Next Article in Journal
The Impact of A1- and A2 β-Casein on Health Outcomes: A Comprehensive Review of Evidence from Human Studies
Previous Article in Journal
Mechatronic System for Maintaining the Homogeneity of Injectable Drugs in Syringe Pumps
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Applied Sciences
Volume 15
Issue 13
10.3390/app15137271
applsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

A Coupled Coordination and Network-Based Framework for Optimizing Green Stormwater Infrastructure Deployment: A Case Study in the Guangdong–Hong Kong–Macao Greater Bay Area

by
Jiayu Zhao
1,
Yichun Chen
1,
Rana Muhammad Adnan Ikram
2,
Haoyu Xu
1,
Soon Keat Tan
3 and
Mo Wang
1,*
1
College of Architecture and Urban Planning, Guangzhou University, Guangzhou 510006, China
2
Water Science and Environmental Research Centre, College of Chemistry and Environmental Engineering, Shenzhen University, Shenzhen 518060, China
3
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore 639798, Singapore
*
Author to whom correspondence should be addressed.
Appl. Sci. 2025, 15(13), 7271; https://doi.org/10.3390/app15137271
Submission received: 13 May 2025 / Revised: 16 June 2025 / Accepted: 17 June 2025 / Published: 27 June 2025
Browse Figures
Versions Notes

Abstract

Green Stormwater Infrastructure (GSI), as a nature-based solution, has gained widespread recognition for its role in mitigating urban flood risks and enhancing resilience. Equitable spatial distribution of GSI remains a pressing challenge, critical to harmonizing urban hydrological systems and maintaining ecological balance. However, the complexity of matching GSI supply with urban demand has limited comprehensive spatial assessments. This study introduces a quantitative framework to identify priority zones for GSI deployment and to evaluate supply–demand dynamics in the Guangdong–Hong Kong–Macao Greater Bay Area (GBA) using a coupled coordination simulation model. Clustering and proximity matrix analysis were applied to map spatial relationships across districts and to reveal underlying mismatches. Findings demonstrate significant spatial heterogeneity: over 90% of districts show imbalanced supply–demand coupling. Four spatial clusters were identified based on levels of GSI disparity. Economically advanced urban areas such as Guangzhou and Shenzhen showed high demand, while peripheral regions like Zhaoqing and Huizhou were characterized by oversupply and misaligned allocation. These results provide a systematic understanding of GSI distribution patterns, highlight priority intervention areas, and offer practical guidance for large-scale, equitable GSI planning.

1. Introduction

Urban agglomerations—characterized by high building density and strong inter-city connectivity—are particularly vulnerable to urban flooding due to extensive impervious surfaces and intensified land use, which significantly increase stormwater runoff volumes [1,2]. Uncoordinated development has historically resulted in the proliferation of impervious surfaces, exacerbating stormwater runoff and impairing natural hydrological processes. These disruptions significantly affect the social, economic, and ecological stability of urban environments [3,4,5]. To address these challenges, various decentralized stormwater management approaches have been adopted globally. Notable examples include the United States’ Low Impact Development (LID) practices [6]—which emphasize site-level techniques to mimic natural runoff patterns—and Japan’s Water Balancing Hydrological System. In recent years, these concepts have evolved into the broader framework of Green Stormwater Infrastructure (GSI), which builds on LID principles while incorporating additional ecological and social co-benefits such as habitat creation, temperature regulation, and enhanced urban amenity. In China, the “Sponge City” initiative integrates green, gray, and blue infrastructure for sustainable flood mitigation, with cities like Guangzhou and Shantou serving as demonstration areas for implementing GSI [7,8].
GSI refers to nature-based infrastructure solutions—such as green roofs, rain gardens, permeable pavements, and vegetated swales—that facilitate the retention, infiltration, and filtration of stormwater at the source. Beyond hydrological benefits, GSI contributes to biodiversity conservation, urban heat island mitigation, and improved public health and livability [9]. However, several implementation barriers persist, including spatial constraints, cost limitations, variability in soil permeability, and public resistance [10,11].
A critical barrier to effective GSI deployment is the spatial mismatch between supply and demand. In this study, GSI supply is defined as the environmental and infrastructural capacity of a district to deliver stormwater management services—captured by metrics such as green coverage, permeable surfaces, and waterway density. In contrast, GSI demand refers to the socio-hydrological need for such services, determined by factors including flood risk exposure, population vulnerability, and urban development intensity [12,13]. Equitable alignment between supply and demand is essential to ensure efficient infrastructure performance and social resilience. Yet a universally accepted framework to quantify and spatially evaluate this balance remains lacking, especially at sub-city or district scales.
The Greater Bay Area (GBA) of China—a megalopolis comprising 11 cities and 642 administrative districts—is a globally significant economic zone that faces acute hydrological stress (Figure 1). Geographically situated in a subtropical monsoon climate, the GBA is frequently subjected to extreme precipitation, typhoons, and seasonal storms. These meteorological extremes, coupled with rapid urbanization and extensive land development, have significantly reduced the region’s natural absorption capacity. Large swathes of impervious surfaces—especially in urban cores such as Guangzhou, Shenzhen, and Hong Kong—prevent effective infiltration, leading to high surface runoff volumes and increased urban flood vulnerability. Historical events such as Typhoon Mangkhut in 2018 and the intense rainstorms of 2022 have highlighted these vulnerabilities through widespread inundations and infrastructure disruptions. At the same time, the region exhibits several favorable conditions for the implementation of GSI, including policy support under China’s “Sponge City” initiative, an expanding urban ecological network in peripheral zones, and ongoing investment in integrated water management. The coexistence of extreme hydrological pressures and policy-driven opportunities makes the GBA a representative and urgent case for evaluating the spatial coordination of GSI systems. A systematic diagnosis of where GSI supply and demand are misaligned is therefore critical for mitigating flood risks and supporting resilient urban development across the region.
Hydrological models have contributed substantially to GSI planning, but their applicability is often constrained by high data requirements and low spatial granularity [14]. As an alternative, spatial indicators derived from biophysical and socio-economic datasets provide scalable means for demand assessment and scenario comparison across urban districts [15,16,17]. Nevertheless, few studies systematically integrate these data to evaluate localized supply–demand mismatches, leaving a critical gap in actionable spatial diagnostics for GSI planning.
To address this limitation, the present study introduces an integrated framework that combines Coupling Degree (CD) and Coupled Coordination Degree (CCD) models with proximity network clustering to assess spatial mismatches in GSI supply–demand dynamics across the Guangdong–Hong Kong–Macao GBA. The CD index captures internal interactions between supply and demand subsystems, while the CCD reflects their alignment with broader urban socio-ecological systems [18,19]. Proximity-based clustering further enables classification of districts with similar demand characteristics, thereby facilitating context-sensitive intervention strategies [20].
This study is guided by the hypothesis that significant spatial heterogeneity exists in GSI supply–demand relationships within the GBA, and that mismatches can be systematically identified through integrated CD/CCD modeling and network analysis. The objectives are threefold: (1) to construct a quantitative framework for assessing spatial alignment of GSI supply and demand; (2) to characterize district-level disparities using CD and CCD metrics; and (3) to identify priority demand clusters via proximity matrix analysis to inform more equitable and adaptive GSI planning.

2. Materials and Methods

Figure 2 illustrates the integrated analytical framework developed to evaluate the spatial alignment between the supply and demand of GSI in the GBA. The framework comprises three main components: I. Data Sourcing and Processing: This initial stage involves the compilation, georeferencing, and standardization of spatial indicators relevant to GSI supply and demand. Supply-side indicators include vegetation cover, waterway networks, and parks, while demand-side indicators encompass socio-demographic vulnerability, urban infrastructure, and flood exposure. All indicators are normalized using min–max scaling, with entropy weighting and Analytic Hierarchy Process (AHP) applied to assign integrated weights. II. Simulation of Supply–Demand Coupling: In this stage, composite indices for GSI supply and demand are calculated at the district level. The CD is used to measure the interaction strength between supply and demand subsystems, while the CCD assesses the overall alignment within broader urban socio-ecological contexts. Districts are classified into coordination categories based on CD and CCD thresholds, revealing degrees of imbalance. III. Proximity Matrix Construction and Network Clustering: To further interpret spatial disparities, a proximity matrix is generated based on the Revealed Comparative Advantage (RCA) of demand indicators. Using this matrix, a network of districts is formed through a maximum spanning tree algorithm and community detection (Fast-Newman method). The resulting clusters reflect similar demand profiles and serve as references for targeted, context-specific GSI planning.

