The Spatiotemporal Dynamics and Evolutionary Relationship Between Urbanization and Eco-Environmental Quality: A Case Study in Hangzhou City, China

Abstract

The rapid expansion of urban spaces driven by accelerating urbanization has profoundly impacted the eco-environmental quality. However, the dynamic relationship between urbanization and eco-environmental quality remains insufficiently understood. This study quantifies urbanization intensity and eco-environmental quality using the impervious surface distribution density (ISDD) and Remote Sensing-based Ecological Index (RSEI). By examining the spatiotemporal dynamics and evolutionary relationships of these indicators in Hangzhou from 1985 to 2020, we found that urban expansion drove ecological degradation in expansion areas, whereas ecological quality in the old city significantly improved. The ecological response to urbanization intensity exhibited spatial variation: in low-intensity urbanized expansion areas, ecological quality declined with increasing urbanization, whereas in the high-intensity urbanized old city, ecological quality improved. Additionally, the degree of coupling coordination between urbanization and ecological quality steadily increased over time, underscoring the importance of rational urban planning and ecological management in achieving sustainable development. This study provides a scientific foundation for urban ecological environment management and offers practical insights for fostering green development in rapidly urbanizing regions.

1. Introduction

Urbanization is a defining feature of modern society, driving economic growth while significantly affecting ecosystems and the eco-environment [1,2]. The urban ecological environment, a critical foundation for human activities, is intricately linked to public health and daily life [3,4]. Rapid urbanization, characterized by land use changes and intensified human activities, has brought about various ecological and environmental challenges, including reduced vegetation coverage, urban heat island effects, water shortages, biodiversity loss, and ecosystem degradation [5,6,7,8,9]. The effects of urbanization on the ecological environment unfold gradually over time, exhibiting substantial variability across spatial and temporal scales. Understanding this complex relationship has become a key focus in contemporary research.

Urbanization evaluation involves multiple levels of indicators and methods, with the indicator system being the core of urbanization research. Urbanization intensity, which reflects changes in land use, population density, and economic activities, is a critical metric for quantifying the extent of urbanization development [10,11]. The impervious surface ratio (ISR) is one of the most traditional methods for assessing urbanization intensity. It calculates the total proportion of impervious surfaces in a specific area, providing a basic measure of urban expansion [12,13,14]. The Mixed Pixel Decomposition (MPD) method represents a more advanced approach in urbanization intensity assessment. Unlike the ISR, which relies on single-pixel classification, MPD addresses the mixed pixel problem that arises in low-resolution remote sensing data, providing a more precise estimate of the impervious surface abundance [15,16,17]. However, ISR and MPD do not account for the spatial distribution or the neighborhood effects of impervious surfaces, which can limit their applicability in understanding the ecological impacts of urbanization. While effective at providing a general measure of urban development, they do not offer insights into the spatial aggregation or clustering of impervious surfaces [18,19,20]. In contrast to the ISR and MPD, the impervious surface distribution density (ISDD) proposed by Meng et al. [21] accounts for the spatial aggregation of impervious surfaces by incorporating distance weighting. This method reflects how the spatial configuration of impervious surfaces can influence the pressure of urbanization on ecosystems. It is particularly useful for analyzing the neighborhood effects and spatial heterogeneity of urban growth, making it a more detailed indicator for evaluating urbanization’s ecological effects [22,23].

Ecological quality evaluation is essential for assessing the health and sustainability of ecosystems. Ecological indices have become one of the primary tools for evaluating ecological quality. Initially, simple ecological indices were introduced, using single ecological factors such as vegetation indices or land surface temperature to evaluate the ecological environment [24,25,26]. However, they often fail to capture the complexity of ecological systems and provide a comprehensive assessment of ecological quality [27,28]. To address this limitation, comprehensive ecological indices have been developed and applied to analyze ecological environment changes. Notably, Xu et al. [29] introduced the Remote Sensing-based Ecological Index (RSEI), which integrates multiple ecological indicators. As a straightforward yet effective metric for evaluating regional ecological conditions, RSEI has been widely applied in urban ecological studies [30]. For instance, Xu et al. [31] evaluated Fuzhou’s urban ecological quality from 2001 to 2016 using RSEI, while Huang et al. [32] monitored ecological changes in Lhasa, and Firozjaei et al. [33] evaluated and compared the ecological equality in six cities of the U.S.A utilizing RSEI.

The impact of urbanization on ecological systems is a global issue, and numerous studies from different regions offer valuable insights. In Europe, Anestis et al. [34] discuss the pressures of urbanization across the continent, highlighting the challenges of managing urban growth in ecologically sensitive areas. Similarly, Sahoo et al. [35] explored the impact of urbanization on ecological footprints and air quality, with a focus on newly industrialized countries. Almulhim et al. [36] have examined how urbanization drives land use changes in the Middle East, influencing ecological sustainability. Li et al. [37] analyzed the spatiotemporal relationship between urbanization and the eco-environment in Kashgar, China, and found that urban expansion led to decreased vegetation and environmental quality. Previous studies have established significant correlations between urban expansion and environmental degradation, including reduced air and water quality and biodiversity loss [38,39]. However, recent research also highlights the potential positive effects of urbanization, such as improved ecological quality through sustainable urban planning, green infrastructure, and environmental policies [40,41]. The emerging concept of “eco-friendly urbanization” advocates for development that balances economic growth and environmental protection [42]. Measures such as expanding urban green spaces, adopting renewable energy, and enforcing pollution controls have shown promise in improving urban ecological quality [43,44,45].

