Long-Term Clustering Analysis of Construction Land Reclamation in Hangzhou: Patterns and Impacts

Abstract

Given the challenges of rapid urbanization and ecological degradation, construction land reclamation (CLR) has become one of the critical strategies for optimizing land use and balancing urban development with ecological preservation. However, the long-term impacts and patterns of CLR remain underexplored, necessitating a comprehensive analysis to inform sustainable land management practices. This study investigates the patterns and impacts of CLR in Hangzhou over a fifteen-year period (2006–2020). Using an improved K-means clustering algorithm, the study analyzes the land reclamation status of 199 streets, identifying nine distinct categories that reflect varying stages of urbanization, ecological protection, and land-use management. The results show a clear transition from rapid, large-scale reclamation during the early years to more targeted and sustainable reclamation efforts in later periods. The study’s innovation lies in its use of a long-term, time-series analysis at the street level, enhancing understanding of CLR dynamics and improving prediction accuracy by 11%. It emphasizes the importance of adaptive land management in balancing urban growth with ecological preservation and agricultural sustainability. The clustering analysis provides insights into the evolving urbanization process, with the highest concentration of construction land reclamation occurring along urban–rural boundaries. These findings offer valuable lessons for urban planning and sustainable urbanization strategies.

1. Introduction

Since the 1950s, the rapid acceleration of global urbanization has intensified human disturbances to the natural environment, exacerbating resource scarcity, environmental pollution, and ecosystem degradation. These challenges have become major obstacles to sustainable development, with both local and global consequences [1,2]. As the world’s largest developing country, China has experienced rapid industrialization and urbanization since the reform and opening-up, leading to significant ecological and environmental challenges that pose risks to the sustainability of its socio-economic development [3,4]. In response to these challenges, China has implemented national development strategies such as urban–rural development, new-type urbanization, beautiful countryside construction, and rural revitalization [5,6,7,8,9]. These strategies focus on the coordinated development of various factors, such as population, land, and industries, aiming to promote balanced development between urban and rural areas [10,11]. In the future, urban development will move away from blind expansion and chaotic construction, entering a new phase centered on adjusting the structure of existing land resources [12,13].

Land consolidation, as a vital component of comprehensive land governance, has become one of the most significant approaches to changing land use patterns and improving the man–land relationship in China [14,15,16]. Early land consolidation efforts focused on quantity over quality, largely neglecting ecological concerns [17,18]. However, with the gradual promotion of ecological civilization construction, land consolidation is undergoing an ecological transformation. During China’s 12th and 13th Five-Year Plans, the government released “National Land Consolidation Plan (2011~2015)” and “National Land Consolidation Plan (2016~2020)” to coordinate the promotion of land consolidation. And inspired by the pilot project in Zhejiang Province, the Ministry of Natural Resources of China proposed the Whole-region Comprehensive Land Consolidation Initiative in 2019 [2]. The focus of land consolidation has shifted to adjusting regional land use patterns in order to create a more rational and efficient land use structure, thereby enhancing the overall stability and functionality of ecosystems [4,19]. This shift has become a pressing necessity in contemporary development [14,20]. As such, evaluating the effectiveness of land consolidation is now a critical element in advancing these land consolidation practices [21,22,23]. Research on the effectiveness of land consolidation typically examines three areas: social, economic, and ecological benefits [24,25,26,27]. Such studies often created index systems to assess these effects [28], with a predominant focus on measuring the economic benefits of land consolidation—such as land valuation, food security, and land fragmentation [29,30,31,32]. However, scholars differed in their views on the ecological benefits of land consolidation. Some argued that land consolidation positively impacts the ecological environment [33,34,35], while others suggested that it could have negative consequences [5]. Some studies analyzed the effects of land consolidation by investigating land use changes [36,37]. Most research, however, tended to focus on individual land consolidation projects, with little long-term, time-series analysis [27,38,39]. Overall, current evaluation methods for land consolidation effectiveness remain relatively singular, with inconsistent standards for selecting indicators and unclear definitions of benefits. Moreover, there is a lack of long-term studies to assess the sustained benefits, which hinders the ability to inform policy formulation effectively [40]. Therefore, selecting appropriate perspectives and indicators for scientifically analyzing and evaluating the implementation of land consolidation practices, identifying patterns, and discovering issues is crucial for promoting the long-term sustainability of these efforts [41].

