth. The water yield was calculated following Equations (1)–(4).

$$Y\left(x\right)=\left(1-{\displaystyle \frac{\mathrm{A}\mathrm{E}\mathrm{T}\left(x\right)}{P\left(x\right)}}\right)\times P\left(x\right)$$

(1)

where $Y\left(x\right)$ represents the annual water yield (mm); $\mathrm{A}\mathrm{E}\mathrm{T}\left(x\right)$ is the annual actual evaporation of grid unit x (mm); $P\left(x\right)$ is the annual precipitation for grid unit x (mm).

$${\displaystyle \frac{\mathrm{A}\mathrm{E}\mathrm{T}\left(x\right)}{P\left(x\right)}}=1+{\displaystyle \frac{\mathrm{P}\mathrm{E}\mathrm{T}\left(x\right)}{P\left(x\right)}}-{\left[1+{\left({\displaystyle \frac{\mathrm{P}\mathrm{E}\mathrm{T}\left(x\right)}{P\left(x\right)}}\right)}^{\omega }\right]}^{\frac{1}{\omega }}$$

(2)

$$\mathrm{P}\mathrm{E}\mathrm{T}\left(x\right)=K_{\mathrm{c}}\left(x\right)\times E{T}_{\mathrm{o}}\left(x\right)$$

(3)

$$\omega \left(x\right)={\displaystyle \frac{\mathrm{A}\mathrm{W}\mathrm{C}\left(x\right)}{P\left(x\right)}}\times Z+1.25$$

(4)

where $\mathrm{P}\mathrm{E}\mathrm{T}\left(x\right)$ is the potential evapotranspiration for grid unit x; $E{T}_{\mathrm{o}}\left(x\right)$ is the reference vegetation evapotranspiration; $K_{\mathrm{c}}\left(x\right)$ is the crop evapotranspiration coefficient; $\mathrm{A}\mathrm{W}\mathrm{C}\left(x\right)$ is the plants’ ability to utilize water content; $\omega \left(x\right)$ represents the empirical parameters; $Z$ is the empirical constant.

2.

Soil retention

Situated on the Yangtze Delta alluvial plain, Shanghai features predominantly soft soils and silty clays with low erosion resistance. During urbanization, large-scale land leveling, construction activities, and road development destroy surface vegetation, leaving bare soils highly susceptible to washout under heavy rainfall conditions [39]. Soil erosion consequences extend beyond nutrient loss to increased burden on urban drainage systems, elevating flooding risks. The InVEST model evaluates soil retention capacity to identify high-risk erosion zones and optimize green space configuration for sediment reduction. Soil retention was quantified using the Universal Soil Loss Equation (USLE), shown in Equations (5)–(7).

$$\mathrm{R}\mathrm{K}\mathrm{L}\mathrm{S}=R\times K\times LS$$

(5)

$$\mathrm{U}\mathrm{S}\mathrm{L}\mathrm{E}=R\times K\times LS\times P\times C$$

(6)

$$SDR=RKLS-USLE$$

(7)

where $\mathrm{R}\mathrm{K}\mathrm{L}\mathrm{S}$ represents the potential soil erosion amount; $\mathrm{U}\mathrm{S}\mathrm{L}\mathrm{E}$ represents the actual amount of soil erosion; $SDR$ represents the soil retention amount; $R$ is the rainfall erosivity factor; $K$ is the soil erodibility factor; $LS$ is the slope length factor; $P$ is the soil and water yield measurement factor; $C$ is the vegetation cover management factor.

3.

Carbon sequestration

As China’s largest economic center, Shanghai accounts for approximately 1.5% of the nation’s total carbon emissions. Research indicates that the urban heat island intensity in Shanghai’s central urban area reaches up to 3.5 °C [40]. Urban vegetation delivers the dual climate benefits of carbon sequestration through photosynthesis and microclimate regulation via transpiration. The carbon sequestration module within the InVEST model quantifies ecosystem carbon stocks, enabling the identification of high-carbon-density zones and optimization of green space planning. This approach achieves synergistic outcomes in both carbon fixation and urban cooling. The following formula was used:

$$C_{\mathrm{t}\mathrm{o}\mathrm{t}}=C_{\mathrm{a}\mathrm{b}\mathrm{o}\mathrm{v}\mathrm{e}}+C_{\mathrm{b}\mathrm{e}\mathrm{l}\mathrm{o}\mathrm{w}}+C_{\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}}+C_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d}}$$

(8)

where $C_{\mathrm{t}\mathrm{o}\mathrm{t}}$ represents the total carbon storage; $C_{\mathrm{a}\mathrm{b}\mathrm{o}\mathrm{v}\mathrm{e}}$ represents the aboveground biochar storage; $C_{\mathrm{b}\mathrm{e}\mathrm{l}\mathrm{o}\mathrm{w}}$ represents the underground biochar storage; $C_{\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}}$ represents the soil carbon storage; $C_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d}}$ represents the dead organic carbon storage.

4.