Despite these advances, several research gaps persist. Firstly, the complex interactions between urbanization and the ecological environment have not been fully elucidated. Most studies focus on the negative impact of urbanization on ecological quality, while neglecting the indirect positive effects brought by human management practices [46]. Capturing this nonlinear relationship between urbanization and ecological quality requires advanced remote sensing data and quantitative models. Secondly, the heterogeneity and variability of urbanization impacts across spatiotemporal scales have been insufficiently explored. Many studies focus on city-wide changes, neglecting differences across urbanization stages and intensity gradients [47,48]. A deeper understanding of the similarities and differences in ecological responses across areas with varying urbanization intensities is needed. Finally, there is limited research on ecological restoration potential in rapidly urbanizing regions, particularly strategies for achieving ecosystem restoration alongside urban expansion [49,50]. Addressing these issues is crucial for designing sustainable urban development strategies that balance growth with environmental preservation.

To address the complex relationship between urbanization and ecological quality, we propose an integrated framework that combines ISDD and RSEI. ISDD provides a refined spatial measure of urbanization intensity by incorporating the clustering effect of impervious surfaces, while RSEI offers a comprehensive evaluation of ecological quality based on greenness, wetness, dryness, and surface temperature. We investigate the dynamic relationship between urbanization and ecological quality in Hangzhou by leveraging time-series Landsat imagery from 1985 to 2020 and applying correlation analysis and the coupling coordination degree (CCD) model. Moreover, we analyze the differentiated ecological responses across the old city, expansion areas, and suburban areas, which enables a deeper understanding of the coupling coordination mechanisms between urban growth and environmental sustainability. This study aims to achieve the following:

(1)

Integrate ISDD and RSEI to establish a spatiotemporally explicit framework for assessing the ecological effects of urbanization;

(2)

Conduct a long-term, high-resolution analysis of urbanization and ecological changes in Hangzhou from 1985 to 2020 using 30 m Landsat imagery;

(3)

Reveal spatially differentiated ecological responses across the old city, expansion areas, and suburban areas, and quantify their coordination level, providing scientific support for sustainable urban development.

2. Materials and Methods

2.1. Study Area

Hangzhou (118°21′–120°30′E, 29°11′–30°33′N), located in Zhejiang Province, is a key city in the Yangtze River Delta, renowned for its historical and cultural significance as well as its scenic and tourism attractions. Covering a total land area of 16,850 km², the city’s topography is predominantly flat in the northeast and east, while mountainous regions dominate the west and south. The city experiences a subtropical monsoon climate, characterized by distinct seasons and abundant rainfall.

Hangzhou’s population has grown significantly, from 5.43 million in 1985 to 12.37 million by the end of 2022. In the same year, the city’s Gross Domestic Product (GDP) reached 1.8753 trillion CNY, accounting for 24.1% of Zhejiang Province’s total GDP (from Hangzhou Statistics Bureau, https://tjj.hangzhou.gov.cn/, accessed on 21 September 2023).

Over the past three decades, rapid socioeconomic development has driven multiple adjustments to Hangzhou’s administrative divisions. According to the city’s Comprehensive Land Use Plan (2006–2020) (from Hangzhou Bureau of Planning and Natural Resources, http://ghzy.hangzhou.gov.cn/, accessed on 21 September 2023), Hangzhou consists of one central urban area—comprising Xihu (Xh) district; Yuhang (Yh) district; Gongshu (Gs) district; Xiaoshan (Xs) district; Jianggan (Jg) district; Binjiang (Bj) district; Shangcheng (Sc) district; Xiacheng (Xc) district—along with three county-level cities (Fuyang, Linan, Jiande), and two counties (Tonglu, Chun’an). Given the advanced urbanization of the central urban area, this study focuses on these districts (Figure 1) to examine the effects of urban expansion on the ecological environment.

2.2. Datasets and Pre-Processing

This research relies on two key datasets: land use data and remote sensing imagery. The land use information, sourced from the China Land Cover Dataset (CLCD), was employed to extract impervious surface distribution and calculate the ISDD. Meanwhile, remote sensing imagery, obtained from the Landsat Collection 2 Level 2 (C2L2) products, was applied in the computation of the RSEI.

2.2.1. Impervious Surface Data

Impervious surfaces are a critical component of the terrestrial surface, effectively representing urbanization processes [12]. In this study, impervious surface data were derived from the CLCD, a dataset developed by Professor Huang Xin’s research team at Wuhan University. This dataset, based on Landsat data, provides land use information for China from 1985 to 2020 at a 30 m spatial resolution, with an overall accuracy of 79.31%. The dataset is freely accessible online (https://doi.org/10.5281/zenodo.4417810, accessed on 21 September 2023) [51]. The CLCD categorizes land surfaces into nine classes: impervious surfaces, forest, grassland, water bodies, shrubland, bare land, farmland, wetlands, and snow/ice. For this study, the impervious surface category was extracted for Hangzhou, and the ISDD was calculated to quantify urbanization intensity.

2.2.2. Landsat Data

Landsat provides continuous, long-term surface observation data, making it ideal for monitoring land cover and ecological changes over extended periods. The Landsat C2L2 dataset utilized in this study comprises land surface temperature data and land surface reflectance data, acquired from the TM, ETM+, and OLI sensors aboard the Landsat 5, 7, and 8 satellites, which were sourced from the USGS. The dataset spans 1985 to 2020 and consists of 342 scenes from Landsat 5 (1985–2011), 11 scenes from Landsat 7 (2012), and 121 scenes from Landsat 8 (2013–2020) (Figure 2). To ensure accurate and consistent results, Landsat images were carefully selected and processed as follows: Firstly, images from April to September were selected to minimize variations in solar angles and vegetation phenology during the growing season. Secondly, cloud and shadow pixels were masked using the CFMASK algorithm. Thirdly, Median Compositing was applied to create annual mosaics, using the median value of each pixel across available images to minimize cloud effects and preserve band relationships. For Landsat 7 ETM+ data, affected by SLC failure, time-series interpolation was used to fill in the missing pixels. Linear interpolation between adjacent years was applied, ensuring temporal continuity. All preprocessing steps were performed on the Google Earth Engine (GEE) platform.