2.1. Data Sourcing and Processing

Through a systematic quantification of supply and demand indicators, this study identifies mismatches in GSI allocation in the GBA, enabling more targeted resource distribution. Meteorological and hydrological datasets, including rainfall and streamflow records, were sourced from governmental archives and local reports. All spatial data were processed and geo-referenced using ArcGIS, with the regional hydrological network comprising approximately 2596 watercourses (see Figure 1).
The assessment of GSI in this study was anchored in a comprehensive set of supply and demand indicators. Data was sourced from established governmental and scientific repositories to ensure validity. Supply-side variables included parks, vegetation cover, impervious surface rate, road networks, and waterway density (Table 1).
To ensure data compatibility across formats, all spatial indicators were aggregated at the district level. Raster datasets (e.g., vegetation cover, impervious surface ratio) were averaged over each administrative unit. Polygon features (e.g., road and waterway networks) were converted to length density (km/km2). Point-based features (e.g., parks and medical facilities) were expressed as counts per km2. This standardized spatial resolution enabled integration into a unified indicator framework for supply and demand assessment. Due to the limitations in polygonal land parcel data, parks were represented by point features from OpenStreetMap and counted per district. While this simplifies spatial representation, it is acknowledged that using area-based metrics would provide more accurate GSI supply estimations.
These factors capture the biophysical capacity of each district to manage stormwater through retention, infiltration, and conveyance mechanisms. Demand-side variables covered urban flood risk, GDP, infrastructural facilities (e.g., transport and healthcare, as well as science and education institutions), historical heritage, and vulnerable population demographics (e.g., female, children, elderly, and low-education groups), as shown in Table 2.
Urban green spaces, such as parks, can function as buffers to absorb excess runoff and reduce peak discharge [23], provided they meet essential hydrological and design specifications. These include sufficient pervious surface area, soil infiltration capacity, vegetation cover, slope control, and surface detention features. The effectiveness of such spaces in stormwater attenuation is well captured in hydrodynamic models like the Storm Water Management Model (SWMM), which simulate interactions between precipitation, catchment characteristics, and drainage response. Vegetation cover enhances natural infiltration [21], while high impervious surfaces ratios are consistently associated with reduced absorptive capacity and greater flood risk [24]. Likewise, the density and configuration of roads and drainage networks shape stormwater conveyance efficiency and influence overall GSI functionality [25].
In terms of demand, population density and social vulnerability are key drivers of GSI needs [26]. Districts with a high concentration of children, elderly, or under-educated residents face elevated flood risks due to their reduced adaptive capacity—such as limited mobility, lower awareness of disaster protocols, and economic vulnerability—which constrains their ability to prepare for, respond to, and recover from flood events. Furthermore, areas rich in cultural or historical assets require robust GSI measures to protect against flood damage [17,27]. Economic variables, particularly GDP, influence both exposure levels and the institutional capacity to implement GSI [28,29].
To ensure consistency across diverse indicators, data were standardized using a min–max normalization technique [30]. Negative indicators, such as impervious surface rate, were inverted during processing. Entropy weighting (EW) was applied to reflect the informational contribution of each indicator objectively [31,32]. Higher entropy indicates greater variability and, hence, a higher weight in the composite index. The entropy formula and subsequent weight calculations followed standard multi-criteria decision-making procedures. In parallel, the Analytic Hierarchy Process (AHP) was employed to incorporate expert judgment. Indicators were structured into hierarchical levels, and pairwise comparisons were conducted to generate a judgment matrix. A consistency ratio (CR) test ensured the logical coherence of expert inputs.
To balance subjectivity and objectivity, EW and AHP weights were integrated using a linear combination approach [33,34]. The combined weight was adopted for each of the 16 indicators, as described in Equation (1):
Z j = a H j + 1 − a W j
where Z j , H j , and   W j represent the combined weightings, subjective weightings (AHP), and objective weightings (EW), respectively, and j is the supply and demand impact factor, where 1 ≤ j ≤ 16. The value of the coefficient a ranges from 0 to 1 and is taken as 0.5, since both subjective and objective factors are of equal importance.

2.2. Simulation of Supply and Demand Coupling

Composite indices for supply and demand were calculated using standardized values and integrated weights. The CD was computed to assess the internal interaction between supply and demand subsystems, while the CCD evaluated the degree of synchronization with external urban factors. The CD–CCD model categorizes districts into five coordination levels, ranging from severe imbalance (0.0–0.2) to strong synergy (0.7–1.0), based on established benchmarks [35].
The supply and demand levels of GSI in the GBA were calculated based on the respective normalized values as well as combined weightings. Whereas the coupling coordination model is composed of CD and CCD. The category of the coupling coordination level was determined based on reported work [35,36]. As such, the coupled coordination model can be used to elucidate the state of balance of GSI [8,37]. In addition, the magnitude of CD and CCD values can reflect the state of equilibrium between the two subsystems of GSI supply and demand. The GSI supply and demand levels are computed using Equation (2):
S x = ∑ j = 1 n Z j X j
where Z j denotes the combined linear weighting of indicator j, S x represents the combined level of supply and demand, and X j is the normalized value.
The GSI supply and demand relationship contains internal similarity and turning points, and can be used to classify clusters of different types. In light of the intrinsic characteristics of the data, natural divisions were employed to optimize the determination of classification breakpoints, full information utilization, and data visualization [8,38]. The GSI supply and demand levels were categorized as high, medium, and low using ArcGIS 10.5.
CD is formulated in accordance with Equation (3):
C D = S x D y S x + D y 2 2 2
where S x denotes the composite coefficient of the supply system, D y represents the composite coefficient of the demand system, C D is the degree of coupling; the larger the CD value, the better the state of balance of GSI supply–demand.
CCD is calculated using Equations (4) and (5):
T = α S x + β D y
C C D = C D × T
where CCD denotes the coordination index; T is a coefficient indicating the level of integrated supply and demand of GSI; and α and β represent the weightings of the supply and demand of GSI, and both values are set to 0.5 following the work of [39,40]. As such, the category of coupled coordination can be classified into five groups according to the range of the index values. The significance of these indices will be illustrated based on the findings of the case study.

2.3. Proximity Matrix and Network Clustering

To complement the CD and CCD assessments of supply–demand coordination, a Revealed Comparative Advantage (RCA)-based proximity matrix was constructed to identify and classify districts with disproportionately high GSI demand [41]. Unlike the CD/CCD framework, which evaluates the internal and external alignment between supply and demand systems, RCA is used here to diagnose demand-driven imbalance at the district level by comparing each indicator’s intensity to the regional average. Specifically, RCA scores were computed to determine whether a given district exhibited relative dominance in any particular vulnerability dimension (e.g., flood exposure or socio-demographic sensitivity). A district was considered to have a comparative demand advantage if its RCA score for an indicator exceeded 1. These scores formed the basis for calculating conditional proximity values, which quantify the likelihood of similarity between districts based on their vulnerability profiles. Pairwise proximity was then encoded in a matrix and converted into a proximity network using a maximum spanning tree algorithm [42,43]. The full modeling workflow is provided in Supplementary Materials—Modeling Workflow. This approach ensures that all nodes (districts) are connected via the most representative paths while minimizing redundancy. A total of 399 districts were included in this analysis. The proximity between districts c and d was calculated as follows:
R C A i ,   c = x i ,   c / ∑ i x i ,   c ∑ c x i ,   c / ∑ i ∑ c x i ,   c
P r c d = ∑ i R C A i ,   c · R C A i ,   d ∑ i R C A i ,   c 2 · ∑ i R C A i ,   d 2
P r o x i m i t y c ,   d = P r c d
where x i ,   c is the value of demand indicator i in district c and the denominator of Equation (6) represents the regional average of that indicator. These equations compute comparative vulnerability and mutual proximity, which inform the subsequent network clustering analysis for demand-driven GSI optimization.
The Fast-Newman modularity optimization method was used to detect community structures, identifying clusters of districts with similar vulnerability profiles. These clusters formed the basis for prioritizing GSI investment, offering spatial references and benchmarks for demand-based planning [44], see Equation (9):
Q = 1 2 m ∑ c , d B c d − k c k d 2 m δ h c , h d
where Q indicates the modularity of the system; B c d represents the weighting of the edge between district c and district d; k c is the sum of the weightings of the edges connected to the district c; and h c denotes the cluster of district c.