Land reclamation is one of the key tools for implementing land consolidation and plays an important role in reducing inefficient construction land and improving land use efficiency [42]. It focuses on the protection, governance, and utilization of land resources at the site level. In the critical phase of adjusting the structure of existing land resources, effectively utilizing land reclamation is central to unlocking the potential of land resources and their value. This is especially true for regions at the interface of urban and rural areas in developing countries, where land use attributes frequently change [14]. Land reclamation mainly includes agricultural land reclamation, construction land reclamation (CLR), and other land reclamation [43]. Among them, CLR refers to the process of implementing corresponding engineering measures to reclaim inefficient, idle, or abandoned construction land that meets the conditions for reclamation, converting it into arable land or other types of agricultural land. During the period of 2006–2020, urban development is shifting from “incremental expansion” to “exploiting existing stock”, with existing built-up land becoming a key resource for urban spatial adjustment. CLR, focusing mainly on inefficiently used land, with a significant emphasis on “reducing” the scale of construction land, has become an important implementation strategy for adjusting the structure of urban construction land. Analyzing the long-term impacts of land reclamation on land use and the ecological environment, summarizing patterns, and assessing the effectiveness of urban and rural CLR could provide a new angle for evaluating the effectiveness of land consolidation. Currently, research on land reclamation is largely focused on individual land projects, such as a certain mine or temporary land use, with less attention given to urban and rural CLR [27,28,42,44]. Some studies used spatial analysis methods and index models to estimate the potential of land reclamation under different scenarios [17]. Based on these analyses, scholars categorized land consolidation into types such as urban expansion, agricultural development, ecological enhancement, and farmer well-being. Methods like the Analytic Hierarchy Process, Entropy Weight Method, and Multi-Factor Comprehensive Evaluation Method are then applied to explore the effects of land reclamation on agriculture, society, and the economy [45]. Other studies focused more on the mechanisms of land reclamation, conducting comprehensive analyses of land use changes and the roles of key stakeholders, such as farmers, government agencies, and real estate developers, to examine the effects of land use changes on rural revitalization [14,46]. However, due to the sensitivity of land data, academic research on existing land tends to focus on qualitative analysis and policy mechanisms. There remains a gap in quantitative analysis based on time-series data to uncover regional policy mechanisms and evaluate their implementation. From the perspective of research scale, current studies focus on the micro-level analysis of individual reclamation sites or on macro-level evaluations at the provincial or municipal level [47]. There is a lack of meso-scale research focused on streets as the unit of analysis.

In land consolidation practices, Zhejiang Province has consistently been at the forefront nationwide, with relatively complete achievements and data. In 2003, it pioneered the “One Thousand Demonstration Villages and Rectification of Ten Thousand Villages”, integrating land consolidation, reclamation, and development, marking the initial exploration of land improvement efforts. After 2010, the province further advanced land consolidation by focusing on the urban–rural construction land balance policy. Hangzhou, closely following provincial land consolidation policies, launched the first pilot project for whole-region comprehensive land consolidation in Shuangpu Town, Xihu District, in 2017 [48]. This initiative marked a shift from fragmented land development and consolidation to large-scale, integrated land management, accumulating extensive practical experience. As the capital of Zhejiang Province, Hangzhou faces increasing land resource constraints, particularly under the dual pressures of urban expansion and ecological protection. Efficiently reclaiming construction land and improving land use efficiency have become critical challenges. Therefore, studying the effectiveness of CLR in Hangzhou not only helps summarize past experiences and provide scientific support for future land consolidation efforts but also serves as a valuable reference for optimizing land resource utilization in other regions. This study focuses on land reclamation in Hangzhou from 2006 to 2020, using long-term time series data. Due to different land consolidation policies during China’s “11th Five-Year Plan”, “12th Five-Year Plan”, and “13th Five-Year Plan” periods, research also explores the spatiotemporal dynamics of land reclamation according to the three planning periods. Using an improved K-means clustering algorithm based on machine learning, the study clusters over 100 streets in Hangzhou and compares the similarities and differences in land reclamation policies across streets with different geographic characteristics. The study also investigates the factors influencing the retention rate of cultivated land and assesses the impact of CLR on urban land use structure and ecosystems, and provides a robust data foundation for advancing strategies related to farmland protection, urban–rural integration, and rural revitalization.

2. Materials and Methods

2.1. Study Area

Hangzhou, the capital city of Zhejiang Province in China, is a mega city located in the Yangtze River Delta region (Figure 1). It is also the economic, cultural, and scientific center of Zhejiang Province. As of the end of 2023, the city’s permanent population reached 12.52 million, with an urbanization rate of 84.2%. Its topography is complex and varied, with an overall decline in elevation from west to east. The western, central, and southern parts of the city are characterized by low mountains and hills, while the northeastern part belongs to the northern Zhejiang plain. The city boasts a natural environment where rivers, lakes, and mountains converge, making it a typical and representative city within Zhejiang. Hangzhou’s urban development plays a key role in promoting coordinated growth in the Yangtze River Delta and serves as a guiding force for the development of other cities in the region.

2.2. Data Source

The land use data are sourced from the 30-m resolution dataset provided by the Chinese Academy of Sciences, including 6 primary types and 25 s-class types. It is primarily used to analyze land use changes in plots after construction land reclamation. The digital elevation map (DEM) data of Hangzhou with a grid size of 30 × 30 m is obtained from the Geospatial Data cloud (http://www.gscloud.cn/, 10 Sebtember 2024) of the Chinese Academy of Sciences. The construction land reclamation data are sourced from the Hangzhou Bureau of Natural Resources and Planning. This dataset includes information on the reclamation history of 7726 plots, which have undergone significant land use changes driven by construction land reclamation policies implemented over the period from 2006 to 2020. These plots are distributed across nine administrative districts and counties within the Hangzhou metropolitan area. This rich dataset enables a nuanced analysis of regional variations in land reclamation activities and offers insights into the complex interactions between land use change, policy implementation, and urban development.