Habitat quality

The implementation of Shanghai’s “14th Five-Year Plan” for ecological space prioritizes establishing well-connected ecological networks. Habitat quality assessment identifies ecological corridor locations and optimizes the ecological functions of waterfront spaces along the Huangpu River and Suzhou Creek (“One River, One Creek”) and urban green space systems [41]. The Habitat Quality module within the InVEST model evaluates regional habitat quality by calculating degradation levels and habitat suitability. This effectively analyzes negative impacts of urbanization stressors—including road networks and building density—on habitats. The following formula was used:

$$Q_{xj}=H_j\times \left[1-\left({\displaystyle \frac{D_{xj}^z}{D_{xj}^z+k^z}}\right)\right]$$

(9)

where $Q_{xj}$ represents the habitat quality index of grid x in land use j; $H_j$ is the habitat suitability of habitat type j, with a value range of 0–1; $D_{xj}$ is the habitat degradation index; $k$ is the semi saturation constant; $z$ is the normalization constant.

2.4. Multi-Scenario Configuration

This study employed the PLUS model to simulate Shanghai’s future land use under different climate and land planning scenarios for 2040. To address climate change challenges and assess future climate trajectories, the Sixth Assessment Report of the IPCC introduced new carbon emission scenarios. Building upon CMIP5’s four representative RCP scenarios, CMIP6 expanded its RCP scenario dataset by adding three additional concentration-based scenarios (RCP1.9, RCP3.4, RCP7.0) to fill gaps in CMIP5’s RCP framework. The report also introduced the concept of Shared Socioeconomic Pathways (SSPs), which integrate population, economic, social, and urbanization factors. Using Integrated Assessment Models (IAMs), CMIP6 developed new SSP-RCP combinations: SSP1-1.9 (low radiative emissions), SSP2-4.5 (moderate radiative forcing), SSP3-7.0 (moderately high radiative forcing), SSP4-3.4 (low radiative forcing), SSP4-6.0 (moderate radiative emissions), and SSP5-8.5 (high radiative forcing). This study selects SSP1-1.9, SSP2-4.5, and SSP5-8.5 to represent low-, medium-, and high emission pathways, respectively [42]. Land planning scenarios include the natural development mode (ND) and ecological protection mode (EP) [43]. The natural development scenario refers to future land use planning that adheres to existing patterns of land use change, without imposing restrictions on areas outside of nature reserves, allowing for free development. Meanwhile, within the study area, nature reserves are treated as constraints. The ecological conservation scenario, on the other hand, emphasizes the enhanced protection of the ecological environment by restricting urban expansion, promoting land use transformation towards more natural states, strictly prohibiting habitat degradation, and designating regional water bodies as constraints to prevent arbitrary conversion. Six combined scenarios for 2040 were designed: ND scenarios under three climate pathways, ND-SSP1-1.9(ND119), ND-SSP2-4.5(ND245), and ND-SSP5-8.5(ND585), and EP scenarios under three climate pathways, EP-SSP1-1.9(EP119), EP-SSP2-4.5(EP245), EP-SSP5-8.5(EP585).

3. Results

3.1. LUCC in Shanghai from 2000 to 2020

Table 2. Land use and cover change in Shanghai in 2000–2020.

Land Use Type	Land Use Area (km2/Proportion)
	2000	2010	2020
Farmland	5064.11	4361.00	4168.51
	72.86%	62.75%	59.98%
Built-up land	1302.73	2137.41	2477.77
	18.74%	30.75%	35.65%
Water	575.81	435.38	294.70
	8.28%	6.26%	4.24%
Woodland	7.46	12.81	8.90
	0.11%	0.18%	0.13%
Bare ground	0.07	0.31	0.25
	0.00%	0.00%	0.00%
Grassland	0.04	3.32	0.08
	0.00%	0.05%	0.00%

presents the area and proportion of each land use type in Shanghai from 2000 to 2020. The overall land use changes reveal the following patterns. Farmland maintained the largest share among all land use categories, with areas of 5064.11 km² (72.86%), 4361 km² (62.75%), and 4168.51 km² (59.98%) in 2000, 2010, and 2020, respectively. It exhibited a continuous decline over the two-decade period. Built-up land ranked second in area proportion, expanding from 1302.73 km² (18.74%) in 2000 to 2137.41 km² (30.75%) in 2010 and further to 2477.77 km² (35.65%) in 2020, demonstrating a pronounced upward trend. Water area decreased in area from 575.81 km² (8.28%) in 2000 to 435.38 km² (6.26%) in 2010 and then to 294.70 km² (4.24%) in 2020, reflecting a consistent decline. Woodland showed an initial increase followed by a decrease, with areas of 7.46 km² (0.11%), 12.81 km² (0.18%), and 8.90 km² (0.13%) across the three time points. Grassland and bare land accounted for small proportions of the total area, with no significant changes observed.

The spatial distributions of land use in Shanghai from 2000 to 2020 are shown in Figure 2. Analysis reveals that the city’s land use pattern was predominantly characterized by built-up areas and farmland, with urban expansion radiating outward from Yangpu District as a key feature. Built-up land was concentrated in central urban zones and surrounding key regions, continuously expanding between 2000 and 2020 and progressively encroaching upon agricultural and other land spaces.