3. Results

3.1. Supply and Demand Levels of GSI

Figure 3 presents the spatial distribution of GSI supply and demand indicators across the GBA. The GSI supply indicators are conceptually divided into two categories: natural ecological services (e.g., vegetation cover and waterway networks) and artificial ecological services (e.g., parks and permeable pavements). This distinction facilitates a more nuanced interpretation of how different landscape components contribute to stormwater management. The spatial patterns of these indicators reveal substantial heterogeneity among the urban districts (Figure S1). In general, peripheral cities such as Zhaoqing and Huizhou demonstrate stronger capacities for natural ecological services, largely due to the presence of preserved green space and water bodies. Conversely, core cities like Guangzhou, Shenzhen, and Hong Kong feature more developed artificial infrastructure, including urban green parks and engineered drainage systems, reflective of their economic advancement (Figure S1a–e). These observations differ somewhat from an earlier study, which categorized natural GSI based on ecological land types (e.g., forests and wetlands), and artificial GSI based on constructed features such as green roofs and bioswales [39].
Demand-side indicators exhibit similar spatial clustering. High GSI demand is concentrated in the southeast and central urban zones of the GBA, including Guangzhou, Shenzhen, Hong Kong, and Dongguan. These areas face elevated flood risk due to high impervious surface ratios, aging drainage systems, and dense populations (Figure S1f–p). Districts rich in cultural heritage assets, such as historic architecture and protected landmarks, also exhibit heightened vulnerability and consequently a stronger demand for GSI interventions [27,45].
Figure 3 depicts the classified spatial distribution of GSI supply and demand across the region. Only nine districts exhibit a high level of GSI supply, all located in Hong Kong. In contrast, the majority of districts—particularly in Guangzhou (74%), Jiangmen (74%), and Shenzhen (46%)—fall into low or moderate supply categories. Demand patterns follow a similar trend; while 163 districts register high demand, 479 are categorized as having low to moderate demand. Notably, high-demand districts are concentrated in Guangzhou (56%), Shenzhen (31%), Foshan (41%), and Macao (50%), whereas low-demand zones are prominent in Huizhou (74%), Jiangmen (78%), Zhaoqing (89%), and Zhuhai (54%). Figure 3 further highlights the extent of spatial mismatches. In locations where supply exceeds demand, resources may be underutilized, pointing to inefficiencies in planning and investment. Conversely, districts with high demand but insufficient supply—such as urban cores—face compounded flood risk and infrastructure strain. This imbalance underscores the need for targeted GSI expansion in high-demand areas, particularly those with insufficient green coverage and vulnerable populations.
Across the GBA, the highest observed GSI supply score (0.53) is in Kowloon, Hong Kong, reflecting its proactive planning and strong institutional support. The lowest (0.076) is found in Da’ao, Jiangmen, likely due to limited technical capacity and lack of investment. For demand, the peak value (0.60) is recorded in Guangta Street, Yuexiu District, Guangzhou, a densely populated area with scarce green space. The lowest demand (0.002) is found in Baokou Town, Huizhou, which benefits from abundant vegetation and natural drainage capacity.

3.2. Coupling and Coordination of GSI Systems

Figure 4 illustrates the spatial distribution of the CD across the GBA. The average CD value is 0.68, indicating a moderate level of coupling between GSI supply and demand systems. However, significant regional disparities persist. The highest CD value, 0.99, is observed in Zhonglutan Town in Baiyun District, Guangzhou, reflecting near-optimal integration between supply capacity and demand intensity. In contrast, Baokou Town in Huidong County, Huizhou, records the lowest CD (0.00047), revealing a critical decoupling likely associated with misaligned infrastructure deployment.
254 districts are categorized as having “favorable coupling”, while 213 are identified as exhibiting “primary coupling”. These two groups constitute the majority of the region and are concentrated in economically advanced and administratively proactive areas, including Guangzhou, Shenzhen, Foshan, Dongguan, Hong Kong, and Macao. The elevated CD values in these cities can be attributed to a combination of integrated green infrastructure systems, strategic spatial planning, and supportive policy frameworks [46].
By contrast, lower levels of coupling are predominantly found in Huizhou, Zhaoqing, and Jiangmen. Although these areas possess substantial natural ecological resources, the lack of structured urban stormwater management systems hampers their capacity to respond effectively to flood risk. The underutilization of natural assets in these regions underscores the need for enhanced institutional capacity and policy intervention to improve GSI coordination [47].
The CCD, which integrates both internal coupling and external systemic alignment, reveals further dimensions of spatial disparity. While several districts with high CD also demonstrate high CCD—reflecting consistent alignment across ecological, infrastructural, and socio-economic domains—other districts remain weakly coordinated despite favorable environmental conditions. High CCD values are mainly concentrated in the Pearl River Delta, where cities exhibit strong governance, active investment in ecological infrastructure, and synchronized planning with social services. Kowloon City, Hong Kong, ranks highest with a CCD value of 0.65, falling within the “favorable coupling coordination” category. The district’s high population density, strong economic performance, and advanced green infrastructure contribute to its efficient and sustainable GSI system. Conversely, Baokou Town again registers the lowest CCD (0.008), placing it in the “extreme decoupling coordination” category. Despite a rich endowment of green space and water bodies, low population density and limited development result in minimal demand for GSI services, leading to inefficiency and underutilization of infrastructure. Mild to severe decoupling coordination is prevalent in peripheral areas such as Zhaoqing, Huizhou, and Jiangmen. These locations, although ecologically endowed, face systemic weaknesses in aligning GSI deployment with socio-economic needs.
To further delineate spatial mismatches, a quadrant analysis was conducted by combining CD and CCD classifications. This approach categorizes districts into four types based on their coordination profiles: (I) Coupled and Coordinated, (II) Not Coupled but Coordinated, (III) Not Coupled and Not Coordinated, and (IV) Coupled but Not Coordinated (see Figure 5).
A total of 266 districts fall into the third quadrant, “Not Coupled and Not Coordinated”, signaling critical misalignments between GSI infrastructure and local needs. These include substantial clusters in Zhaoqing (76 districts), Huizhou (37), Guangzhou (94), and Jiangmen (26). For instance, in Guangzhou—an urban core with high population and building density—green space provision has lagged behind rapid development, resulting in pronounced supply–demand imbalances. On the other hand, 62 districts are classified under the first quadrant, “Coupled and Coordinated”. These are primarily located in Shenzhen, Macao, and Hong Kong, where urban design integrates GSI systems effectively into the built environment and planning priorities. The quadrant-based approach facilitates identification of regions with either underutilized ecological potential or uncoordinated infrastructure expansion. Several districts have constructed substantial GSI systems without an adequate understanding of local demand, while others have developed rapidly without incorporating sustainable stormwater management.

3.3. High-Demand Proximity Network Matrix

Figure 6a illustrates the spatial clustering of districts within the GBA that exhibit high GSI demand based on proximity network analysis. Districts with marginal supply and demand characteristics were excluded from this analysis to ensure the clarity and relevance of clustering outcomes. A total of 399 districts were grouped into four clusters, defined by the relative prominence of specific demand indicators as determined through the RCA metric.
Cluster I includes nine districts, primarily located in Hong Kong and Macao (Figure 7). The defining features of this cluster include a high concentration of traffic facilities (90%), medical institutions (80%), and educational institutions (90%). These indicators reflect the vulnerability of compact, high-density built environments, where the disruption of essential services during urban flooding poses significant risks. Cluster II comprises 179 districts, distributed mainly across Guangzhou, Shenzhen, Zhuhai, and Zhongshan. This cluster is strongly associated with population structure indicators, including a high proportion of female residents (70.4%), children under the age of 14 (93.6%), elderly individuals over 60 (79.1%), and residents with less than a high school education (83.6%). These demographic characteristics point to heightened social vulnerability, suggesting that dense populations with limited adaptive capacity are at increased risk from urban flooding.
Cluster III, consisting of 124 districts, is primarily located in Dongguan, Foshan, and Zhaoqing. These districts exhibit extremely high probabilities of flood risk (98.4%), indicating their geographical susceptibility to extreme rainfall and storm events. As such, this cluster reflects physical exposure to hydrological hazards. Cluster IV encompasses 87 districts, mainly distributed across Hong Kong, Zhongshan, Shenzhen, and Zhaoqing. Districts in this group show strong associations with economic indicators, particularly GDP. High GDP may signal densely developed urban cores with significant impervious surfaces, thereby increasing flood vulnerability. Conversely, low GDP levels may imply a lack of institutional or fiscal capacity to invest in and maintain adequate GSI systems, thereby amplifying social vulnerability to stormwater impacts.