2.3. Methods

2.3.1. Improved K-Means Clustering Algorithm Based on Machine Learning

Cluster analysis is an unsupervised classification learning method that divides data into categories with similar features based on their attributes or information in order to analyze the underlying mechanisms and significance. K-means clustering is a popular and widely used clustering algorithm that classifies data based on the distance between samples. It is computationally efficient and suitable for handling large-scale datasets [49]. Due to its straightforward principles, K-means is easy to understand and interpret, making it widely adopted in practical applications. However, traditional K-means clustering has certain limitations, such as sensitivity to outliers, dependence on initial centroids, and assumptions about the shape and size of clusters. This study improves the traditional K-means algorithm using a machine learning-based enhanced K-means clustering method [50,51]. First, an adaptive data preprocessing approach is applied to remove outliers from the source data, improving accuracy, reliability, and the predictive performance of the model. Second, the initial centroids are adjusted iteratively to eliminate issues related to anomalous initial points. Lastly, the number of pre-clusters is iterated to achieve higher predictive performance through an unsupervised approach. In this study, PyCharm was used with Python 3.12 to improve the classical K-means clustering method.

2.3.2. Cluster Analysis and Evaluation

For evaluating clustering results, there are generally two criteria in partitional clustering. One is that the intra-cluster cohesion should be as tight as possible, and the inter-cluster separation should be as large as possible. The second is that the clustering results should align closely with human judgment. This study uses the Silhouette Coefficient as one standard to evaluate clustering results [52]. By comparing the relative quality of clustering results across different distance measures, after dividing the dataset into k categories, for one of the partition units i, a (i) is the average distance from i to all other points in the cluster, and b (i) is the minimum average distance between i and all points in the nearest neighboring cluster. Therefore, the Silhouette Coefficient is:

$$S\left(i\right)={\displaystyle \frac{b\left(i\right)-a\left(i\right)}{max\left\{a\left(i\right),b\left(i\right)\right\}}},$$

(1)

2.3.3. Environmental Impact Assessment

This study uses the Habitat Quality Assessment model from InVEST to evaluate the ecological environmental quality before and after construction land reclamation and analyze the changes. Based on the principles of the model, this study defines forests, water bodies, grasslands, and cultivated land as habitats, while urban construction land, rural construction land, other construction land, and bare land are considered threats to these habitats. Different maximum distances, weights, and sensitivities are assigned to each object based on values recommended by the model and relevant studies, as shown in

Table 1. Maximum distance and weight of the threats affecting habitat quality.

Threat	Maximum Distance (km)	Weight	Decay
Urban construction land	9	1	Exponential
Rural construction land	5	0.6	Exponential
Other construction land	4	0.6	Exponential
Bare land	4	0.3	Exponential

and

Table 2. The sensitivity of habitat types to each threat factor.

Land-Use Type	Habitat Suitability Score	Sensitivity to Threats
		Urban Construction Land	Rural Construction Land	Other Construction Land	Bare Land
Paddy	0.5	0.7	0.6	0.5	0.5
dryland	0.4	0.5	0.35	0.2	0.1
Forest	1	1	0.85	0.6	0.3
Shrub	1	0.6	0.45	0.2	0.2
Sparse woods	1	1	0.9	0.65	0.3
Other forest	1	1	0.95	0.7	0.3
Dense grass	0.8	0.6	0.5	0.2	0.2
Moderate grass	0.75	0.6	0.45	0.25	0.25
Sparse grass	0.7	0.7	0.6	0.3	0.3
River	0.9	0.9	0.8	0.7	0.3
Lake	0.9	0.9	0.75	0.5	0.3
Reservoir/ponds	0.7	0.9	0.75	0.5	0.3
Tidal flat	0.6	0.95	0.75	0.5	0.3
Bottomland	0.6	0.95	0.85	0.55	0.5
Urban construction land	0	0	0	0	0
Rural construction land	0	0	0	0	0
Other construction land	0	0	0	0	0
Bare land	0	0	0	0	0

.

The specific calculation formula for habitat quality is as follows:

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

(2)

$$D_{xj}={\sum }_{r=1}^R{\sum }_{y=1}^{Y_r}\left({\displaystyle \frac{w_r}{{\sum }_{r=1}^R{w}_r}}\right)r_y{i}_{rxy}{\beta }_x{S}_{jr},$$

(3)

$$i_{rxy}=\left\{\begin{array}{c}1-\left({\displaystyle \frac{d_{xy}}{d_{rmax}}}\right), \mathrm{L}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{a}\mathrm{r} \mathrm{d}\mathrm{e}\mathrm{c}\mathrm{a}\mathrm{y}\\ exp\left(-\left({\displaystyle \frac{2.99}{d_{rmax}}}\right)·d_{xy}\right), \mathrm{E}\mathrm{x}\mathrm{p}\mathrm{o}\mathrm{n}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l} \mathrm{d}\mathrm{e}\mathrm{c}\mathrm{a}\mathrm{y}\end{array}\right.$$

(4)

Among them, Q_x is habitat quality, H_j is habitat suitability, D_xj is habitat degradation degree, r is the threat source of the habitat, y is the grid in the threat source r, d_xy is the distance between grid x (habitat) and grid y (threat source), d_rmax is the impact range of threat source r, and i_rxy is the impact of various threat sources on habitat x. According to the attenuation methods (linear attenuation or exponential attenuation) generated by different threat sources, their calculation methods are divided into linear attenuation and exponential attenuation.

3. Results

3.1. Periods in CLR in Hangzhou

CLR in Hangzhou has unfolded across three distinct Five-Year Plan periods from 2006 to 2020: Period I (2006–2010), Period II (2011–2015), and Period III (2016–2020). These periods reflect a broader trend in urban land development that began with extensive land use adjustments, followed by optimization and final consolidation efforts. Analyzing the number and scale of reclamation projects during these periods provides insight into the city’s evolving approach to land management in the context of rapid urbanization (Figure 2).

During Period I, Hangzhou experienced the highest number of reclamation projects, marking an initial phase of large-scale land use adjustments. A total of 3470 projects were implemented, including 807 large-scale reclamation initiatives finalized in the final two years of the period. This phase was characterized by intensive and widespread CLR, as the city sought to rationalize its land use structure, improve the efficiency of land utilization, and promote a more balanced development between urban and rural areas. The substantial reclamation activities during this period were critical in laying the foundation for Hangzhou’s urban expansion, particularly in peripheral districts where land was converted for construction purposes to support the city’s burgeoning population and economic growth.

In Period II, the number of reclamation projects decreased slightly to 3362, although this figure was still relatively high, indicating continued urban expansion and land reorganization. However, a significant shift in the nature of reclamation occurred. The majority of projects in this period were small- to medium-sized, and most were completed within the first two to three years. The period saw the urban core in the northeastern part of Hangzhou become nearly fully urbanized, with further reclamation focused on optimizing and regulating the land use in areas already affected by urbanization. The projects during this period were particularly concentrated in Lin’an District and surrounding areas, where the focus was on fine-tuning land use to enhance urban–rural integration and support the city’s growth while maintaining environmental sustainability. Notably, only five large-scale reclamation projects extended beyond four years, signaling a trend towards more controlled and deliberate land use adjustments.

By Period III, CLR activities began to significantly diminish. A total of 576 projects were completed, reflecting the final stage of reclamation consolidation. During this phase, the focus shifted to more targeted adjustments in specific regions such as Chun’an County, where the need for land reallocation remained, albeit at a smaller scale compared to the earlier periods. At the same time, significant areas of construction land, forests, orchards, and farmland in the urban center reached a state of equilibrium, effectively reducing the necessity for large-scale land-use conversions. The harmonious coexistence of various land uses in these areas demonstrated the success of integrated urban planning policies aimed at minimizing conflicts between urban expansion and ecological protection.

The spatial distribution of CLR areas during these three planning periods is illustrated in Figure 3. During Period I, the largest reclamation areas were observed, particularly in peripheral urban districts such as Lin’an District, Jiande City, Fuyang District, and Tonglu County in the central part of Hangzhou. These areas were targeted for large-scale land use transformations, with suburban districts like Yuhang District, Xiaoshan District, and Linping District undergoing smaller-scale reclamation projects. In rural areas, land was frequently converted into agricultural zones suitable for large-scale farming, while land closer to urban centers was repurposed for construction to accommodate the expanding urban footprint. In contrast, central urban districts such as Xihu District, Gongshu District, and Shangcheng District, which were already largely urbanized, saw only limited reclamation projects during this period.

In Period II, reclamation activities continued at a significant scale but with a more focused approach. Streets with notable reclamation were predominantly located in the central and southern parts of Hangzhou, including areas like Tonglu County, Fuyang District, and Jiande City, with some reclamation occurring in the northeastern Qiantang District. These regions, where urban and rural boundaries were more fluid, saw targeted land-use changes aimed at optimizing urban growth and facilitating land development without disrupting the surrounding ecological systems.

By Period III, the area under CLR sharply declined, reflecting the near completion of Hangzhou’s urbanization process. Reclamation was mostly limited to regions like Chun’an County and Tonglu County, where small-scale adjustments were necessary to fine-tune land use patterns. In these areas, the potential for further large-scale reclamation was limited due to both the saturation of urbanization and the increasing emphasis on sustainable land management practices.

Based on the implementation of reclamation activities across the three planning periods, Hangzhou’s streets can be categorized into four types (Figure 4): streets with no CLR, streets with reclamation in one period, streets with reclamation in two periods, and streets with reclamation in all three periods. Notably, 78 streets experienced no CLR, primarily concentrated in central urban areas and key ecological protection zones, such as the regions around Qiandao Lake, which are preserved for their environmental value. Forty-two streets underwent reclamation in only one period, with small reclamation areas concentrated in urban–rural boundaries like Yuhang District, Lin’an District, and Fuyang District. These areas were constrained by the need to balance urban growth with agricultural and ecological preservation, limiting further reclamation. A larger group of 67 streets underwent reclamation in two periods, primarily in Lin’an District, Fuyang District, and Jiande City, where moderate reclamation activities were carried out in the earlier periods. These streets reflect a more stable phase of urban development, where initial reclamation efforts were followed by careful adjustments to land use, ensuring continued growth while managing environmental impacts. Finally, 12 streets with reclamation in all three periods typically exhibited significant agricultural production potential and were located in suburban districts like Tonglu County and Jiande City. These areas benefited from sustainable reclamation practices, allowing for the integration of production, living, and ecological functions, thus supporting balanced urban expansion in alignment with long-term development goals.

The overall trends in CLR in Hangzhou illustrate the city’s shift from rapid urbanization to more controlled and sustainable land management. As urbanization progresses, the need for large-scale reclamation decreases, giving way to more refined land-use planning that seeks to balance development with environmental protection and agricultural sustainability. These trends highlight the importance of adaptive land management strategies in ensuring that urban growth does not come at the expense of ecological health or the productive capacity of land.

3.2. Impacts of CLR

3.2.1. Impacts on Cultivated Land

Following land reclamation practices, taking the data from 2005, the end of the previous planning period, as the baseline, this analysis examines the changes in cultivated land area at the end of three subsequent planning periods (2010, 2015, and 2020) from 2006 to 2020. The changes are classified into four categories: consistent increase, consistent decrease, keep steady, and fluctuation, as depicted in Figure 5. Consistent increase refers to areas where the cultivated land area steadily increased at the end of all three periods. Consistent decrease refers to areas where the cultivated land area steadily decreased at the end of all three periods. Keep steady refers to areas where the cultivated land area did not change in quantity by the end of any of the three periods. Fluctuation refers to areas where the cultivated land area increased and then decreased, or decreased and then increased, over the three periods. These shifts reflect the interplay between urbanization and agricultural land use, with varying trends across different areas of the city. In total, six streets experienced a consistent increase in cultivated land, primarily located in the more peripheral regions such as Yuhang District and Lin’an District. These areas, situated on the outskirts of the urbanized core, have seen a marginal expansion of agricultural land, possibly due to the availability of relatively underdeveloped spaces where agricultural activities can be maintained or reintroduced. The increase in cultivated land in these areas, however, is relatively small compared to other trends, highlighting the limited scope for agricultural expansion in the face of broader urbanization pressures.

The largest category, encompassing 136 streets, shows a consistent decrease in cultivated land. This pattern is most pronounced in areas undergoing significant urban expansion, where agricultural land has been progressively sacrificed to accommodate the growing demand for residential, industrial, and commercial development. The reduction in cultivated land is a clear reflection of the urbanization process, where the need for infrastructure, housing, and economic activities has increasingly taken precedence over agricultural use. This trend is particularly evident in the urban core and its immediate periphery, where the conversion of farmland into built-up areas continues to be a dominant force driving land use change.

In contrast, nine streets experienced fluctuations in the total area of cultivated land, with these changes scattered across regions such as Qiantang District, Jiande City, and Chun’an County. These areas are generally situated farther from the urban center and are located along the urban–rural interface, where land use transitions between urban and agricultural areas are more dynamic. The fluctuations may reflect a combination of factors, including shifting agricultural practices, land reclamation efforts, as well as land use policies that occasionally promote agricultural preservation or expansion. Such areas, which are in close proximity to both urban developments and rural spaces, may experience periodic reversals in land use trends due to changing socio-economic pressures and land management strategies.

Additionally, 48 streets demonstrated a stable total area of cultivated land, with these areas largely concentrated in the central urban zones and key ecological protection regions around Qiandao Lake. In the central urban areas, this stability can be attributed to deliberate efforts to maintain agricultural land within ecological reserves or zones designated for sustainable development. These regions, often located in or near ecologically sensitive areas, play a crucial role in balancing urban growth with environmental conservation. The stability in cultivated land in these areas suggests that there are concerted efforts to protect farmland and integrate it into urban planning frameworks that prioritize ecological preservation alongside development. The persistence of cultivated land in the Qiandao Lake area is likely influenced by its importance as a scenic and ecological zone, where land use policies may favor agricultural or conservation-based land uses over urban encroachment.