Figure 3 illustrates land use transitions in Shanghai. The spatial distributions of land use changes during 2000–2020 are shown in Figure 3a–c. Transition matrix analysis reveals substantial changes in farmland and built-up land areas: farmland reduction primarily occurred during 2000–2010 (893.01 km² decrease), slowing to 355.22 km² in 2010–2020. Conversion to built-up land constituted the primary driver of farmland loss, though partial water-to-farmland conversion also occurred. Conversely, woodland transitions concentrated in 2010–2020, with deforestation mainly driven by conversions into built-up land. Net woodland change showed a 6.43 km² gain versus 4.98 km² loss during 2000–2020. The area of built-up land significantly increased during the period 2000–2020, mainly converted from substantial farmland (1111.88 km²) and partially from water area (74.26 km²). Grassland and bare land maintained stable total areas with balanced gains and losses. Water area decreased substantially, particularly during 2010–2020 (435.38 km² reduction), being converted primarily into farmland and built-up land.

Further statistical analysis was conducted across Shanghai’s 16 administrative districts regarding land use area and proportion during this period (Table S1 of the Supplementary Materials). Farmland accounted for a high proportion in the early stage. Chongming, Pudong New Area and Fengxian Districts have been dominated by farmland for a long time. With the promotion of urbanization, the farmland in Pudong New Area and Fengxian Districts were significantly reduced from 968.98 km² and 609.42 km² in 2000 to 717.49 km² and 508.45 km² in 2020, with reduction rates of 25.95% and 16.57%, respectively. At the same time, the farmland in the outer suburbs of Shanghai has also been affected, showing a decreasing trend. As an island at the mouth of the Yangtze River, Chongming district has been increasing its farmland area from 1221.85 km² in 2000 to 1298.03 km² in 2020 due to its special geographical location. The proportion of woodland, grassland and other ecological land in each administrative region is relatively low and scattered, mostly distributed in the outer suburbs or protected areas, and has not formed a large-scale centralized distribution. Especially, the area of woodland and grassland in Baoshan, Huangpu, and Jing’an districts in the city center is very small. The spatial distribution of water area is relatively stable. Although Chongming District accounted for the largest water areas, which were 227.60 km², 174.54 km² and 98.33 km² in 2000, 2010 and 2020, respectively, a rapid decreasing trend (56.80%) in the two decades was detected. Jing’an District occupied the smallest water areas, with 0.03 km², 0.12 km² and 0.10 km² in the three years, respectively.

3.2. ES Value in Shanghai from 2000 to 2020

Table 3. Total and average value of ESs in Shanghai from 2000 to 2020.

Year	Water Yield		Soil Retention		Carbon Storage		Habitat Quality
	Total (m3)	Average (mm)	Total (t)	Average (t/hm2)	Total (t)	Average (t/hm2)	Average
2000	37.37 × 108	564.09	22.95 × 106	31.85	35.16 × 106	50.59	0.29
2010	37.68 × 108	568.76	19.54 × 106	27.13	34.75 × 106	49.99	0.25
2020	44.75 × 108	645.47	21.40 × 106	29.69	34.88 × 106	50.18	0.22

shows the total amount and average per area value of ESs in Shanghai in 2000, 2010 and 2020. The water yield in Shanghai is on the rise, soil retention first decreases and then rises slowly, carbon sequestration fluctuates slightly, and habitat quality continues to decline. During the period from 2000 to 2020, the water yield of Shanghai showed an overall upward trend, increasing from 37.37 × 10⁸ m³ to 44.75 × 10⁸ m³. From 2000 to 2010, the average value of water yield increased by 4.67 mm, with relatively moderate growth. From 2010 to 2020, the growth was significant, with the average value increasing by 76.71 mm. The overall trend of soil retention is first decreasing and then increasing. From 2000 to 2010, soil retention decreased from 22.95 × 10⁶ t to 19.54 × 10⁶ t, and slightly increased to 21.40 × 10⁶ t from 2010 to 2020. Carbon sequestration decreased from 35.16 × 10⁶ t to 34.75 × 10⁶ t, and then slightly rose to 34.88 × 10⁶ t, showing a trend of first decreasing and then stabilizing. Habitat quality showed an overall continuously deteriorating trend from 2000 to 2020, dropping from 0.29 to 0.22.

The spatial distributions of ESs are shown in Figure 4. Areas with high water yield were predominantly distributed in central urban zones and coastal regions, whereas western and southern districts generally exhibited lower water yield levels. High-value soil retention areas were primarily concentrated in Chongming District, while most central urban areas demonstrated comparatively low soil retention values. Carbon storage peaked in suburban belts surrounding the central city and southern Chongming District, with low-value clusters mainly found in western suburban districts and northern coastal areas of Chongming. Superior habitat quality prevailed in the western suburban districts, northern Chongming District, and the coastal eastern Pudong New Area. Conversely, central urban cores consistently display inferior habitat quality.

District-level analysis (Table S3 of the Supplementary Materials) further demonstrated key spatial patterns. For water yield, throughout 2000–2020, Jing’an District consistently recorded the highest mean water yield, with values of 862.60 mm (2000), 781.17 mm (2010), and 940.32 mm (2020). Qingpu District witnessed the lowest mean water yield in both 2000 (421.67 mm) and 2010 (419.72 mm), while in 2020, Fengxian District had the lowest value of 549.08 mm.