4. Discussion

With the ongoing impacts of climate change and the rapid pace of urbanization, urban flooding has become increasingly frequent and severe. Addressing this challenge requires the adoption of sustainable and effective strategies that can mitigate flood risks under evolving environmental conditions. Among these strategies, nature-based solutions—such as green parks, swales, bioretention cells, and vegetated filter strips—play a vital role in augmenting stormwater absorption and regulating runoff. At the same time, enhancing the capacity of existing drainage systems, many of which are no longer adequate to handle the intensified hydrological load, is equally important. However, in the context of large-scale urban agglomerations, the retrofitting or expansion of traditional drainage infrastructure poses significant financial and spatial constraints [46]. The inefficient and unbalanced allocation of GSI often undermines the coordinated development of both human settlements and the surrounding ecological environment [11].
The findings reveal that the current pattern of GSI deployment in the GBA is notably uneven. In some districts, this results in wasted public investment and underused infrastructure, while in others, it leads to serious deficiencies that impede sustainable development. Financial limitations, land constraints, and fragmented urban governance exacerbate the difficulty of implementing GSI at scale, especially given the spatial heterogeneity of the GBA [48]. By identifying areas with supply–demand imbalances and aligning investments with localized demand characteristics, planners can enhance flood resilience while promoting equitable resource distribution [49].

4.1. Factors Affecting the Imbalance of Coupling and Coordination

The quadrant analysis conducted in this study reveals that a significant number of districts in the GBA fall into categories characterized by poor coordination between GSI supply and demand. Each typology reflects distinct historical, spatial, and socio-economic influences. The key categories are discussed as follows:

4.1.1. Not Coupled and Not Coordinated

Districts classified under the “Not Coupled and Not Coordinated” category are typically located in older urban cores or suburban fringes. These areas often suffer from legacy planning issues, including excessive impervious surfaces, limited green space, and a dense concentration of culturally or historically significant buildings. In such settings, the stormwater management infrastructure is largely centralized—consisting of pipelines and detention ponds—without adequate provisions for infiltration or retention during extreme rainfall events [50]. Chang [51] argues that decentralized drainage systems are more effective in mitigating runoff volume in urban areas. Thus, for older districts, efforts should focus on preserving existing drainage facilities and public spaces while guiding redevelopment toward lower population densities and increased green coverage. In suburban areas, decentralized GSI interventions should be prioritized to restore ecological function and promote natural hydrological cycles.

4.1.2. Coupled but Not Coordinated

Districts falling into the “Coupled but Not Coordinated” category are primarily economically developed zones, often characterized by industrial parks and high-intensity land use. Examples include Guangming, Longgang, and Pingshan districts in Shenzhen; Chancheng and Sanshui districts in Foshan; and Doumen district in Zhuhai. Although these areas may exhibit technical coupling between GSI components, the spatial distribution of public facilities such as parks, schools, and transport infrastructure is uneven, leading to suboptimal coordination. A potential remedy lies in incorporating eco-efficiency principles into macro-scale planning. By integrating both traditional drainage infrastructure and green spaces, planners can foster synergistic development between human, social, and natural systems [52], transforming these areas into fully “Coupled and Coordinated” districts.

4.1.3. Not Coupled but Coordinated

The “Not Coupled but Coordinated” group is mainly composed of peripheral districts within the GBA. These areas tend to feature some ecological assets, such as natural water networks, but they face constraints in GSI effectiveness due to high population densities and fragmented impervious surfaces. Green space is limited and often dispersed. During redevelopment or land-use transformation, there is considerable potential to enhance GSI performance through strategic planning. Measures might include the expansion of parks, rehabilitation of drainage channels, installation of permeable paving, and improved integration with surrounding infrastructure. Such interventions can raise both the ecological and functional quality of GSI systems [53].

4.1.4. Coupled and Coordinated

Districts in the “Coupled and Coordinated” category—such as Bao’an in Shenzhen, Shunde in Foshan, and administrative regions in Hong Kong and Macao—exemplify a balanced integration of green infrastructure and urban development. These areas are supported by abundant ecological resources, including connected green spaces and well-distributed water networks. Li et al. [4] suggest that adopting micro-renewal strategies and multi-objective optimization techniques can further refine GSI functions at the community and street scales. Efforts such as building small ecological parks and pocket greenery enhance urban resilience while offering replicable models for other municipalities aiming to improve their own stormwater infrastructure systems.

4.2. Clustering Analysis for Identifying GSI Demand Areas in the GBA

As shown in Figure 7, Figure 8, Figure 9 and Figure 10, districts with demand primarily driven by population vulnerability and flood risk probability (i.e., Clusters II and III) tend to exhibit more severe imbalances between GSI supply and demand when compared with those driven by social infrastructure and economic indicators (i.e., Clusters I and IV). This suggests that Clusters II and III are more exposed to urban flood risks and possess weaker adaptive capacity.
The root causes of this disparity may lie in the historical neglect of population dynamics and environmental vulnerability in the urban planning and development phases of Clusters II and III [54]. For example, the rapid pace of urbanization has led to surging population growth, rising socio-economic inequality, intensified urban heat island effects, insufficient wastewater treatment, and overstressed public infrastructure—all of which increase vulnerability to flooding [55,56]. Additionally, accelerated urban expansion has been shown to significantly reduce soil water retention in various districts [57], compounding stormwater management challenges.
To address these imbalances, this study introduces the use of reference network maps as a tool to guide planners toward more equitable and adaptive GSI deployment strategies. These maps serve as graphical frameworks that allow planners to identify “reference” districts with similar demand characteristics and better-balanced supply–demand coordination. The concept is akin to those applied in various fields—such as social mapping [58], academic influence mapping [59], healthcare accessibility [60], and financial risk networks [22].
Each of the four demand-oriented clusters is linked to a representative district that can serve as a reference point for planning. Specifically, Cluster I is anchored by Wuhe District, Cluster II by Luzhou District, Cluster III by Baipenzhu District, and Cluster IV by Gaotan District. These reference districts are located near the geometric centers of their respective clusters, reflecting relatively balanced GSI supply–demand relationships. Planners can look to these benchmark areas when designing infrastructure interventions for districts with similar urban characteristics.
Figure 7, Figure 8, Figure 9 and Figure 10 display the proximity-based reference network maps for each cluster. In these diagrams, districts within the inner circle are strongly associated with the reference district based on high proximity scores, indicating similar vulnerability patterns and planning contexts. For instance, in Wuhe Township (Figure 7), high-probability reference relationships are established with Dong District (0.75), Shenshuipo District (0.80), and Huawangtong District (0.75). In Luzhou Town (Figure 8), similar patterns are observed with Hualong Town (0.70), Huicheng Street (0.83), and Zengjiang Street (0.70). In contrast, districts located in the outer circle of the reference network maps indicate weaker coordination between GSI supply and demand. These areas could benefit from adaptive strategies drawn from their inner circle counterparts or from potential cross-district collaborations to improve collective resilience. In some cases, districts with limited resources or fragmented infrastructure may be considered for integration or cooperative planning with neighboring municipalities.
Reference network maps have shown great potential in stormwater management, particularly when integrated with advanced techniques. For example, Zhong et al. [61] employed reference networks in combination with deep learning and image recognition to estimate urban flood depth, eliminating the need for additional hydrological sensors while improving real-time flood response.