2.3. Methods

The overall process of this study is shown in Figure 3. First, based on the CLCD, the impervious surface distribution data of Hangzhou from 1985 to 2020 were extracted, and the ISDD, which represents urbanization intensity, was calculated. The K-means clustering algorithm was then applied to extract the urban main built-up area (UMBA) of Hangzhou, dividing the city into three zones: the old city, expansion areas, and suburban areas. Next, based on the Landsat C2L2 dataset, four indices representing urban surface biophysical characteristics (NDVI, WET, NDBSI, LST) were calculated. The Principal Component Analysis (PCA) method was used to construct the ecological index (RSEI) representing the surface ecological environment. Finally, the spatiotemporal distribution characteristics and evolution trends of ISDD and RSEI in Hangzhou from 1985 to 2020 were analyzed. Additionally, through correlation analysis and the CCD model, the spatiotemporal response and coupling coordination relationship between RSEI and ISDD were examined.

2.3.1. Calculation of ISDD

ISDD quantifies the extent of artificial impervious surfaces—such as buildings, roads, and parking lots—within a specific area, making it a key indicator of urbanization intensity and land use change [52]. Unlike traditional methods, the distance-weighted ISDD accounts for the spatial clustering of impervious surfaces, thereby offering a more precise reflection of their local impact on the ecological environment [53]. In this method, the ISDD value for a given pixel is determined by the density and spatial arrangement of impervious surfaces within a defined radius around it. The distance from each impervious surface to the central pixel serves as a weight, with closer surfaces receiving higher weights. The calculation formula is expressed as follows [21]:

$${\mathrm{I}\mathrm{S}\mathrm{D}\mathrm{D}}_{\mathrm{s}}\left(\mathrm{r}\right)={\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{B}}_{\mathrm{s}\mathrm{i}}\ast \left(1-\frac{{\mathrm{D}}_{\mathrm{i}}}{2\mathrm{r}}\right)}{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}\left(1-\frac{{\mathrm{D}}_{\mathrm{i}}}{2\mathrm{r}}\right)}}$$

(1)

where s denotes the central pixel; and r represents the specified neighborhood radius; Bsi corresponds to the pixel value within radius r (where impervious surface pixels are assigned a value of 1, while pervious surface pixels are assigned a value of 0); and Di indicates the distance between the central pixel s and the ith pixel.

ISDD is sensitive to both the spatial resolution of input data and the selected neighborhood radius. At coarser resolutions, fine-scale variations in impervious surface clustering may be lost. Additionally, the clustering pattern can vary depending on the chosen neighborhood radius. In this study, we used impervious surface distribution data extracted from Landsat imagery at a 30 m resolution for ISDD calculation. Based on existing studies [22,23], we selected 250 m (approximately 8 pixels) and 500 m (approximately 16 pixels) as the neighborhood radii, evaluating ISDD sensitivity through spatial distribution and statistical characteristics at each radius.

The value range of ISDD is 0–1, where higher values indicate a higher aggregation density of impervious surfaces. To analyze the spatial variation in ISDD, it was classified into five categories: low (0–0.2), relatively low (0.2–0.4), medium (0.4–0.6), relatively high (0.6–0.8), and high (0.8–1).

2.3.2. Extraction of UMBA Based on ISDD

Accurately extracting the UMBA is essential for analyzing urban spatial patterns [54]. A key characteristic of UMBA is the concentrated distribution of impervious surfaces. However, built-up areas also encompass other land use types, including parks, vegetated areas, and water bodies. Additionally, fragmented impervious surfaces can extend into suburban areas, far from the urban core. Simply extracting the UMBA based on impervious surface distribution alone does not guarantee the continuity of the built-up area. In contrast, ISDD-based methods can assess the concentration of urban impervious surfaces while maintaining the integrity of land cover classifications [23]. Common techniques for extracting urban boundary information include the threshold method and cluster analysis [55]. However, variations in physical environments and socio-economic conditions across regions make a single threshold method unreliable for extracting built-up area on a large scale [8]. The K-means clustering algorithm, known for its objectivity, simplicity, and reliability, is frequently used for this purpose [56,57]. Accordingly, this study utilized the K-means clustering method to extract the UMBA, with the results validated against statistical data released by the Ministry of Housing and Urban-Rural Development of the People’s Republic of China (MOHURD).

2.3.3. Urban Spatial Morphology

Examining the spatial dynamics of urban expansion is essential for efficient urban planning and management, contributing to sustainable development and ecological conservation. Utilizing accurate boundary data of the UMBA, we measured spatial morphological changes in Hangzhou during the urbanization process from three dimensions: quantity, intensity, and external structure. These dimensions are represented by four indicators: urban expansion area (UEA), urban expansion intensity (UEI), fractal dimension (FD), and compactness (CP). UEA reflects the scale and speed of urban development, indicating changes in land resource utilization [58]. UEI captures the rate and extent of urbanization, helping policymakers assess the potential ecological impacts of urban development [59]. FD provides a quantitative measure of urban morphology’s complexity, reflecting the diversity and heterogeneity of the city. A higher fractal dimension suggests more disordered urban morphology and greater expansion potential [60]. CP measures the spatial concentration of urban form, reflecting the efficiency of urban land use. Higher compactness usually implies more centralized land use and better accessibility to public facilities and services [61]. The following formulas were used to compute these indicators:

$$UEA=U_b{-U}_a$$

(2)

$$UEI={\displaystyle \frac{UEA}{TA}}\ast {\displaystyle \frac{1}{T}}\ast 100\mathrm{\%}$$

(3)

$${FD}_{it}=2\ast \ln \left(0.25\ast P_{it}\right)/\ln A_{it}$$

(4)

$${CP}_{it}=2\ast \sqrt{\pi A_{it}}/P_{it}$$

(5)

In these formulas, UEA represents the expansion area, and U_b and U_a represent the built-up area at the end and the beginning of the study interval. UEI represents the urban expansion intensity index, TA represents the overall spatial extent of the study area, and T represents the duration of the study period. Additionally, FD_it, P_it, CP_it, and A_it correspond to the FD, perimeter, compactness, and size of the ith built-up area patches in year t. The FD value ranges from 1 to 2, where a higher value indicates a more complex and irregular urban form, while a lower value suggests a simpler and more regular boundary.