For soil retention, from 2000 to 2020, Jinshan District consistently maintained the highest mean soil retention values at 43.60 t/hm² (2000), 40.74 t/hm² (2010), and 43.45 t/hm² (2020). Conversely, Jing’an District recorded the lowest mean soil retention levels throughout the period, with values of 2.27 t/hm² (2000), 1.91 t/hm² (2010), and 2.40 t/hm² (2020).

For carbon sequestration, from 2000 to 2020, Jinshan District consistently demonstrated the highest mean carbon sequestration at 54.36 t/hm² (2000), 53.43 t/hm² (2010), and 53.32 t/hm² (2020). Conversely, Huangpu District maintained the lowest mean carbon sequestration values throughout the period, with values of 42.66 t/hm² (2000), 42.56 t/hm² (2010), and 42.50 t/hm² (2020).

For habitat quality, the highest values emerged in remote, highly vegetated areas, where Chongming maintained the top index values of 0.37 (2000), 0.34 (2010), and 0.31 (2020). The lowest values occurred along major transportation corridors and central urban zones (e.g., Jing’an consistently scored 0.01).

3.3. Multi-Scenario Simulations

3.3.1. Prediction of LUCC

Table 4. Land area in 2040 under different scenario simulations.

Land Use Type	Area(km2)/Proportion
	ND119	EP119	ND245	EP245	ND585	EP585
Farmland	3358.52	3628.51	3358.52	3628.52	3358.52	3628.52
	48.32%	52.21%	48.32%	52.21%	48.32%	52.21%
Woodland	8.17	8.90	8.17	8.90	8.28	8.90
	0.12%	0.13%	0.12%	0.13%	0.12%	0.13%
Grassland	0.06	0.08	0.06	0.08	0.06	0.08
	0.00%	0.00%	0.00%	0.00%	0.00%	0.00%
Water	312.89	294.70	311.33	294.70	311.17	294.70
	4.50%	4.24%	4.48%	4.24%	4.48%	4.24%
Bare ground	0.21	0.23	0.21	0.23	0.21	0.23
	0.00%	0.00%	0.00%	0.00%	0.00%	0.00%
Built-up land	3270.36	3017.80	3271.93	3017.80	3271.98	3017.80
	47.05%	43.42%	47.08%	43.42%	47.08%	43.42%

presents the area and proportion of each land use type under Shanghai’s six development scenarios in 2040. Under the ND scenario, ND119, ND245 and ND585 showed little change in the area and proportion of each land use type, among which farmland, grassland and bare land remained unchanged, with an area of 3358.52 km², 0.06 km² and 0.21 km², respectively. The woodland area remained the same under the ND119 and ND245 scenarios, at 8.17 km², increasing to 8.28 km² under the ND585 scenario. The water area is the largest in the ND119 scenario and the smallest in the ND585 scenario, at 312.89 km² and 311.17 km², respectively. On the contrary, the area of construction land is the lowest in the ND119 scenario and the highest in the ND585 scenario. Under the EP scenario, the land use types in Shanghai in 2040 will be the same as those in the EP119, EP245 and EP585 scenarios. The reason may be that due to the small research scale and short prediction time span in Shanghai, the impact of climate change is not obvious enough in a short period of time, thus resulting in no change in the area of land use types under various climate change scenarios.

Figure 5 illustrates the spatial distribution of land use in Shanghai under different 2040 development scenarios. Farmland and construction land share similar spatial distribution characteristics under the ND119, ND245, and ND585 scenarios. In all ND scenarios, farmland is mainly concentrated in the northern and southern parts of the study area, forming large continuous distribution regions. The area of construction land under the ND119, ND245, and ND585 scenarios is mainly distributed in the central and northern parts of the study area, especially in and around the urban center, showing a highly concentrated feature. Under the ND585 scenario, the distribution range of farmland and construction land expands, especially in the eastern and southern edge areas of the study area.

Farmland and construction land also show similar spatial distribution characteristics under the EP119, EP245, and EP585 scenarios. In all EP scenarios, the area of farmland is 3628.51 km², and its distribution range is wider than that in the ND scenarios. Particularly in the central and southern parts of the study area, the area of farmland has increased significantly. The built-up land under EP119, EP245, and EP585 is relatively scattered, mainly concentrated in the eastern and northern parts of the study area. However, compared with the ND scenarios, both its area and proportion have decreased. Woodland is relatively sparsely distributed in all scenarios, mainly concentrated in the western and northwestern edge areas of the study area, forming sporadically distributed patches. The area of woodland is 8.17 km² in all ND scenarios and 8.90 km² in all EP scenarios, with the area and proportion of woodland in the EP scenarios being slightly higher than those in the ND scenarios. Water areas are mainly distributed in the northern and eastern parts of the study area, showing a strip-like distribution, which is consistent with the locations of rivers and lakes. The water areas in the ND scenarios are slightly larger than those in the EP scenarios, but the overall change is insignificant. Bare land and grassland have a very limited distribution in all scenarios, mainly concentrated in local areas of the study area, with both a small area and proportion.