4.3. Limitation and Future Research Perspectives

This study employed a coupled coordination model to quantitatively assess the spatial balance between GSI supply and demand in the GBA. While the research focuses on ecohydrological processes and their role in mitigating flood risks, the broader multifunctional benefits of GSI—particularly cultural ecosystem services—remain underexplored. These include improvements in air quality, reductions in urban heat island effects, and the provision of recreational green spaces that promote physical and mental health. Such benefits are especially critical in densely populated urban districts with limited access to public green areas and heightened social vulnerability.
GSI systems, by incorporating vegetation, wetlands, and other nature-based features, not only enhance flood resilience but also support broader ecological and social objectives. These include fostering community cohesion, reducing stress, and contributing to long-term public health. Despite this, the current study does not offer a systematic evaluation of these intangible benefits or their implications for social equity. Furthermore, although interactions among ecohydrological variables are modeled, the potential of GSI to support cultural engagement, well-being, and climate adaptation is insufficiently addressed [17]. Future research should consider integrating these dimensions to provide a more comprehensive understanding of GSI’s contribution to sustainable urban development [62,63].
While traditional GIS methods—such as buffer analysis and spatial overlay—are valuable for identifying physical proximity among land use features (e.g., roads, hydrological networks, parks, and heritage sites), this study adopts a more systemic approach to evaluate the functional interdependencies between GSI supply and demand. Specifically, the use of CD and CCD metrics enables the analysis of internal subsystem dynamics and external system alignment, capturing performance beyond spatial adjacency. This approach allows for a district-level understanding of GSI mismatches that accounts for both interaction intensity and alignment quality across biophysical and socio-hydrological indicators.
Moreover, the integration of entropy weighting and AHP-based expert judgment supports a more nuanced synthesis of heterogeneous spatial indicators, accounting for variations in data reliability and contextual importance. Proximity network clustering, based on RCA scores, further refines the analysis by identifying groups of districts with disproportionately high demand relative to the regional average—an insight that standard buffer approaches may overlook. Nevertheless, foundational GIS processing—such as land cover classification, spatial joins, and raster-to-vector conversion—was essential in the construction of these indicators and remains integral to the model’s spatial logic.
In addition, this study assumes relatively limited spatial variation in precipitation intensity across the GBA, based on its shared subtropical monsoon climate and regionally consistent exposure to extreme rainfall events. This assumption enables inter-district comparisons focused on GSI supply–demand dynamics, but also introduces constraints by excluding precipitation-specific variables. In practice, the temporal structure and intensity of rainfall, existing storm sewer network capacity, and the presence and distribution of stormwater reservoirs or detention basins are key factors influencing real-time flood behavior. The exclusion of these variables limits the study’s ability to model hydrodynamic performance and drainage efficiency under actual storm conditions. The absence of hydraulic modeling is largely attributable to the lack of high-resolution, publicly available data on underground drainage infrastructure across the GBA. However, the integration of such data—when accessible—would significantly enhance the analytical precision and operational value of GSI planning tools. Future research should aim to incorporate rainfall time series, sewer system topologies, and detention infrastructure into the CD/CCD framework, creating a fully integrated and hydrologically informed model of urban flood resilience.
Despite these methodological strengths, the study recognizes inherent limitations in the hierarchical evaluation process. Variations and potential biases in weighting methods—such as those introduced by expert judgment in the AHP or entropy weighting—may affect the consistency and accuracy of the coupling and coordination analysis. Additionally, the scope of this work is primarily confined to assessing GSI demand. Although all districts in the GBA were clustered based on indicators such as social infrastructure and population vulnerability, the analysis did not incorporate variations in GSI supply within the clustering framework. Future investigations should aim to include supply-side parameters alongside demand factors and also explore related infrastructure typologies—such as grey-green hybrid systems and three-dimensional urban forms—to develop a more holistic framework for urban stormwater resilience.

5. Conclusions

This study presents a novel integrated framework to assess and optimize the spatial balance of GSI supply and demand across large-scale urban agglomerations, using the GBA as a representative case. The main innovations and findings of the research are summarized as follows:
  • Novel Framework Design
A comprehensive analytical framework was developed by integrating CD, CCD, and proximity network analysis. This multi-method approach enables a nuanced assessment of spatial mismatches between GSI provision and urban demand.
  • Identification of Spatial Imbalances
Analysis of 642 districts in the GBA reveals pronounced spatial heterogeneity in GSI distribution. Central urban cores such as Guangzhou and Shenzhen exhibit under-supply due to dense populations and high impervious surface coverage, while peripheral regions like Zhaoqing and Huizhou show over-supply linked to low development intensity and ecological abundance.
  • Typological Classification of Districts
Districts were categorized into four types based on CD and CCD metrics: (1) Coupled and Coordinated, (2) Not Coupled but Coordinated, (3) Coupled but Not Coordinated, and (4) Not Coupled and Not Coordinated. Notably, 266 districts fall into the fourth category, highlighting areas in urgent need of targeted GSI interventions.
  • Demand-Driven Clustering
Proximity network analysis based on socio-demographic, infrastructural, economic, and flood exposure indicators identified four high-demand clusters. These clusters provide a differentiated understanding of local GSI needs and reveal systemic inequities in infrastructure allocation.
  • Reference Mapping for Planning Guidance
Reference network maps were constructed to visually represent district similarities within each demand cluster. These maps serve as practical tools for planners, offering benchmark cases to guide equitable and adaptive GSI deployment in districts with similar characteristics.

Supplementary Materials

The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/app15137271/s1, Figure S1: Spatial distribution of GSI-related supply and demand indicators in the GBA. (a) Parks; (b) impervious surface ratio; (c) vegetation cover rate; (d) road network density; (e) waterway network density; (f) urban waterlogging points; (g) gross domestic product (GDP); (h) traffic facilities; (i) medical institutions; (j) science and education institutions; (k) historical buildings; (l) cultural relics protection units; (m) population density (female); (n) population density (under 14 years); (o) population density (over 60 years); (p) population density (less than a high school education).; Table S1: Weightings for supply-side indicators (%); Table S2: Weightings for demand-side indicators (%).

Author Contributions

Conceptualization, M.W. and J.Z.; methodology, J.Z., Y.C. and R.M.A.L.; software, J.Z., Y.C. and H.X.; validation, R.M.A.L.; formal analysis, M.W. and J.Z.; investigation, J.Z., Y.C. and R.M.A.L.; resources, M.W. and S.K.T.; data curation, J.Z. and R.M.A.L.; writing—original draft preparation, M.W. and S.K.T.; writing—review and editing, M.W.; visualization, J.Z., Y.C. and H.X.; supervision, M.W.; project administration, M.W.; funding acquisition, M.W. and S.K.T. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by Guangdong Basic and Applied Basic Research Foundation, China [grant number 2023A1515030158, 2025A1515012916], Guangzhou City School (Institute) Enterprise Joint Funding Project, China [grant number 2024A03J0317], and the Major Program of the National Natural Science Foundation of China [grant number 62394334].

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors on request.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
AHPAnalytic Hierarchy Process
CDCoupling Degree
CCDCoupled Coordination Degree
CRConsistency Ratio
EWEntropy Weighting
GBAGreater Bay Area
GSIGreen Stormwater Infrastructure
LIDLow-Impact Development
RCARevealed Comparative Advantage