Taken together, these trends illustrate the complex relationship between urban expansion and agricultural land use in Hangzhou. While peripheral areas may see some expansion in cultivated land, the broader trend is one of decreasing agricultural land due to urban development. However, fluctuations and stability in certain regions highlight the potential for dynamic land use management strategies that can adapt to both urban and environmental needs. This underscores the importance of thoughtful urban planning and land reclamation practices that balance the demands of growth with the need to preserve agricultural and ecological resources for future generations.

3.2.2. Impacts on Ecological Environment

In addition to changes in land use, there have also been notable alterations in the ecological environment. These changes in habitat quality are essential for understanding the broader ecological dynamics of the region. Taking the 2005 data as the baseline, this analysis examines the changes in habitat quality at the end of the three planning periods (2010, 2015, and 2020), as illustrated in Figure 6. Habitat quality in Hangzhou demonstrates a clear spatial variation, with a prominent west-high, east-low pattern. Specifically, the southwestern region, encompassing the Qiandao Lake Scenic Area and its surrounding environments, is characterized by more robust ecological functions. This area serves as a critical ecological buffer with a relatively high degree of natural vegetation, water resources, and biodiversity. In contrast, the northeastern urban areas, which serve as a hub for human activities and urban development, exhibit notably lower levels of habitat quality, reflecting the impacts of urban expansion, industrialization, and infrastructural development on the natural environment.

From a temporal perspective, the overall trend in habitat quality across Hangzhou has remained relatively stable over the study period, with no drastic improvements or declines observed at the regional scale. However, certain localized areas, particularly those on the periphery of the northeastern urban zone, have experienced a noticeable degradation in habitat quality. This decline can be attributed to the intensification of urban sprawl, increased pollution levels, and the encroachment of built environments into previously ecologically sensitive areas. As urbanization continues to unfold, the regions with lower habitat quality exhibit a fluctuating pattern of expansion and contraction. Notably, these areas initially expanded as a result of urban encroachment but are now showing signs of contraction as efforts to mitigate environmental degradation—such as green infrastructure development, ecological restoration, and urban planning—begin to take effect. The central urban areas, however, remain relatively stable, with a continual balance between urban development and ecological preservation, albeit with localized disruptions due to infrastructural projects and population growth. This evolving pattern underscores the complex interaction between urban growth and ecological sustainability, highlighting the need for more integrated approaches to urban planning and environmental management.

3.3. Clustering of CLR

CLR in Hangzhou from 2006 to 2020 was analyzed using a clustering method. The analysis considers 12 variables, with Period I, Period II, and Period III each encompassing four aspects: reclaimed construction land area, cultivated land retention rate, paddy field preservation rate, and habitat quality change. In the clustering experiment, the Silhouette Coefficient was observed to initially increase and then decrease as the number of clusters grew. The highest value of 0.82 was achieved when the number of clusters was set to nine. Based on this result, the study adopts nine clusters for further analysis, as shown in Figure 7. It reflects variations in urbanization stages, ecological protection, and land-use management. These categories are discussed in detail below:

• Category A (Urban Core Areas with No Land Reclamation): 78 streets, primarily located in the city center, represent areas with more mature urbanization. No reclamation occurred here between 2006 and 2020, as these regions were already highly urbanized and densely developed. These areas exhibit poor agricultural functions, with fragmented cultivated land and an ecosystem dominated by artificial landscapes. Habitat quality in these areas remained below 0.4 during the entire 2006–2020 period, which is lower than the average habitat quality found in peripheral urban areas, reflecting the limited ecological value of these highly urbanized zones.
• Category B (Central Urban Periphery with Early-Stage Reclamation): This category includes 24 streets near the main urban areas, such as Yuhang and Lin’an Districts. These streets underwent single-stage reclamation during Period I, driven by the increasing demand for construction land as urbanization expanded outward. Agricultural land near the urban center was gradually converted to construction land, resulting in a continuous decline in the total amount of cultivated land. The retention rate of cultivated land after reclamation was notably low, below 30%, reflecting the early stages of urban expansion where agricultural land was not adequately protected. These areas exhibited relatively high habitat quality (above 0.8) during Period I, suggesting that urbanization, while impactful, did not result in total degradation of ecological functions in these transitional zones.
• Category C (Peripheral Areas with Mid-Stage Land Reclamation): This category includes 15 streets scattered across the peripheral urban areas of Chun’an County, Jiande City, and Fuyang District, where reclamation occurred primarily during Period II. As urbanization expanded beyond the city center, reclamation activities began to affect more distant areas. In these regions, the total amount of cultivated land continued to decrease, with the retention rate of cultivated land during the Period I period remaining below 40%. Despite these trends, habitat quality remained relatively high (above 0.75) throughout Period I. This suggests that, while reclamation was occurring, ecological protection measures were effectively preserving habitat quality. These streets represent the early stages of peripheral urban expansion, where agricultural land was gradually converted to construction land, but ecological functions remained relatively unaffected.
• Category D (Ecologically Sensitive Fringe Areas): This category consists of three streets, primarily located around Qiandao Lake in Chun’an and parts of Yuhang. These areas underwent single-stage reclamation during Period III, representing a smaller subset of reclaimed areas. The retention rate of cultivated land in Yuhang reached 93% during this period, reflecting high retention in urban fringe areas. However, the agricultural land remained unstable, with a continuous decline in total cultivated land. Habitat quality in these regions remained close to 0.85, indicating relatively good ecological conditions despite the land use changes. In contrast, the Qiandao Lake area, which prioritizes ecological functions, exhibited much lower retention rates of cultivated land (below 30%), reflecting its emphasis on maintaining ecological functions.
• Category E (Transitional Urban–Rural Areas with Two-Stage Land Reclamation): A total of 51 streets, primarily located in the peripheral urban areas of Fuyang District, Jiande City, and around Qiandao Lake, underwent two-stage land reclamation during Period I and II. Although cultivated land continued to decrease, the retention rate of cultivated land improved in Period II, rising to above 60%. However, much of the newly reclaimed land was subject to further adjustments, making long-term retention difficult. Throughout 2006 to 2020, habitat quality remained above 0.8, reflecting relatively stable ecological conditions despite the ongoing transition from rural to urban uses. This suggests that while CLR played a central role in the area’s development, efforts were made to balance growth with ecological preservation, although concerns about the long-term sustainability of agricultural land use remain.
• Category F (Ecologically Dominant Areas with Two-Stage Land Reclamation): This category includes five streets, primarily in Lin’an and Chun’an, where two-stage land reclamation occurred during Period I and III. These areas are ecologically significant, and maintaining agricultural land was particularly challenging due to their high ecological value. As a result, retention rates for cultivated land remained below 30% in both periods, with much of the reclaimed land being converted to areas that support ecological functions. Despite this, the regions consistently exhibited the highest habitat quality in Hangzhou, maintaining values above 0.9 from 2006 to 2020. This may be attributed to the region’s focus on ecological preservation during the reclamation process, which helped maintain the dominance of ecological functions and contributed to the exceptional habitat quality.
• Category G (Land-Use Balanced Areas): This category includes five streets in Xiaoshan and Fuyang District, where CLR occurred both in Periods II and III. Due to the gradual implementation of the land-use balance policy, these areas maintained a high retention rate for cultivated land of 95%. However, while agricultural land was preserved, the conversion of forested and other ecological lands for urban development contributed to a decline in habitat quality, which remained consistently below 0.6 throughout the study period.
• Category H (Ecologically Integrated Areas with Three-Stage Land Reclamation): This category consists of 11 streets, primarily located in Tonglu County, Jiande City, and Chun’an County. These areas are similar to Category 5 regions, characterized by a dominant focus on ecological functions. They underwent three stages of land reclamation from 2006 to 2020, with favorable reclamation conditions. However, the protection of cultivated land still showed room for improvement. In Period I, the retention rate for cultivated land was relatively low at 34.8%, but it gradually improved, reaching 54.2% in Period Ⅱ and 71.3% in Period Ⅲ. Despite challenges in agricultural land protection, habitat quality remained consistently above 0.85, ranking second highest in Hangzhou. This category highlights regions where urban development was effectively balanced with ecological conservation and agricultural land preservation.
• Category I (Areas with Increasing Cultivated Land): This category includes seven streets, primarily located in Chun’an County and Xiaoshan District. These areas are unique in Hangzhou, being the only category where the total amount of cultivated land has been increasing. The retention of cultivated land has been particularly strong, with retention rates remaining between 80% and 90%, the highest in Hangzhou during Periods I and II. Streets in Xiaoshan District, part of the urban area, exhibit higher habitat quality compared to other urban regions, with values ranging from 0.4 to 0.5. In contrast, Chun’an County features a well-balanced integration of production and ecological functions, with habitat quality approaching that of Category F.

The clustering analysis highlights diverse trajectories in Hangzhou’s land reclamation processes, reflecting varying stages of urbanization, ecological sensitivity, and land-use management. These categories provide insight into the city’s evolving approach to land reclamation, where earlier phases of rapid urban expansion gave way to more targeted and sustainable land-use practices. The protection of cultivated land and ecological functions became increasingly important as the city transitioned from large-scale reclamation to more refined, context-specific interventions. The patterns of reclamation, retention rates, and habitat quality across these categories underscore the need for adaptive land management strategies that balance urban growth with the preservation of ecological and agricultural functions. These findings are valuable for urban planners and policymakers aiming to navigate the challenges of sustainable development amidst rapid urbanization.

4. Discussion

4.1. Phases of CLR and Urban Development

Hangzhou’s CLR efforts, unfolding in three distinct phases—Concentrated Remediation, Sustained Remediation, and Adjustment and Optimization—reflect the city’s evolving approach to urbanization. Period I focused on large-scale reclamation, primarily in suburban and rural areas, to accommodate rapid urban growth. This period emphasized the expansion of residential and commercial zones, resulting in significant alterations to the landscape and land use structure. By contrast, Period II marked a shift toward more targeted reclamation, addressing urban–rural borders and focusing on optimizing land use around the city’s periphery. In Period III, reclamation efforts were smaller in scale but more refined, aimed at adjusting and fine-tuning urban areas with a focus on long-term sustainability.

The trends of CLR in different stages are highly similar to the trends of urban expansion. During rapid urban growth, the density within the city center continues to increase. However, as available space in the central areas becomes limited, high-density zones are increasingly appearing on the outskirts of the city, and the expansion speed gradually slows down [53]. Hangzhou’s progression highlights how urban growth can shift from quantity-driven development to quality-driven adjustments as a city matures. Future land reclamation strategies could benefit from incorporating more sustainability-oriented measures, such as green infrastructure and environmental restoration.

4.2. Urban–Rural Integration and Socio-Economic Impacts

A central feature of Hangzhou’s CLR is its focus on suburban and urban–rural transition zones, particularly through Categories B and E streets. These areas, which underwent concentrated remediation between 2006 and 2015, represent Hangzhou’s efforts to balance urban expansion with rural development. Such integration is crucial in mitigating the social and economic gaps between urban and suburban populations. As the city expands outward, land reclamation can provide opportunities for infrastructure improvements, residential development, and job creation in formerly rural areas, thus improving the socio-economic conditions of these communities.

However, land reclamation also raises concerns about the environmental impact on local ecosystems, especially in the conversion of agricultural land to urban use. While reclamation can foster economic growth, it could also lead to the loss of biodiversity and agricultural capacity. To mitigate these challenges, it is essential that future reclamation projects adopt a holistic approach that considers both environmental sustainability and equitable development. Strategies such as integrating green infrastructure, creating buffer zones, and restoring ecosystems can help reduce the environmental footprint of urbanization. Additionally, careful planning is necessary to ensure that the benefits of land reclamation are distributed equitably across both urban and rural communities, fostering social inclusion and resilience of ecosystems.

As cities like Hangzhou, with ongoing and sustained expansion, it will be crucial to assess the long-term socio-economic impacts of CLR, including the effects on employment, income inequality, and community well-being. Future studies should explore how urban–rural integration can be optimized to promote sustainable development, balancing urban growth with the preservation of agricultural land and the protection of ecosystems.

4.3. Effectiveness of the Improved K-Means Clustering Algorithm

The application of the Improved K-means clustering algorithm in this study has proven to be an effective tool for categorizing CLR areas in Hangzhou. By enhancing the traditional K-means algorithm with adaptive data preprocessing and dynamic centroid adjustment, the method improved the silhouette coefficient by 11%, demonstrating better accuracy in clustering and providing more reliable insights into the patterns of CLR across the city. Compared to the K-means model in related studies, which typically involves clear clustering object mechanisms for feature variable selection and model improvement, this study adjusts the K-means model through machine learning to select feature variables for unknown mechanisms and causal relationships. After selecting from different variable combinations within the range of classification K-values, the study further refines the model using dynamic centroid adjustment, with the silhouette coefficient reflecting the current variable combination’s explanatory power for CLR [54].

This result underscores the value of advanced computational methods in urban planning, where clustering algorithms can help decision-makers identify key areas for intervention and optimize resource allocation. However, the algorithm’s effectiveness is still dependent on the quality of input data and the resolution of any outliers. Further refinement of clustering methods, including the integration of additional variables like demographic or economic factors, could lead to even more precise predictions and improve the decision-making process in land reclamation projects.

5. Conclusions

This study provides a comprehensive analysis of the patterns and impacts of CLR in Hangzhou over a fifteen-year period (2006–2020), utilizing an improved K-means clustering algorithm. The key innovation of this research lies in its application of a long-term, time-series analysis at the street level, offering valuable insights into the spatial and temporal dynamics of CLR. By identifying nine distinct categories based on varying stages of urbanization, ecological protection, and land-use management, this study reveals the complex transitions from rapid, large-scale reclamation to more targeted and sustainable land-use practices.

The clustering method used in this research significantly improves prediction accuracy by 11%, providing more reliable insights compared to traditional approaches. This advancement enables a deeper understanding of the evolving land-use patterns in Hangzhou, especially along the urban–rural boundaries where CLR is most concentrated. Furthermore, the study highlights the critical role of adaptive land management in balancing urban expansion with ecological preservation and agricultural sustainability, which is increasingly relevant as cities strive for sustainable growth in the face of environmental pressures.

This research not only contributes to the understanding of CLR in Hangzhou but also offers a methodological framework that can be applied to other rapidly urbanizing cities facing similar challenges. The findings underscore the importance of targeted land reclamation strategies that integrate ecological considerations, offering valuable guidance for urban planning and policy formulation aimed at achieving sustainable development. Due to data limitations, this study’s consideration of the CLR model mechanism remains insufficiently comprehensive. In the future, based on this study, more data such as population, economy, and relevant policies will be collected to further analyze the mechanisms of CLR changes and their influencing factors in order to more accurately predict the evolution of land use types and improve the decision-making process for land reclamation projects.