Supplementary district-level analysis (Table S2 of the Supplementary Materials) reveals that Chongming District maintains the highest farmland proportion (36.5% under ND scenarios vs. 34.5% under EP scenarios). Pudong New District exhibits the largest built-up land share, though its area decreases from approximately 860 km² (ND scenarios) to 788 km² (EP scenarios). Chongming District also contains the most substantial water areas, with areas near 104 km² (peaking at 104.98 km² under ND119) under NP scenarios, while EP scenarios show slightly reduced but consistent areas around 98 km².

Figure 6 display projected 2020–2040 transitions under different scenarios. Farmland to built-up land conversion dominates across SSP scenarios, reaching 795.24 km²–796.28 km² under ND scenarios but decreasing to 540.77 km²–540.88 km² under EP scenarios due to China’s ecological conservation redline policy [37,44], which restricts urban sprawl. Shanghai is a highly urbanized area; its expansion inertia is economically driven, while the impact of differences in SSPs is relatively limited under the contexts of natural development and ecological protection. The area of farmland conversion under the different SSPs is similar, at approximately 796 km² in ND and 541 km² in EP. Under ND scenarios, apart from the conversion of farmland to construction land, a small portion of farmland is converted into water areas, which are 18.19 km², 16.63 km², and 16.47 km² under the SSP119, SSP245, and SSP585 scenarios, respectively. Under EP scenarios, farmland is regarded as an integral part of the ecological network, and its hydrological functions are maintained through ecological design, eliminating the need for conversion into independent water areas. Therefore, the conversion of farmland into water areas is completely inhibited.

3.3.2. ESs Values in Shanghai in Different Scenarios in 2040

The spatial distribution of ESs in Shanghai under different scenarios for 2040 is shown in Figure 7. Regarding the spatial patterns of ESs, water yield shows similar distributions across all six scenarios, with high-value areas concentrated primarily in Shanghai’s central urban area and the southeastern part of Chongming District. In terms of soil retention in 2040, the highest values are predominantly observed in Chongming District. Under the ND119, EP119, ND585, and EP585 scenarios, high-value zones are limited and scattered across southwestern Shanghai and northern Chongming Island. In contrast, the ND245 and EP245 scenarios demonstrate expanded high-value areas. Carbon sequestration in 2040 is generally higher in the southern regions and Chongming District, whereas lower levels are concentrated in the central urban and western areas. Notably, under ND245 and EP245 scenarios, low-carbon-storage zones expand in southern areas and Chongming Island but shrink in the central urban core. As for habitat quality in 2040, areas of high quality are mainly located in western Shanghai and northeastern Chongming Island, while low-quality zones form a concentrated pattern in the urban center, extending outward into suburban peripheries. However, under the ND245 and EP245 scenarios, the extent of low-habitat-quality areas significantly decreases—confined mainly around the central urban core—while high-value zones expand in southern Shanghai and northern Chongming District.

We further conducted statistical analyses of the total and average ESs values in Shanghai under different scenarios (

Table 5. Total and average value of ESs in Shanghai under different scenarios.

Scenario	Water Yield (mm)		Soil Retention (t/hm2)		Carbon Sequestration (t/hm2)		Habitat Quality
	Total (m3)	Average (mm)	Total (t)	Average (t/hm2)	Total (t)	Average (t/hm2)	Average
ND119	4.28 × 109	645.32	1.88 × 107	26.09	3.43 × 107	49.39	0.20
EP119	4.22 × 109	637.40	1.96 × 107	27.16	3.46 × 107	49.79	0.21
ND245	2.66 × 109	400.93	1.58 × 107	21.87	3.49 × 107	50.17	0.32
EP245	2.61 × 109	393.17	1.56 × 107	21.67	3.64 × 107	52.38	0.28
ND585	3.61 × 109	544.99	1.50 × 107	20.82	3.43 × 107	49.40	0.20
EP585	3.56 × 109	537.07	1.56 × 107	21.68	3.46 × 107	49.79	0.21

). Under ND scenarios, the highest total water yield in 2040 will occur in ND119 (4.28 × 10⁹ m³), decreasing to 2.66 × 10⁹ m³ in ND245 and 3.61 × 10⁹ m³ in ND585. Under EP scenarios, EP119 will yield a value of 4.22 × 10⁹ m³, which is higher than that yielded by EP245 (2.61 × 10⁹ m³) and EP585 (3.56 × 10⁹ m³). This indicates that under identical climate scenarios, ND will consistently produce higher water yields than EP. For soil retention in 2040, EP119 will peak at 1.96 × 10⁷ t, followed by ND119 (1.88 × 10⁷ t). Values under ND245, EP245, ND585, and EP585 will remain comparable, with ND585 recording the lowest (1.50 × 10⁷ t). Carbon sequestration will reach its maximum under EP245 (3.64 × 10⁷ t), followed by ND245 (3.49 × 10⁷ t). Minimal variations will occur under ND119, EP119, ND585, and EP585, with ND119 having the lowest (3.43 × 10⁷ t). Habitat quality will peak under ND245 (0.32) while reaching lowest value under ND119 and ND585 (0.20).

Figure 8 presents line distribution charts of the mean values of various ESs across different administrative districts in Shanghai under six development scenarios for the year 2040. The average water yield capacity across districts will range between 300 mm and 850 mm, exhibiting significant spatial heterogeneity (Figure 8a). Jing’an District will show the highest average water yield capacity among all districts, while Qingpu District will have the lowest. Minimal differences were observed in our projections of water yield between ND and EP scenarios under the SSP119, SSP245, and SSP585 pathways. However, our projection for the districts demonstrated significantly lower average water yield under SSP245 compared to SSP119 and SSP585, with the highest values occurring under SSP119.