References

  1. Jiang, R.; Lu, H.; Yang, K.; Chen, D.; Zhou, J.; Yamazaki, D.; Pan, M.; Li, W.; Xu, N.; Yang, Y.; et al. Substantial increase in future fluvial flood risk projected in China’s major urban agglomerations. Commun. Earth Environ. 2023, 4, 389. [Google Scholar] [CrossRef]
  2. Wang, M.; Fu, X.; Zhang, D.; Chen, F.; Liu, M.; Zhou, S.; Su, J.; Tan, S.K. Assessing urban flooding risk in response to climate change and urbanization based on shared socio-economic pathways. Sci. Total Environ. 2023, 880, 163470. [Google Scholar] [CrossRef] [PubMed]
  3. Heidari, B.; Prideaux, V.; Jack, K.; Jaber, F.H. A planning framework to mitigate localized urban stormwater inlet flooding using distributed Green Stormwater Infrastructure at an urban scale: Case study of Dallas, Texas. J. Hydrol. 2023, 621, 129538. [Google Scholar] [CrossRef]
  4. Li, J.; Burian, S.J. Evaluating real-time control of stormwater drainage network and green stormwater infrastructure for enhancing flooding resilience under future rainfall projections. Resour. Conserv. Recycl. 2023, 198, 107123. [Google Scholar] [CrossRef]
  5. Nakamura, S.; Abanokova, K.; Dang, H.-A.H.; Takamatsu, S.; Pei, C.; Prospere, D. Is Climate Change Slowing the Urban Escalator Out of Poverty? Evidence from Chile, Colombia, and Indonesia. Int. J. Environ. Res. Public Health 2023, 20, 4865. [Google Scholar] [CrossRef]
  6. Dietz, M.E. Low Impact Development Practices: A Review of Current Research and Recommendations for Future Directions. Water Air Soil. Pollut. 2007, 186, 351–363. [Google Scholar] [CrossRef]
  7. Griffiths, J.; Chan, F.K.S.; Shao, M.; Zhu, F.; Higgitt, D.L. Interpretation and application of Sponge City guidelines in China. Philos. Trans. A Math. Phys. Eng. Sci. 2020, 378, 20190222. [Google Scholar] [CrossRef]
  8. Wang, M.; Chen, F.; Zhang, D.; Rao, Q.; Li, J.; Tan, S.K. Supply–Demand Evaluation of Green Stormwater Infrastructure (GSI) Based on the Model of Coupling Coordination. Int. J. Environ. Res. Public Health 2022, 19, 4742. [Google Scholar] [CrossRef]
  9. Bjørn, M.C.; Howe, A.G. Multifunctional bioretention basins as urban stepping stone habitats for wildflowers and pollinators. Urban For. Urban Green. 2023, 90, 128133. [Google Scholar] [CrossRef]
  10. Adhikari, B.; Perlman, R.; Rigden, A.; Walter, M.T.; Clark, S.; McPhillips, L. Field assessment of metal and base cation accumulation in green stormwater infrastructure soils. Sci. Total Environ. 2023, 875, 162500. [Google Scholar] [CrossRef]
  11. Solins, J.P.; Phillips de Lucas, A.K.; Brissette, L.E.G.; Morgan Grove, J.; Pickett, S.T.A.; Cadenasso, M.L. Regulatory requirements and voluntary interventions create contrasting distributions of green stormwater infrastructure in Baltimore, Maryland. Landsc. Urban Plan. 2023, 229, 104607. [Google Scholar] [CrossRef]
  12. Herreros-Cantis, P.; McPhearson, T. Mapping supply of and demand for ecosystem services to assess environmental justice in New York City. Ecol. Appl. 2021, 31, e02390. [Google Scholar] [CrossRef] [PubMed]
  13. Kuller, M.; Bach, P.M.; Ramirez-Lovering, D.; Deletic, A. What drives the location choice for water sensitive infrastructure in Melbourne, Australia? Landsc. Urban Plan. 2018, 175, 92–101. [Google Scholar] [CrossRef]
  14. Ambrogi Ferreira do Lago, C.; Abtin, S.H.; Mário, M.E.; Hofheinz Giacomoni, M. Simulation and optimization framework for evaluating the robustness of low-impact development placement solutions under climate change in a small urban catchment. Hydrol. Sci. J. 2023, 68, 2057–2074. [Google Scholar] [CrossRef]
  15. Bai, Y.; Wong, C.P.; Jiang, B.; Hughes, A.C.; Wang, M.; Wang, Q. Developing China’s Ecological Redline Policy using ecosystem services assessments for land use planning. Nat. Commun. 2018, 9, 3034. [Google Scholar] [CrossRef]
  16. Prettyman, K.; Babbar-Sebens, M.; Parrish, C.; Babbar-Sebens, J. A feasibility study of uninhabited aircraft systems for rapid and cost-effective plant stress monitoring at green stormwater infrastructure facilities. J. Hydroinform. 2020, 23, 417–437. [Google Scholar] [CrossRef]
  17. Sun, X.; Liu, H.; Liao, C.; Nong, H.; Yang, P. Understanding recreational ecosystem service supply-demand mismatch and social groups’ preferences: Implications for urban–rural planning. Landsc. Urban Plan. 2024, 241, 104903. [Google Scholar] [CrossRef]
  18. Kong, Q.; Kong, H.; Miao, S.; Zhang, Q.; Shi, J. Spatial Coupling Coordination Evaluation between Population Growth, Land Use and Housing Supply of Urban Agglomeration in China. Land 2022, 11, 1396. [Google Scholar] [CrossRef]
  19. Pan, Z.; Gao, G.; Fu, B.; Liu, S.; Wang, J.; He, J.; Liu, D. Exploring the historical and future spatial interaction relationship between urbanization and ecosystem services in the Yangtze River Basin, China. J. Clean. Prod. 2023, 428, 139401. [Google Scholar] [CrossRef]
  20. Wang, M.; Fu, X.; Zhang, D.; Lou, S.; Li, J.; Chen, F.; Li, S.; Tan, S.K. Urban agglomeration waterlogging hazard exposure assessment based on an integrated Naive Bayes classifier and complex network analysis. Nat. Hazards 2023, 118, 2173–2197. [Google Scholar] [CrossRef]
  21. Zhang, Q.; Wu, Z.; Zhang, H.; Dalla Fontana, G.; Tarolli, P. Identifying dominant factors of waterlogging events in metropolitan coastal cities: The case study of Guangzhou, China. J. Environ. Manag. 2020, 271, 110951. [Google Scholar] [CrossRef] [PubMed]
  22. Chen, B.-X.; Sun, Y.-L. Financial Market Connectedness between the U.S. and China: A New Perspective Based on Non-linear Causality Networks. J. Int. Financ. Mark. Inst. Money 2023, 90, 101886. [Google Scholar] [CrossRef]
  23. Orta-Ortiz, M.S.; Geneletti, D. Prioritizing urban nature-based solutions to support scaling-out strategies: A case study in Las Palmas de Gran Canaria. Environ. Impact Assess. Rev. 2023, 102, 107158. [Google Scholar] [CrossRef]
  24. Zhao, J.; Ke, E.; Wang, B.; Zhao, Y. An optimization model for the impervious surface spatial layout considering differences in hydrological unit conditions for urban waterlogging prevention in urban renewal. Ecol. Indic. 2024, 158, 111546. [Google Scholar] [CrossRef]
  25. Fňukalová, E.; Zýka, V.; Romportl, D. The Network of Green Infrastructure Based on Ecosystem Services Supply in Central Europe. Land 2021, 10, 592. [Google Scholar] [CrossRef]
  26. Fang, X.; Li, J.; Ma, Q. Integrating green infrastructure, ecosystem services and nature-based solutions for urban sustainability: A comprehensive literature review. Sustain. Cities Soc. 2023, 98, 104843. [Google Scholar] [CrossRef]
  27. Suárez, M.; Rieiro-Díaz, A.M.; Alba, D.; Langemeyer, J.; Gómez-Baggethun, E.; Ametzaga-Arregi, I. Urban resilience through green infrastructure: A framework for policy analysis applied to Madrid, Spain. Landsc. Urban Plan. 2024, 241, 104923. [Google Scholar] [CrossRef]
  28. Bodus, B.; O’Malley, K.; Dieter, G.; Gunawardana, C.; McDonald, W. Review of emerging contaminants in green stormwater infrastructure: Antibiotic resistance genes, microplastics, tire wear particles, PFAS, and temperature. Sci. Total Environ. 2024, 906, 167195. [Google Scholar] [CrossRef]
  29. Wang, M.; Yuan, H.; Zhang, D.; Qi, J.; Rao, Q.; Li, J.; Keat Tan, S. Supply-demand measurement and spatial allocation of Sponge facilities for Sponge city construction. Ecol. Indic. 2023, 148, 110141. [Google Scholar] [CrossRef]
  30. Xu, H.; Ma, C.; Lian, J.; Xu, K.; Chaima, E. Urban flooding risk assessment based on an integrated k-means cluster algorithm and improved entropy weight method in the region of Haikou, China. J. Hydrol. 2018, 563, 975–986. [Google Scholar] [CrossRef]
  31. Cabrera, J.; Lee, H.S. Flood risk assessment for Davao Oriental in the Philippines using geographic information system-based multi-criteria analysis and the maximum entropy model. J. Flood Risk Manag. 2020, 13, e12607. [Google Scholar] [CrossRef]
  32. Wu, J.; Chen, X.; Lu, J. Assessment of long and short-term flood risk using the multi-criteria analysis model with the AHP-Entropy method in Poyang Lake basin. Int. J. Disaster Risk Reduct. 2022, 75, 102968. [Google Scholar] [CrossRef]