2.3.4. Calculation of RSEI

This study employed the RSEI proposed by Xu et al. [62] to evaluate the eco-environmental quality of Hangzhou from 1985 to 2020. RSEI is a comprehensive indicator derived from remote sensing data, integrating four key factors: greenness, moisture, heat, and dryness (Equation (6)) [63]. Greenness is indicated by the normalized difference vegetation index (NDVI), a widely recognized indicator closely linked to vegetation biomass [64]. Moisture is determined by the third component of the tasseled cap transformation [65,66]. Heat is represented by the surface temperature obtained from the Landsat Collection 2 Land surface temperature (LST) product [67], which was produced by the Earth Resources Observation and Science (EROS) Center with a spatial resolution of 30 m. Dryness is mainly associated with built-up and barren land and is quantified through the normalized difference built-up and soil index (NDBSI), which is synthesized from the Index-Based Built-Up Index (IBI) and Soil Index (SI) [68].

Table 1. Surface biophysical parameters used in RSEI.

Index	Explanation	Equation
NDVI	Represents vegetation cover information	$NDVI={\displaystyle \frac{NIR-Red}{NIR+Red}}$
Wet	Represents land surface humidity	Tasseled cap transformation (TCT) component 3
NDBSI	Represents land surface dryness	NDBSI = (IBI + SI)/2 $IBI={\displaystyle \frac{\frac{2SWIR1}{SWIR1+NIR}-\left(\frac{NIR}{NIR+Red}+\frac{Green}{Green+SWIR1}\right)}{\frac{2SWIR1}{SWIR1+NIR}+\left(\frac{NIR}{NIR+Red}+\frac{Green}{Green+SWIR1}\right)}}$ $SI={\displaystyle \frac{\left(SWIR1+Red\right)-\left(NIR+Blue\right)}{\left(SWIR1+Red\right)+\left(NIR+Blue\right)}}$
LST	Represents land surface temperature	Landsat collection 2 LST product (single-channel algorithm)

Note: Red, green, blue, NIR, and SWIR1 correspond to the respective spectral band values—red, green, blue, near-infrared, and short-wavelength infrared 1—in Landsat imagery.

summarizes the four components of RSEI: NDVI (greenness), wet (moisture), NDBSI (dryness), and LST (heat).

PCA is a dimensionality reduction method that identifies a subset of key variables from a larger set through linear transformation, effectively addressing the issue of collinearity among variables. PCA compresses multidimensional information into a few principal components by successively rotating the coordinate axes perpendicularly. Each principal component typically represents specific feature information, with the first component (PC1) capturing the majority of the characteristic information from all variables [69,70]. In this study, PC1 derived from PCA was utilized to construct the RSEI (Equation (7)).

Since the eigenvector direction in PCA is not unique, the choice of eigenvectors can influence the resulting RSEI. To address this issue, the moisture component, which is less affected by seasonal variations, was selected to define the eigenvector direction of the PC1 [71]. The final RSEI values were then scaled to a normalized range of [0, 1]. To assess spatial differences in ecological quality, RSEI values were evenly divided into five categories at 0.2 intervals: poor, fair, moderate, good, and excellent.

$$RSEI=f(Greenness,Moisture,Heat,Dryness)$$

(6)

$$RSEI=\left\{\begin{array}{c}PC1\left[f\left(NDVI,Wet,LST,NDBSI\right)\right],V_{Wet}\ge 0\\ 1-PC1\left[f\left(NDVI,Wet,LST,NDBSI\right)\right],V_{Wet}<0\end{array}\right.$$

(7)

2.3.5. Method for Analyzing Spatiotemporal Evolution Trends

This study employed a linear regression model to examine the spatiotemporal trends of urbanization intensity and ecological quality at the pixel scale. Furthermore, the F-test was applied to determine the statistical significance of these trends [72]. The relevant calculation formulas are provided below:

$$slope={\displaystyle \frac{n\ast {\sum }_{i=1}^{n}i\ast V_i-{\sum }_{i=1}^{n}i{\sum }_{i=1}^n{V}_i}{n\ast {\sum }_{i=1}^n{i}^2-{\left({\sum }_{i=1}^{n}i\right)}^2}}$$

(8)

$$F={\displaystyle \frac{U\ast \left(n-2\right)}{Q}}$$

(9)

In Equation (8), the slope represents the trend of change, where a positive value (slope > 0) indicates an upward trend in ISDD and RSEI, while a negative value (slope < 0) signifies a declining trend. The parameter n denotes the total number of monitoring years; i corresponds to a specific year (i = 1, 2, …, n); V_i represents the value of variable V (either ISDD or RSEI) in the ith year. In Equation (9), the sum of the squared errors is given by $U={\sum }_{i=1}^n{\left(\widehat{V_i}-\overline{V_i}\right)}^2$, while the explained sum of squares is represented as $Q={\sum }_{i=1}^n{\left(V_i-\widehat{V_i}\right)}^2$. Here, $\overline{V_i}$ represents the mean value of variable V, and $\widehat{V_i}$ denotes its regression-estimated value.

This study classified the spatiotemporal evolution of ISDD and RSEI into five categories by analyzing both the trend slope and F-test results: significantly reduced/degraded (p < 0.01 and slope < 0), reduced/degraded (0.01 < p < 0.05 and slope < 0), basically stable (p > 0.05), increased/improved (0.01 < p < 0.05 and slope > 0), significantly increased/improved (p < 0.01 and slope > 0).