Average soil retention capacity will range from 0 to 45 t/hm² across districts, showing little variation among development scenarios (Figure 8b). Jinshan District will consistently exhibit the highest average soil retention across all scenarios, whereas Jing’an District will consistently record the lowest. Most districts will show higher soil retention under ND119 and EP119 scenarios. However, the values for Minhang, Baoshan, and Jiading districts will peak under ND245 and EP245 scenarios. Jing’an District will display minimal variation in soil retention across scenarios.

Average carbon sequestration will primarily range from 42 t/hm² to 56 t/hm², with most districts reaching peak values under the EP245 scenario (Figure 8c). Districts including Pudong New Area, Jinshan, Songjiang, Fengxian, and Chongming will maintain consistently high carbon sequestration across all scenarios, particularly under EP245 and EP585. Conversely, Putuo, Hongkou, and Huangpu districts will consistently show lower values. Significant fluctuations under EP245 and ND245 will be observed in Qingpu, Minhang, Baoshan, Jiading, Pudong New Area, Jinshan, Songjiang, Fengxian, and Chongming districts.

The average habitat quality index across Shanghai’s districts will range from 0 to 0.5. Most districts will achieve their highest values under ND245, closely followed by EP245 (Figure 8d). Qingpu, Chongming, Fengxian, Jinshan, and Songjiang districts will maintain generally high habitat quality across scenarios, while Putuo, Changning, and Jing’an districts will consistently show lower values. Qingpu District will demonstrate the highest habitat quality relative to other districts in all scenarios, though the projected values varied significantly across development pathways—with a particularly prominent peak under ND245.

4. Discussion

4.1. LUCC in Shanghai During 2000–2020

The overall pattern of land use intensity in Shanghai from 2000 to 2020 exhibited a “center-suburb-exurban” gradient. The urban core was highly developed with dense built-up areas, while suburban regions saw intensified development featuring interwoven farmland and construction zones. Exurban areas developed relatively slowly and remained predominantly agricultural. Over these two decades, significant urban expansion occurred as built-up land continuously encroached upon farmland, with ecological land maintaining low coverage and showing signs of fragmentation.

This evolution was driven by multiple factors including socioeconomic forces, urban planning policies, and climate change. Urban expansion and economic development served as primary drivers for the substantial increase in built-up land. As China’s economic hub and global city, Shanghai underwent rapid urbanization, where population growth, industrial agglomeration, and infrastructure development fueled surging spatial demands—particularly through outward expansion from the urban core that consumed surrounding farmland and water area. This trend aligns with patterns observed in developing economies where rapid economic growth often occurs at the expense of ecological and agricultural spaces [45].

Policy interventions played crucial roles in shaping land use transitions. Post-2000 industrial restructuring policies (“retreating secondary industries and advancing tertiary industries”) accelerated built-up land expansion by converting industrial zones into zones for commercial and residential use [46]. Conversely, Chongming District implemented strict farmland protection policies under its ecological island designation (“Chongming World-Class Ecological Island Development Thirteenth Five-Year Plan”), increasing farmland area against the broader trend of agricultural land loss in other districts.

The drastic shrinkage of water area correlated closely with urban river-filling, river channel hardening, and climate change. Studies indicate that 62% of Shanghai’s water loss during 2000–2015 resulted from river burial [47], particularly evident in Pudong New Area—a focal point for urbanization. In Chongming, rapid water decline was likely linked to sediment deposition in the Yangtze Estuary and land reclamation [48].

Ecological land fragmentation reflected the insufficient prioritization of ecological principles in urban planning. Although afforestation projects briefly increased woodland coverage around 2010, subsequent urban expansion again compressed ecological spaces. This “increase-then-decrease” pattern demonstrates that ecological restoration measures lacking long-term institutional safeguards struggle to withstand urbanization pressures [49].

4.2. LUCC Under Different Scenarios

The changes and spatial distribution of land use in Shanghai by 2040 are primarily shaped by the combined effects of natural geographical conditions, urbanization processes, and ecological conservation policies. Under ND scenarios (ND119, ND245, and ND585), farmland and built-up land dominate as predominant land types, with their spatial distribution closely tied to agricultural protection policies and urban expansion [50]. Farmlands are primarily concentrated in northern Chongming District and southern districts like Jinshan and Fengxian. This pattern stems from both the superior natural conditions—including the presence of alluvial plains with fertile soils suitable for agricultural production—and the fact that their suburban locations are distant from the city center, therefore facing relatively lower urbanization pressure. Built-up land predominantly clusters in central urban areas and adjacent zones (e.g., Pudong New Area), reflecting demands from urban sprawl and associated infrastructure development. This pressure is particularly intensified under the high-emission ND585 scenario, where climate warming may exacerbate land development [51]. In contrast, ecological protection scenarios (EP119, EP245, and EP585) restrain built-up land expansion through policy interventions while increasing farmland area, demonstrating that planning measures like ecological conservation redlines can effectively curb disorderly urban encroachment [52]. The distribution of woodlands and water area adheres to natural constraints: woodlands scatter sporadically across western Qingpu District’s low hilly terrain, while water area aligns with the Yangtze and Huangpu River systems. Their area variations mainly result from strengthened wetland conservation efforts under ecological protection scenarios. From the perspective of administrative division, Chongming district and Pudong New Area have obvious characteristics. Chongming district’s prominent farmland proportion aligns with its ecological island positioning and agricultural focus, while Pudong New Area’s high built-up land coverage underscores its role as an economic core. The scarcity of bare land and grassland areas reflects the degradation of natural vegetation cover in this highly urbanized region [53].