  33. Sun, N.; Cailin, L.; Baoyun, G.; Xiaokai, S.; Yukai, Y.; Wang, Y. Urban flooding risk assessment based on FAHP–EWM combination weighting: A case study of Beijing. Geomat. Nat. Hazards Risk 2023, 14, 2240943. [Google Scholar] [CrossRef]
  34. Zeng, J.; Huang, G. Set pair analysis for karst waterlogging risk assessment based on AHP and entropy weight. Hydrol. Res. 2017, 49, 1143–1155. [Google Scholar] [CrossRef]
  35. Sun, Y.; Liu, S.; Dong, Y.; An, Y.; Shi, F.; Dong, S.; Liu, G. Spatio-temporal evolution scenarios and the coupling analysis of ecosystem services with land use change in China. Sci. Total Environ. 2019, 681, 211–225. [Google Scholar] [CrossRef]
  36. Yan, S.; Chen, H.; Quan, Q.; Liu, J. Evolution and coupled matching of ecosystem service supply and demand at different spatial scales in the Shandong Peninsula urban agglomeration, China. Ecol. Indic. 2023, 155, 111052. [Google Scholar] [CrossRef]
  37. Bi, Y.; Zheng, L.; Wang, Y.; Li, J.; Yang, H.; Zhang, B. Coupling relationship between urbanization and water-related ecosystem services in China’s Yangtze River economic Belt and its socio-ecological driving forces: A county-level perspective. Ecol. Indic. 2023, 146, 109871. [Google Scholar] [CrossRef]
  38. Milic, N.; Popovic, B.; Mijalkovic, S.; Marinkovic, D. The Influence of Data Classification Methods on Predictive Accuracy of Kernel Density Estimation Hotspot Maps. Int. Arab J. Inf. Technol. 2019, 16, 1053–1062. [Google Scholar]
  39. Xin, R.; Skov-Petersen, H.; Zeng, J.; Zhou, J.; Li, K.; Hu, J.; Liu, X.; Kong, J.; Wang, Q. Identifying key areas of imbalanced supply and demand of ecosystem services at the urban agglomeration scale: A case study of the Fujian Delta in China. Sci. Total Environ. 2021, 791, 148173. [Google Scholar] [CrossRef]
  40. Zhao, L.-T.; Li, F.-R.; Wang, D.-S. How to achieve the common developments of green finance and clean energy in China? Evidence from coupling coordination evaluation. Ecol. Indic. 2023, 155, 111011. [Google Scholar] [CrossRef]
  41. Boschma, R.; Minondo, A.; Navarro, M. The Emergence of New Industries at the Regional Level in Spain: A Proximity Approach Based on Product Relatedness. Econ. Geogr. 2013, 89, 29–51. [Google Scholar] [CrossRef]
  42. Klinger, B.; Barabasi, A.-L.; Hausmann, R. The Product Space Conditions the Development of Nations. Science 2007, 317, 482–487. [Google Scholar] [CrossRef]
  43. Wang, C.; Hou, Y.; Zhang, J.; Chen, W. Assessing the groundwater loss risk in Beijing based on ecosystem service supply and demand and the influencing factors. Sci. Total Environ. 2023, 872, 162255. [Google Scholar] [CrossRef] [PubMed]
  44. Fortunato, S.; Hric, D. Community detection in networks: A user guide. Phys. Rep. 2016, 659, 1–44. [Google Scholar] [CrossRef]
  45. D’Ambrosio, R.; Longobardi, A. Adapting drainage networks to the urban development: An assessment of different integrated approach alternatives for a sustainable flood risk mitigation in Northern Italy. Sustain. Cities Soc. 2023, 98, 104856. [Google Scholar] [CrossRef]
  46. Liu, W.; Zhang, X.; Feng, Q.; Yu, T.; Engel, B.A. Analyzing the impacts of topographic factors and land cover characteristics on waterlogging events in urban functional zones. Sci. Total Environ. 2023, 904, 166669. [Google Scholar] [CrossRef]
  47. Lee, T.-C.; Wong, W.-K.; Tam, K.-H. Urban-focused weather and climate services in Hong Kong. Geosci. Lett. 2018, 5, 18. [Google Scholar] [CrossRef]
  48. Zhang, C.; Xu, T.; Wang, T.; Zhao, Y. Spatial-temporal evolution of influencing mechanism of urban flooding in the Guangdong Hong Kong Macao greater bay area, China. Front. Earth Sci. 2023, 10, 1113997. [Google Scholar] [CrossRef]
  49. Pallathadka, A.; Sauer, J.; Chang, H.; Grimm, N.B. Urban flood risk and green infrastructure: Who is exposed to risk and who benefits from investment? A case study of three U.S. Cities. Landsc. Urban Plan. 2022, 223, 104417. [Google Scholar] [CrossRef]
  50. Wang, M.; Jiang, Z.; Zhang, D.; Zhang, Y.; Liu, M.; Rao, Q.; Li, J.; Keat Tan, S. Optimization of integrating life cycle cost and systematic resilience for grey-green stormwater infrastructure. Sustain. Cities Soc. 2023, 90, 104379. [Google Scholar] [CrossRef]
  51. Chang, H.-S.; Lin, Z.-H.; Hsu, Y.-Y. Planning for green infrastructure and mapping synergies and trade-offs: A case study in the Yanshuei River Basin, Taiwan. Urban For. Urban Green. 2021, 65, 127325. [Google Scholar] [CrossRef]
  52. Wessels, N.; Sitas, N.; O’Farrell, P.; Esler, K. Inclusion of ecosystem services in the management of municipal natural open space systems. People Nat. 2023, 6, 301–320. [Google Scholar] [CrossRef]
  53. Sun, F.; Zhang, J.; Xu, Y.-H.; Hu, W.; Cao, Y. Analysis of the relationship between supply–demand matching of selected ecosystem services and urban spatial governance: A case study of Suzhou, China. Environ. Sci. Pollut. Res. 2023, 30, 79789–79806. [Google Scholar] [CrossRef]
  54. Lyu, Y.; Wu, C. Managing the supply-demand mismatches and potential flows of ecosystem services from the perspective of regional integration: A case study of Hangzhou, China. Sci. Total Environ. 2023, 902, 165918. [Google Scholar] [CrossRef] [PubMed]
  55. He, B.-J. Cause-related injustice, process-related injustice, effect-related injustice and regional heat action planning priorities: An empirical study in Yangtze River Delta and Chengdu-Chongqing urban agglomerations. Landsc. Urban Plan. 2023, 237, 104800. [Google Scholar] [CrossRef]
  56. Jeon, Y.; Jung, S. Spatial Equity of Urban Park Distribution: Examining the Floating Population within Urban Park Catchment Areas in the Context of the 15-Minute City. Land 2024, 13, 24. [Google Scholar] [CrossRef]
  57. Wang, M.; Li, Y.; Yuan, H.; Zhou, S.; Wang, Y.; Adnan Ikram, R.M.; Li, J. An XGBoost-SHAP approach to quantifying morphological impact on urban flooding susceptibility. Ecol. Indic. 2023, 156, 111137. [Google Scholar] [CrossRef]
  58. Hughes, E.A.; Ellis, S.; Smith, J.R. Connecting groups and behaviours: A network analysis of identity-infused behaviours. Br. J. Soc. Psychol. 2024, 63, 205–233. [Google Scholar] [CrossRef]
  59. Gao, T.; Liu, J.; Pan, R.; Wang, H. Citation counts prediction of statistical publications based on multi-layer academic networks via neural network model. Expert Syst. Appl. 2024, 238, 121634. [Google Scholar] [CrossRef]
  60. Vieira, T.A.; Trapp, F.B.; Souza, C.F.M.; Faccini, L.S.; Jardim, L.B.; Schwartz, I.V.D.; Riegel, M.; Vargas, C.R.; Burin, M.G.; Leistner-Segal, S.; et al. Information and Diagnosis Networks—Tools to improve diagnosis and treatment for patients with rare genetic diseases. Genet. Mol. Biol. 2019, 42, 155–164. [Google Scholar] [CrossRef]
  61. Zhong, P.; Liu, Y.; Zheng, H.; Zhao, J. Detection of Urban Flood Inundation from Traffic Images Using Deep Learning Methods. Water Resour. Manag. 2024, 38, 287–301. [Google Scholar] [CrossRef]
  62. Sadeghi, K.M.; Kharaghani, S.; Tam, W.; Gaerlan, N.; Loáiciga, H. Green Stormwater Infrastructure (GSI) for Stormwater Management in the City of Los Angeles: Avalon Green Alleys Network. Environ. Process. 2019, 6, 265–281. [Google Scholar] [CrossRef]
  63. Luo, Z.; Tian, J.; Zeng, J.; Pilla, F. Assessing the spatial pattern of supply-demand mismatches in ecosystem flood regulation service: A case study in Xiamen. Appl. Geogr. 2023, 160, 103113. [Google Scholar] [CrossRef]
Applsci 15 07271 g001
Figure 1. Geographic extent of the study area and the distribution of primary waterway networks across the Guangdong–Hong Kong–Macao Greater Bay Area.
Figure 1. Geographic extent of the study area and the distribution of primary waterway networks across the Guangdong–Hong Kong–Macao Greater Bay Area.
Applsci 15 07271 g001
Applsci 15 07271 g002
Figure 2. Framework for optimizing the spatial balance of Green Stormwater Infrastructure (GSI) in the Greater Bay Area (GBA), comprising three components: (I) data sourcing and processing; (II) simulation of supply–demand coupling; and (III) proximity matrix construction and network clustering.