2.3.6. Spatiotemporal Analysis of RSEI Responses to ISDD

The Pearson Correlation Coefficient (PCC) was employed to investigate the relationship between RSEI and ISDD over the period from 1985 to 2020. To determine the statistical significance of the correlation, a t-test was conducted [73]. This analysis aimed to reveal how the ecological environment responds to changes in urbanization intensity over time and space. The corresponding calculation formula is presented below:

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

(10)

In Equation (10), r denotes the PCC between variables x and y, with values ranging from −1 to 1. A value of r approaching 1 signifies a strong positive correlation, while a value close to −1 implies a strong negative correlation. If r equals 0, it indicates no correlation between the two variables. The parameter n represents the total number of monitoring years, and i refers to a specific year (i = 1, 2, …, n); The terms x_i and y_i represent the values of variables x and y in the ith year, while $\overline{x}$ and $\overline{y}$ indicate their respective mean values.

2.3.7. CCD Model

The CCD model is commonly used to quantify the interactions between multiple systems [74]. In this study, it was employed to explore the coupling and coordination mechanism between urbanization intensity and the ecological environment. While the CCD model effectively evaluates the extent of coupling and coordinated development between two systems [75,76], it does not provide insight into their relative development status [77]. To address this limitation, the relative development index was incorporated to assess the comparative development levels of urbanization and ecological quality. The formulas for both the CCD model and the relative development index are presented as below:

$$C={\left\{{\displaystyle \frac{F_1\times F_2}{{\left[\left(F_1+F_2/2\right)\right]}^2}}\right\}}^{{\displaystyle \frac{1}{2}}}$$

(11)

$$T={\alpha }_1{F}_1+{\alpha }_2{F}_2$$

(12)

$$CCD=\sqrt{CT}$$

(13)

$$R={\displaystyle \frac{F_1}{F_2}}$$

(14)

In the above formulas, C represents the coupling degree; while F₁ and F₂ correspond to the ISDD and the RSEI, respectively; the comprehensive coordination index of F₁ and F₂ is represented by T, where α₁ and α₂ serve as their respective weights, ensuring that α₁ + α₂ = 1. In this study, the ecological environmental quality and urbanization intensity were considered equally significant, leading to the assignment of equal weights (α₁ = α₂ = 0.5). The CCD quantifies the degree of coordinated development between urbanization and ecological environmental quality, with values ranging from 0 to 1. A higher CCD value indicates a more balanced and harmonious relationship between the two systems. Additionally, R is the relative development index. According to the classification of physical coupling stages and related studies [77,78,79], we categorized the CCD between urbanization intensity and ecological environment into 7 classes and 3 broad types.

Table 2. Classification of CCD between ISDD and RSEI.

Category	CCD Range	Classes	R	Type
Uncoordinated development	(0–0.2]	Serious imbalance (SI)	R > 1 R ≈ 1 R < 1	Ecological development lag Harmonized growth Urbanization growth lag
	(0.2–0.4]	Moderate imbalance (MOI)	R > 1 R ≈ 1 R < 1	Ecological development lag Harmonized growth Urbanization growth lag
	(0.4–0.5]	Mild imbalance (MII)	R > 1 R ≈ 1 R < 1	Ecological development lag Harmonized growth Urbanization growth lag
	(0.5–0.6]	Near Coordination (NC)	R > 1 R ≈ 1 R < 1	Ecological development lag Harmonized growth Urbanization growth lag
Transformation development	(0.6–0.7]	Primary Coordination (PC)	R > 1 R≈1 R < 1	Ecological development lag Harmonized growth Urbanization growth lag
	(0.7–0.8]	Moderate Coordination (MC)	R > 1 R ≈ 1 R < 1	Ecological development lag Harmonized growth Urbanization growth lag
Coordinated development	(0.8–1.0]	Good Coordination (GC)	R > 1 R ≈ 1 R < 1	Ecological development lag Harmonized growth Urbanization growth lag

presents the classification results.

3. Results

3.1. Spatiotemporal Distribution Characteristics of Impervious Surfaces

Impervious surface data for Hangzhou from 1985 to 2020 were derived from the CLCD. The spatial distribution of impervious surfaces at five-year intervals is illustrated in Figure 4. During the urbanization process, impervious surfaces exhibited a radial expansion pattern, spreading outward from the city center toward the surrounding areas.

A statistical analysis of the interannual changes in impervious surfaces (Figure 5) reveals that the total impervious surface area increased from 198.19 km² in 1985 to 1250.57 km² in 2020, reflecting a growth of approximately 6.3 times. The dynamics of impervious surface expansion can be categorized into three distinct phases based on annual changes in surface area, with respective slopes of 26.01, 44.61, and 21.04 for each stage. During the first stage (1985–1995), the impervious surface area increased slowly and was primarily concentrated in peri-urban zones surrounding the central city. During the second stage (1996–2013), the impervious surface area experienced rapid growth, with impervious surfaces spreading quickly into suburban areas. During the third stage (2014–2020), the impervious surface area showed a slower growth, predominantly in the outer suburbs.

3.2. Impervious Surface Distribution Density and Urban Spatial Morphology

The ISDD was derived from the impervious surface distribution data (Figure 4) and is presented in Figure 6. At both r = 250 m and r = 500 m, the ISDD displayed similar distribution characteristics, with values decreasing from the city center toward the outskirts. In the central areas of Hangzhou (including Shangcheng, Xiacheng, Gongshu, Binjiang, Jianggan, as well as the northern part of Xihu and the central part of Xiaoshan), development is relatively concentrated, resulting in ISDD values greater than 0.75. In contrast, the outer regions (such as western Yuhang, southern Xihu, and the western and southern parts of Xiaoshan) exhibited ISDD values between 0.1 and 0.5.

From a spatial distribution perspective (Figure 6), at r = 250 m, ISDD provided a fine-scale reflection of impervious surface distribution; however, due to significant fragmentation, capturing the overall trend remained challenging. In contrast, at r = 500 m, the ISDD patch appeared larger, offering a clearer representation of the large-scale distribution of impervious surfaces in Hangzhou. From a statistical standpoint, as shown in Table S1 in the Supplementary Materials, the coefficient of variation (CV) for ISDD at r = 500 m consistently remained lower than that at r = 250 m. This indicates greater result stability, making it more reliable for capturing the overall urbanization trend. Consequently, ISDD with a 500 m radius was selected as input for UMBA extraction.

Based on the ISDD data, we extracted the UMBA of Hangzhou from 1985 to 2020 and validated the results using statistical data released by the MOHURD. The results are presented in Table S2 of the Supplementary Materials. The maximum relative error between the extracted data and the statistical data is less than 20%, which is within an acceptable margin of error.

Figure 7a shows the boundary of Hangzhou’s main built-up area. To distinguish the spatial patterns of urbanization impacts, the UMBA (representing the urban extent after urbanization in 2020) was segmented into two zones: the OC, which corresponds to the urban area prior to 1985, and the EA, encompassing urban expansion from 1985 to 2020. For comparison, the region outside the UMBA was defined as the SA, covering all areas outside both the OC and EA.

Figure 7b illustrates the changes in UEA and UEI over 35 years. From 1985 to 2010, both expansion area and intensity steadily increased, peaking between 2005 and 2010, with an expansion area of 210.09 km² and an expansion intensity of 1.25%. From 2010 to 2020, both the expansion area and intensity showed a decreasing trend.

Figure 7c,d depicts changes in the spatial morphology of the UMBA. From 1985 to 2020, the FD increased from 1.44 to 1.68, indicating a rise in the complexity of urban morphology. Meanwhile, compactness decreased from 0.35 to 0.10, suggesting a more dispersed spatial distribution. From 1985 to 1991, urbanization was in its early stages, with relatively concentrated urban development and high compactness. From 1992 to 2005, the FD increased significantly, while compactness continued to decline, indicating that rapid urban growth during this period led to a more fragmented urban structure. From 2006 to 2020, both the FD and compactness began to stabilize, indicating that urbanization slowed, and the urban spatial form became more stable during this period.

3.3. Spatiotemporal Distribution Characteristics and Evolution Trend of ISDD

Figure 8 illustrates the spatiotemporal distribution of ISDD in Hangzhou from 1985 to 2020, highlighting significant spatiotemporal heterogeneity. High-density areas are primarily centered in the old city and gradually diffuse toward the suburban areas. As distance from the urban center increases, the ISDD decreases.

Figure 9 presents the areal ratios of ISDD across five categories: high-density, relatively high-density, medium-density, relatively low-density, and low-density, over the period from 1985 to 2020. As urbanization progressed, the proportion of high-density ISDD areas increased substantially, rising from 0.92% in 1985 to 12.03% in 2020, with these areas primarily concentrated in the old city and expansion zones. Medium-density areas expanded from 0.91% in 1985 to 12.21% in 2020, mainly appearing along the periphery of the built-up areas. Conversely, the proportion of low-density ISDD areas decreased from 94.69% in 1985 to 47.72% in 2020, predominantly in the suburban regions.

The mean ISDD (Figure 10a) reflects the spatial distribution of urbanization intensity in Hangzhou from 1985 to 2020, demonstrating a radial decline in ISDD from the old city center outward. Figure 10b,c displays the trend and significance of ISDD changes. A clear increasing trend in ISDD is observed, with 96.2% of the area showing a positive slope (slope > 0), and 73.5% showing a significant increase. In both the old city and suburban regions, urbanization intensity remained relatively stable (0 < slope ≤ 0.04), whereas expansion areas experienced a substantial increase (slope > 0.12). Conversely, only 3.8% of the area exhibited a decrease in ISDD, with 0.1% showing a significant reduction (Figure 10c), primarily in the old city and eastern outskirts.

Figure 11 presents the annual changes in ISDD for the old city, expansion areas, and suburban areas in Hangzhou. The annual variation in ISDD in the old city (Figure 11a) exhibits two distinct phases. From 1985 to 2000 (Phase 1), ISDD in the old city increased rapidly, with a slope of 0.0103. From 2000 to 2020 (Phase 2), the growth rate slowed, and ISDD tended to saturate, with a much smaller slope of 0.0006. In the expansion areas (Figure 11b), ISDD exhibited three distinct stages: From 1985 to 1990, ISDD remained relatively stable. From 1990 to 2013, there was a rapid increase in ISDD, with a slope of 0.0267. From 2013 to 2020, the growth rate slowed down, with a slope of 0.0113. For the suburban areas (Figure 11c), ISDD showed a consistent, slow growth trend with a slope of 0.0048.

3.4. Spatiotemporal Patterns and Evolutionary Trends of Urban Ecological Quality

PCA was applied to construct the RSEI, and the first principal component (PC1) was selected because it explains the majority of the data variance. As shown in Table S3 of the Supplementary Materials, the variance contribution of PC1 exceeded 70% from 1985 to 2020, peaked at 93.75% in 2001, and reached a minimum of 70.66% in 1995. The high contribution of PC1 highlights its effectiveness in capturing key ecological features.

Figure 12 illustrates the spatial distribution of urban ecological quality in Hangzhou from 1985 to 2020, showing significant spatiotemporal heterogeneity. The RSEI analysis indicates that ecological quality in the old city and expansion areas was generally poorer compared to the suburban areas.

Figure 13 presents the statistical analysis of the five RSEI evaluation levels over the past 35 years. The proportion of areas classified as “excellent” exhibited significant fluctuations, peaking at 34.92% in 1985 and reaching its lowest point of 1.96% in 1995. The proportion of “good” areas decreased from 53.96% in 1985 to 26.83% in 2020. The “moderate” category was the most widespread, initially increasing and then decreasing: 9.09% in 1985, peaking at 47.15% in 2000, and falling to 31.01% in 2020. The proportion of “fair” areas increased significantly, rising from 2.01% in 1985 to 27.63% in 2020. Although “poor” areas remained a relatively small fraction, they showed a rising trend, expanding from 0.01% in 1985 to 4.06% in 2020.

To further analyze the spatial distribution and trends of RSEI, its mean value and slope were calculated, as shown in Figure 14. The mean RSEI (Figure 14a) reveals that the ecological quality in Hangzhou’s main built-up areas predominantly fell into the “fair” and “moderate” categories. Regions with “excellent” ecological quality were primarily located in the northwestern and southern suburb areas, while the old city generally exhibited “fair” ecological conditions.

Figure 14b reflects the trend of RSEI changes over time. The results show that 58.1% of areas exhibited a negative slope (slope < 0), indicating a degradation of ecological quality, while 41.9% of areas had a positive slope (slope > 0), suggesting improvement. This indicates that, from 1985 to 2020, the rate of ecological degradation in Hangzhou outpaced the rate of improvement.

Figure 14c shows the significance of the RSEI variation trend. It reveals that 7.4% of areas experienced improvements in ecological quality, with 2.1% showing significant improvements, primarily concentrated in the old city. Conversely, 19.5% of areas experienced degradation, and 9.1% showed significant degradation, predominantly located in the expansion areas.

Figure 15 illustrates the interannual variations in RSEI in the old city, expansion areas, and suburban areas from 1985 to 2020. In the old city (Figure 15a), the ecological environment quality initially declined and then improved. Specifically, from 1985 to 1990, the quality deteriorated with a slope of −0.0286, while from 1991 to 2020, the quality improved with a slope of 0.0065. In the expansion areas (Figure 15b), ecological quality exhibited an overall trend of degradation, with a slope of −0.0058. In the suburban areas (Figure 15c), the ecological quality experienced a slight decline but remained relatively stable, with a slope of −0.0005.

3.5. Spatiotemporal Response of Urban Ecological Environment to Urbanization Intensity

This study investigates the spatiotemporal response of the ecological environment to urbanization intensity through correlation analysis. Figure 16 illustrates the correlation and statistical significance between ISDD and RSEI in Hangzhou. As shown in Figure 16a, a negative correlation between ISDD and RSEI was observed in 65.0% of the area, primarily concentrated in the expansion areas and sporadically in the suburban areas. Among these, 6.4% of the areas showed a significant negative correlation, while 16.5% exhibited a highly significant negative correlation. In contrast, 35.0% of the area showed a positive correlation, with 2.4% showing a significant positive correlation and 2.2% showing a highly significant positive correlation, mainly located in the old city and suburban areas. This indicates that the response of urban ecological quality to urbanization intensity exhibits spatial heterogeneity: in the old city, the positive effects of urbanization on the ecological environment outweigh the negative impacts, whereas in expansion areas, the negative effects are more dominant.

Figure 17 displays the detailed response of RSEI to ISDD in the old city, expansion areas, and suburban areas of Hangzhou. It is evident that the relationship between ISDD and RSEI is nonlinear.

In the early stage of urbanization (1985–1995): In the old city (Figure 17a), ISDD increased from 0.78 to 0.88, and RSEI declined as ISDD increased; in the expansion areas (Figure 17b), ISDD rose from 0.08 to 0.18, and RSEI sharply decreased from 0.72 to 0.44; in the suburban area (Figure 17c), ISDD increased from 0.02 to 0.04, and RSEI rapidly dropped from 0.76 to 0.58.

In the mid-to-late stage of urbanization (1996–2020): In the old city (Figure 17a), ISDD increased from 0.88 to 0.94, reaching saturation, and RSEI began to rise as ISDD continued to increase; in the expansion areas (Figure 17b), ISDD increased from 0.20 to 0.75, while RSEI fluctuated and declined within the range of 0.47 to 0.35; in the suburban areas (Figure 17c), ISDD increased from 0.04 to 0.16, and RSEI exhibited slight fluctuations between 0.53 and 0.67.

3.6. Results of CCD Between Urbanization and Eco-Environmental Quality

Figure 18 illustrates the coupling coordination dynamics between urbanization and the eco-environmental quality across different urban zones in Hangzhou from 1985 to 2020.

In the old city (Figure 18a), the coupling coordination index remained in a transitional development stage throughout the urbanization process. Initially, the coordination index decreased from 0.75 in 1985 to 0.64 in 1991, then increased to 0.81 in 2019. This trend represents a progression from moderate coordination (MC) to primary coordination (PC) and eventually to good coordination (GC). During the entire development period, the relative development degree remained greater than 1 (R > 1), indicating that the eco-environmental quality was in a lagging state relative to urbanization.

In the expansion areas (Figure 18b), the coupling coordination stage evolved from a conflict stage (1985–1997) to a transitional development stage (1998–2020). The coordination progressed from mild imbalance (MII) to moderate imbalance (MOI) and then to moderate coordination (MC). Regarding the relative development degree, the relationship between urbanization and the ecological environment changed over time. From 1985 to 2002, urbanization lagged behind the ecological environment (R < 1); between 2003 and 2006, the two systems tended toward balance (R ≈ 1); however, from 2007 to 2020, ecological development lagged behind urbanization (R > 1), indicating a shift in the dominant trend.

4. Discussion

4.1. Potential Drivers of Eco-Environmental Quality Change

This study reveals different trends in eco-environmental quality across areas with varying levels of urbanization intensity. The findings underscore the complexity of urbanization and its varied impacts, suggesting that the underlying factors driving these differences merit further investigation. Understanding these driving factors and their impacts on Hangzhou’s ecological environment is crucial for developing sustainable urban planning strategies and ecological restoration efforts.

The eco-environmental quality in the old city has improved during urbanization (Figure 14 and Figure 15a). Several factors contribute to this phenomenon: (1) Green Infrastructure Development: Urban planning advancements have led to the widespread implementation of green infrastructure in the old city, such as parks, green belts, and vertical greenery, which have significantly enhanced local ecological quality [80,81,82]. Increased vegetation coverage helps mitigate the urban heat island effect and improves air quality [83,84]. (2) Environmental Policies and Management: Over time, Hangzhou has strengthened its environmental management capabilities. Measures such as low-carbon city planning, sustainable development policies, and pollutant emission restrictions have effectively improved the ecological environment [85,86]. (3) Socio-Economic Influence: Increased public awareness and government-community collaboration have driven efforts to improve urban ecology [87].

In contrast, expansion areas face a decline in eco-environmental quality (Figure 15b). The main factors may be attributed to (1) Industrial Expansion: As part of Hangzhou’s urban growth strategy, large-scale industrial zones and parks have been developed outside the historical city center. The expansion of industrial zones has resulted in significant land use changes, including the conversion of agricultural land, forests, and wetlands into urban and industrial areas [88]. This transformation has had implications for local ecological quality. (2) Infrastructure Development: The rapid expansion of industrial zones has been supported by the development of extensive transportation infrastructure. Highways, railways, and the Hangzhou Metro have connected industrial areas to the city center, facilitating the movement of goods and people. The residential and commercial developments have also followed the growth of industrial areas [89]. These have resulted in urban sprawl, with new districts emerging in the suburbs to accommodate the growing population and workforce. The rapid expansion of industrial activities and the development of infrastructure have significantly impacted the local ecology.

4.2. CCD Model Analysis

The results of the CCD model indicate a gradual improvement in the coordination between urbanization and ecological quality (Figure 18). However, the reasons behind this shift require deeper analysis. The improved CCD in the old city can largely be attributed to effective urban management policies, including the implementation of green infrastructure and sustainable development strategies. These policies allowed the old city to better manage its ecological environment despite urban expansion. Conversely, the expansion areas exhibited ecological lag at certain stages, as urbanization pressures (e.g., industrial development, rapid infrastructure growth) initially outpaced the development of environmental management measures. This lag can be attributed to the delayed integration of ecological considerations into urban planning and the absence of comprehensive ecological restoration efforts in these areas.

4.3. Limitations and Future Prospects

(1)

While ISDD is a powerful tool for quantifying urbanization intensity, its effectiveness can be limited by factors such as spatial resolution, boundary effects, and environmental context. To enhance its accuracy and applicability, ISDD can be improved by integrating higher-resolution data, employing adaptive analytical approaches, and linking it with ecological and socioeconomic indicators. Addressing these challenges will strengthen ISDD’s robustness, making it a more effective tool for urban environmental research and the planning of sustainable urban development.

(2)

While urbanization is often measured by the expansion of impervious surfaces, it also has significant effects on aquatic environments. The current use of RSEI fails to account for water bodies, which are vital components of urban ecosystems and play critical roles in regulating temperature, supporting biodiversity, and mitigating flooding. In urban areas, where water features are integral to both the urban landscape and environmental quality, the neglect of these areas can result in an incomplete assessment of the overall urbanization impact. Future studies could integrate water quality and quantity indicators into the urbanization analysis. This could be achieved by using additional remote sensing indices specifically designed for monitoring water bodies, such as the normalized difference water index (NDWI), or using hyperspectral imagery to assess water health, including pollution levels, vegetation along shorelines, and changes in water surface area.

4.4. Enlightenment

The findings of this study suggest several policy recommendations for Hangzhou’s urban and ecological development. For the expansion areas, where ecological degradation is most pronounced, policies focusing on greenway construction, pollution control, and eco-friendly urban planning are essential. The development of green infrastructure, such as urban parks, tree planting, and permeable surfaces, can help mitigate the negative effects of rapid urbanization [90]. In contrast, the old city, which has shown signs of ecological improvement, would benefit from continued efforts in low-carbon urban planning, including the promotion of energy-efficient buildings, renewable energy adoption, and sustainable transportation systems. Further, the city should strengthen environmental management policies to maintain and enhance the ecological quality of the old urban area [91]. These measures will contribute to the long-term balance between economic development and ecological protection, ensuring sustainable urban growth.

5. Conclusions

In this study, we used impervious surface data and Landsat time-series data from 1985 to 2020 to construct two key indicators: ISDD and RSEI, representing urbanization intensity and eco-environmental quality, respectively. We revealed the spatiotemporal evolution trends of ISDD and RSEI in Hangzhou and explored their dynamic response and coupling coordination relationships across different urban zones. The key findings of this study are summarized as follows:

(1)

Interannual Trends in ISDD and RSEI: Urbanization has generally contributed to a deterioration in eco-environmental quality in Hangzhou. However, the trends vary spatially: ecological quality in expansion areas has significantly declined, while in the old city, it has markedly improved.

(2)

Dynamic Response of Ecological Quality to Urbanization Intensity: The ecological response to urbanization intensity differs by zone. In expansion areas with low urbanization intensity, ecological quality deteriorates as urbanization progresses. Conversely, in the old city, where urbanization intensity is higher, ecological quality improves as urbanization grows.

(3)

Coupling Coordination Between ISDD and RSEI: The relationship between urbanization and ecological quality has become more coordinated over time, indicating a shift towards more balanced development that integrates urban growth with environmental sustainability.

The findings of this study have significant implications for urban planning in Hangzhou, emphasizing the importance of integrating ecological considerations into urban development strategies. Such integration is essential for mitigating the adverse effects of urbanization on ecological quality and fostering a more sustainable path of urban growth. Moreover, Hangzhou’s experience provides valuable insights for other rapidly urbanizing regions. Cities facing similar challenges can adopt similar strategies to better achieve a sustainable urban future while maintaining ecological balance. This underscores the importance of targeted policies that address the unique needs and challenges of each region, ensuring that urban development does not come at the expense of environmental health.