4.3. ESs During 2000–2020

Figure 9 illustrates the spatial dynamics of Shanghai’s ESs during 2000–2020. Water yield showed an upward trend during this period. The primary reason is the expansion of impervious surfaces (e.g., paved roads, rooftops) resulting from urbanization-driven built-up land expansion. These surfaces impede rainwater infiltration, rapidly converting precipitation into surface runoff and significantly increasing water yield. Concurrently, adjustments in precipitation patterns due to climate change—such as increased extreme precipitation events—also contributed to this growth [54].

Soil retention exhibited a pattern of initial decline followed by recovery. Rapid urbanization encroached upon ecological lands like woodlands and grasslands, reducing vegetation coverage and diminishing resistance to soil erosion. The subsequent rebound stemmed from the implementation of ecological restoration projects, which partially restored the soil retention function of ecosystems, resulting in an overall upward trend. Areas showing increased soil retention were mainly concentrated in northwestern Chongming Island, where favorable natural ecosystems support soil retention.

Carbon sequestration displayed a decline in the early period followed by stabilization. The initial decrease was primarily caused by urban development encroaching on carbon-sink lands such as woodlands and wetlands, weakening carbon absorption and storage capacity. Later stabilization was achieved through urban greening initiatives (e.g., new parklands, forest management), which enhanced carbon sinks and partially offset urbanization-induced losses. This demonstrates the compensatory and enhancing effects of ecological restoration policies on carbon sequestration [55]. Carbon sequestration increases were predominantly observed in northern Chongming Island and coastal areas of Pudong New Area.

Habitat quality demonstrated a continuous downward trend. Urban expansion gradually fragmented and encroached upon natural habitat patches, reducing habitat connectivity and weakening ecosystem resilience to disturbances. Spatially, habitat degradation was most pronounced in northern Chongming Island and eastern coastal areas of Pudong New Area. While extensive green space construction and artificial forests significantly boosted regional carbon sinks, urbanization-driven infrastructure development (e.g., road networks) simultaneously caused habitat fragmentation and environmental pollution [56]. This reduced biodiversity and ecosystem integrity, ultimately leading to habitat quality decline.

4.4. Evaluation of ESs Under Different Scenarios

The spatial heterogeneity of ESs in Shanghai by 2040 under different scenarios will be primarily influenced by the dual effects of climate change–land use interactions and ecological conservation policies. Under the SSP119 scenario, Shanghai’s average water yield peaks at 645.32 mm, while it drops to its lowest of 393.17 mm under SSP245. This decline is primarily attributed to altered precipitation patterns and intensified urban heat island effects under this moderate-emission scenario, leading to significantly reduced water yield. This aligns with IPCC conclusions on hydrological stability under mild climate change scenarios and validates the negative impact of moderate-emission scenarios on hydrological regulation services [57].

Shanghai’s soil retention under SSP119 substantially exceeds that under SSP245 and SSP585, reaching its highest level particularly in EP119. This results from the synergistic effects of relatively mild climatic conditions and proactive ecological conservation policies. Relevant studies confirm that moderate and evenly distributed rainfall under SSP119, combined with increased forest coverage and wetland restoration in ecological protection scenarios, significantly enhances vegetation’s capacity for rainfall interception and root-based soil stabilization [58]. In contrast, soil retention is lowest under ND585 due to extreme precipitation compounded by urban expansion. Research indicates that the increased frequency of extreme precipitation events under SSP585 intensifies raindrop splash and runoff erosion [59]. Under natural development, continuous urban expansion—particularly in traditional agricultural areas like Pudong New Area and Jiading District—reduces farmland and forested areas critical for soil retention [60].

Variations in Shanghai’s carbon sequestration across scenarios are mainly driven by land use changes and climate-induced shifts in vegetation productivity. Carbon sequestration peaks under EP245 among the six scenarios, benefiting from moderate climate change pressures under SSP245 that maintain high net primary productivity (NPP), alongside stringent ecological policies enhancing vegetation carbon sequestration. Studies demonstrate that suburban park development and the 25% forest coverage target elevate carbon sequestration in specific areas, with nature-based solutions (NbS) in wetland restoration contributing an additional 12% carbon sink [61]. This underscores the significant synergistic effects of ecological conservation policies on ESs.

Habitat quality reaches its highest level under ND245, with an average index of 0.32, owing to optimal climatic conditions and the spatial self-organization effect of urban expansion. Research suggests that SSP245 provides more stable habitats than SSP119 (limited by potential vegetation productivity) or SSP585 (under stress from extreme heatwaves) [62]. Furthermore, SSP245 restricts intensive construction along transit corridors [63], preserving continuous habitat patches in Qingpu–Songjiang’s western ecological corridor and northern Chongming Island—key factors contributing to ND245’s superior habitat quality. This demonstrates how synergistic interplay between moderate climate pressures and scientific spatial planning sustains high habitat quality, effectively enhancing urban ecosystem stability.

4.5. Limitation of the Current Study

This study reveals the impacts of land use change and climate change on ESs in Shanghai, providing insights for the city’s ecological conservation and sustainable development. However, certain limitations should be acknowledged. Firstly, while the study evaluated four major ecosystem services, it did not encompass cultural services within ecosystem services, overlooking the landscape esthetic value and tourism benefits provided by urban ecosystems. Secondly, predictive models for land use change still require further refinement. Most simulation studies based on CA, such as the PLUS model, fail to account for the magnitude or quantitative differences in the impacts of spatial suitability, neighborhood aggregation effects, and random factors on land use change. Consequently, they do not fully elucidate the driving and inhibitory mechanisms between different land use types [64]. Future research should focus on uncovering interaction networks among various land use types based on simulation consistency and enhancing the predictive capability for future land use changes. Additionally, it is necessary to consider Shanghai’s uniqueness as a coastal megacity and further develop a more comprehensive ES assessment framework. This will provide more holistic and detailed support for ecosystem management and decision-making, assisting coastal cities in achieving a balance between ecological conservation and socioeconomic development in the future.

5. Conclusions

This study investigates land use changes and their impacts on ESs in Shanghai from 2000 to 2020, subsequently employing the PLUS model to simulate 2040 land use projections under six integrated scenarios combining two land planning modes (ND and EP) with three climate pathways (SSP1-1.9, SSP2-4.5, SSP5-8.5). Through controlled-variable simulations, we predicted the spatial distribution of four key ESs—water yield, soil retention, carbon storage, and habitat quality—providing scientific references for Shanghai’s ecological land management and urban sustainable development planning. The key findings are as follows:

(1)

During 2000–2020, Shanghai’s land use was dominated by built-up areas and farmland, exhibiting marked transitions: farmland proportion declined continuously while built-up land expanded substantially. Water area decreased significantly, with woodland and grassland maintaining minimal, stable shares. Spatially, a “core-suburban-peripheral” gradient emerged: intensive development in urban cores, farmland-built-up mosaics in suburbs, and relatively preserved yet encroached farmland in the peripheries. The expansion of construction land brought by urbanization has led to the increasing fragmentation of ecological land.

(2)

PLUS model simulations of Shanghai’s 2040 land use under six scenarios indicate that natural development scenarios sustain built-up/farmland dominance—especially under ND585, where built-up expansion accelerates and water area decline, while woodland increases slightly due to enhanced climate adaptability. Ecological protection scenarios effectively constrain built-up growth, elevate farmland proportion, and stabilize woodland/water areas by counterbalancing climate impacts through policy interventions. Spatially, farmland concentrates in central–southern zones, built-up land sprawls eastward, and woodland/water is distributed fragmentedly. Ecological policies demonstrably curb urban sprawl and maintain ecological land stability.

(3)

From 2000 to 2020, regional water yield increased, soil retention declined then gradually recovered, carbon sequestration fluctuated mildly, and habitat quality deteriorated persistently. Spatially, high-value areas for water yield and soil retention clustered in peri-urban ecological zones, contrasting with low values in urban cores and developed areas—revealing an “urban-rural divergence” pattern that reflects urbanization’s profound ecosystem impacts.

(4)

The projected 2040 ESs vary substantially across scenarios: water yield and soil retention will peak under SSP119, while carbon sequestration and habitat quality will be optimal under SSP245, underscoring climate adaptation’s critical role in enhancing ESs. Spatially, water yield will peak in the northeast of the urban center where carbon storage, soil retention, and habitat quality decline—inversely mirroring southwestern patterns. This northeast disparity stems from excessive built-up expansion, indicating economic overreliance on land development. Future strategies should diversify eco-economic models, rigorously safeguard farmland and ecological redlines, and rationally utilize undeveloped land to advance high-quality regional coordination.

(5)

District-level ESs in 2040 scenarios show pronounced heterogeneity: water yield will be optimal under SSP119, with Qingpu District becoming the lowest-value area due to rapid urbanization. Soil retention will be minimized in Jing’an District, while Minhang and Baoshan will achieve higher values via ecological restoration under ND245/EP245. Carbon sequestration will exhibit a “core-periphery” gradient, and ecological advantage zones like Pudong and Chongming will maximize their potential under EP245, whereas central urban districts (e.g., Xuhui) will show modest gains under ecological protection. Habitat quality will prove most scenario-sensitive in Qingpu (peaking under ND245), while suburban differences will gradually converge, indicating the delayed effects of conservation measures. Overall, peripheral regions with favorable ecological conditions exhibit superior ESs functionality. Moreover, in these areas, ESs demonstrate heightened sensitivity to climate change and policy interventions. Conversely, highly urbanized central districts manifest generally diminished ESs, where responses to climate change and policies remain markedly subdued. This contrast thereby underscores the critical importance of spatial heterogeneity management in urban development.