Figure 2. Framework for optimizing the spatial balance of Green Stormwater Infrastructure (GSI) in the Greater Bay Area (GBA), comprising three components: (I) data sourcing and processing; (II) simulation of supply–demand coupling; and (III) proximity matrix construction and network clustering.
Applsci 15 07271 g002
Applsci 15 07271 g003
Figure 3. Classified spatial distribution of GSI supply and demand across the GBA. (a) GSI supply levels. (b) GSI demand levels.
Figure 3. Classified spatial distribution of GSI supply and demand across the GBA. (a) GSI supply levels. (b) GSI demand levels.
Applsci 15 07271 g003
Applsci 15 07271 g004
Figure 4. Spatial patterns of GSI coupling and coordination in the GBA. (a) District-level classification of CD (Coupling Degree); (b) proportional distribution of CD levels; (c) district-level classification of CCD (Coupled Coordination Degree); and (d) proportional distribution of CCD levels.
Figure 4. Spatial patterns of GSI coupling and coordination in the GBA. (a) District-level classification of CD (Coupling Degree); (b) proportional distribution of CD levels; (c) district-level classification of CCD (Coupled Coordination Degree); and (d) proportional distribution of CCD levels.
Applsci 15 07271 g004
Applsci 15 07271 g005
Figure 5. Combined classification of GSI coupling and coordination across districts. (a) Quadrant-based typology derived from CD and CCD indices; and (b) spatial distribution of districts by quadrant classification in the GBA.
Figure 5. Combined classification of GSI coupling and coordination across districts. (a) Quadrant-based typology derived from CD and CCD indices; and (b) spatial distribution of districts by quadrant classification in the GBA.
Applsci 15 07271 g005
Applsci 15 07271 g006
Figure 6. Results of proximity network analysis identifying high-demand GSI clusters in the Greater Bay Area. (a) Network diagram showing the clustering of 399 districts based on proximity scores derived from their RCA in GSI demand indicators. Each node represents a district; node size reflects the relative intensity of GSI demand; node color indicates cluster membership; and edge thickness represents the strength of proximity (i.e., similarity in demand profile). (b) Geographic distribution of the clustered districts, where colors correspond to cluster membership as defined in panel (a).
Figure 6. Results of proximity network analysis identifying high-demand GSI clusters in the Greater Bay Area. (a) Network diagram showing the clustering of 399 districts based on proximity scores derived from their RCA in GSI demand indicators. Each node represents a district; node size reflects the relative intensity of GSI demand; node color indicates cluster membership; and edge thickness represents the strength of proximity (i.e., similarity in demand profile). (b) Geographic distribution of the clustered districts, where colors correspond to cluster membership as defined in panel (a).
Applsci 15 07271 g006
Applsci 15 07271 g007
Figure 7. Network reference maps of districts in Cluster I of the GBA.
Figure 7. Network reference maps of districts in Cluster I of the GBA.
Applsci 15 07271 g007
Applsci 15 07271 g008
Figure 8. Network reference maps of districts in Cluster II of the GBA.
Figure 8. Network reference maps of districts in Cluster II of the GBA.
Applsci 15 07271 g008
Applsci 15 07271 g009
Figure 9. Network reference maps of districts in Cluster III of the GBA.
Figure 9. Network reference maps of districts in Cluster III of the GBA.
Applsci 15 07271 g009
Applsci 15 07271 g010
Figure 10. Network reference maps of districts in Cluster IV of the GBA.
Figure 10. Network reference maps of districts in Cluster IV of the GBA.
Applsci 15 07271 g010
Table 1. Data sources and specifications for supply-side indicators of GSI in the Greater Bay Area.
Table 1. Data sources and specifications for supply-side indicators of GSI in the Greater Bay Area.
IndicatorData FormatSpatial ResolutionData SourceDescription/Unit
ParksShapefile (point)District-levelOpenStreetMap (www.openhistoricalmap.org, accessed on 16 May 2024)Number of green park points per km2 (proxy for pervious area)
Impervious Surface RatioRaster30 m[21]% of impermeable surface per district (%)
Vegetation Cover RateTIF30 mLandsat 8 Operational Land Imager% vegetation coverage per district (%)
Road NetworksShapefile (polygon)District-levelOpenStreetMap (www.openhistoricalmap.org, accessed on 18 June 2024)Road length density (km/km2)
Waterway NetworksShapefile (polygon)District-levelGuangdong Planning and Natural Resources Bureau, China (https://nr.gd.gov.cn/, accessed on 26 June 2024)Waterway length density (km/km2)
Table 2. Demand-side indicators with spatial structure and relation.
Table 2. Demand-side indicators with spatial structure and relation.
IndicatorData FormatSpatial ResolutionSpatial Relation to Real SpaceData Source
Science and Education InstitutionsShapefile (Point)District-levelCount of institutions per km2 (spatial join to district polygons)Gaode Map (https://lbs.amap.com/, accessed on 6 July 2024)
Traffic FacilitiesShapefile (Point)District-levelCount of transport nodes per km2 (e.g., bus/train stations)Gaode Map (https://lbs.amap.com/, accessed on 6 July 2024)
Medical InstitutionsShapefile (Point)District-levelCount per km2 (including hospitals, clinics)Gaode Map (https://lbs.amap.com/, accessed on 6 July 2024)
Population Density (Less than High School)ExcelDistrict-levelResidents with less than high school education per km2 (census-mapped)National Bureau of Statistics of China (www.stats.gov.cn, accessed on 28 July 2024)
Population Density (Under 14 Years)ExcelDistrict-levelChildren under 14 per km2 (census-mapped)National Bureau of Statistics of China (www.stats.gov.cn, accessed on 28 July 2024)
Population Density (Over 60 Years)ExcelDistrict-levelElderly over 60 per km2 (census-mapped)National Bureau of Statistics of China (www.stats.gov.cn, accessed on 28 July 2024)
Population Density (Female)ExcelDistrict-levelFemale population per km2 (census-mapped)National Bureau of Statistics of China (www.stats.gov.cn, accessed on 28 July 2024)
Historical BuildingsShapefile (Point)District-levelCount per km2 (mapped via spatial join)Guangdong Provincial Department of Housing and Urban-Rural Development (http://zfcxjst.gd.gov.cn, accessed on 30 July 2024)
Cultural Relics Protection UnitsShapefile (Point)District-levelCount per km2 (mapped via spatial join)Guangdong Provincial Department of Housing and Urban-Rural Development (http://zfcxjst.gd.gov.cn, accessed on 30 July 2024)
Gross Domestic ProductRaster1000 mAverage GDP per km2 extracted by zonal statistics to district polygonsGlobal GDP Gridded Dataset [22]
Urban Waterlogging RateShapefile (Point)District-levelCount of flood risk points per km2 (event-mapped spatial points)[8]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Zhao, J.; Chen, Y.; Ikram, R.M.A.; Xu, H.; Tan, S.K.; Wang, M. A Coupled Coordination and Network-Based Framework for Optimizing Green Stormwater Infrastructure Deployment: A Case Study in the Guangdong–Hong Kong–Macao Greater Bay Area. Appl. Sci. 2025, 15, 7271. https://doi.org/10.3390/app15137271

AMA Style

Zhao J, Chen Y, Ikram RMA, Xu H, Tan SK, Wang M. A Coupled Coordination and Network-Based Framework for Optimizing Green Stormwater Infrastructure Deployment: A Case Study in the Guangdong–Hong Kong–Macao Greater Bay Area. Applied Sciences. 2025; 15(13):7271. https://doi.org/10.3390/app15137271

Chicago/Turabian Style

Zhao, Jiayu, Yichun Chen, Rana Muhammad Adnan Ikram, Haoyu Xu, Soon Keat Tan, and Mo Wang. 2025. "A Coupled Coordination and Network-Based Framework for Optimizing Green Stormwater Infrastructure Deployment: A Case Study in the Guangdong–Hong Kong–Macao Greater Bay Area" Applied Sciences 15, no. 13: 7271. https://doi.org/10.3390/app15137271

APA Style

Zhao, J., Chen, Y., Ikram, R. M. A., Xu, H., Tan, S. K., & Wang, M. (2025). A Coupled Coordination and Network-Based Framework for Optimizing Green Stormwater Infrastructure Deployment: A Case Study in the Guangdong–Hong Kong–Macao Greater Bay Area. Applied Sciences, 15(13), 7271. https://doi.org/10.3390/app15137271

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop