Spatial structure and accessibility of urban parks: A multi-scale integration of GIS and space syntax in Xi'an

Abstract

Urban parks are critical for ecological sustainability, residents' well-being, and spatial equity. Using Xi'an, China, as a case study, this paper integrates park POI, census, and street network data within a GIS platform and applies space syntax theory to build a multi-scale segment model. Kernel density estimation, standard deviation ellipse, and syntactic indicators―Integration (NAIN), Choice (NACH), and Synergy, were employed to evaluate the spatial distribution and accessibility of municipal, district, and community parks across walking, cycling, and driving scales. Results reveal a “central agglomeration-peripheral sparsity” pattern, with clear differences across park hierarchies: municipal parks align with high-choice arterial axes, district parks cluster at medium-integration nodes, and community parks are embedded within locally integrated neighbourhoods. However, some corridors with high movement potential remain underserved, while peripheral areas show service gaps. To address these disparities, the study proposes an optimisation strategy of “arterial linkage-green core radiation-nodal coverage,” coupling the hierarchical park system with the city's intrinsic spatial logic. The findings demonstrate the value of space syntax in green infrastructure planning and equity assessment, advancing theoretical understanding of urban spatial governance while offering practical guidance for building resilient and inclusive green space networks.

Graphical abstract

Keywords

Urban parks / Space syntax / Spatial equity / Accessibility / GIS / Sustainable planning

1 Introduction

Urban parks, as an essential component of modern urban public space, play an irreplaceable role in improving the ecological environment, enhancing residents' well-being, and contributing to the creation of liveable cities (Chiesura, 2004; Thompson, 2002; Waitt and Knobel, 2018). Frederick Law Olmsted introduced the modern concept of the “park” in his works, emphasising the integration of the city with nature through planning and assigning multiple functions to public space, including aesthetics, recreation, education, and social interaction (Olmsted Jr and Kimball, 1822). In the context of contemporary urban development, urban parks are not only a vital part of the ecological system but also accommodate diverse needs in daily life, social interaction, and cultural expression.

However, with the continuous advance of urbanisation and the steady increase in population density, the imbalance between the supply and demand of urban green space has become increasingly prominent. In China, the national average per capita park green space has reached 14.8 m ². Yet, due to uneven spatial distribution, residents in certain central or peripheral urban areas still face difficulties in conveniently accessing green space services in their daily lives. This problem is particularly evident in mega-cities. For instance, Xi'an, as the core city of Northwest China, had achieved a built-up green coverage rate of 38.2% by 2021. Nevertheless, data from the Xi'an Municipal Bureau of Natural Resources and Planning indicate that the coverage of park green space within the 15-min living circle remains uneven. In several densely populated areas, green space resources are scarce, resulting in insufficient actual accessibility for residents.

The planning and evaluation of existing urban green space systems have long relied on macro-level indicators such as the green space ratio, green coverage rate, and per capita green space. Although these measures reflect the overall quantity of greenery, they are insufficient for accurately assessing the performance of green spaces in terms of spatial structure and service equity (Daniels et al., 2018; Özgüner, 2011; Peters, 2010). This limitation is particularly pronounced in large cities with complex spatial structures and diverse functions, where single indicators fail to reveal the actual roles and values of different levels of green space in residents' daily use.

In response to the above issues, recent studies have increasingly adopted a spatial-structural perspective to examine the rationality of urban park layouts and the spatial logic of service coverage. Existing research has shown that when the urban green space system forms a coherent, accessible, and well-structured spatial network, it not only enhances the integrity of the ecological system but also improves the equity and efficiency of green space services (Konijnendijk et al., 2013). Therefore, it is imperative to develop an analytical framework that integrates spatial structure, accessibility, and user demand, to inform more scientific strategies for urban park planning.

Taking Xi'an as a case study, this research integrates the characteristics of urban spatial form and population distribution, while accounting for differences in the service hierarchy of the park system. Urban parks are categorised into three types: municipal parks, which are large in scale and serve functions of cultural representation and ecological conservation; district parks, which provide services for neighbourhood residents and reflect regional historical and landscape features; and community parks, which cater to daily activities with an emphasis on accessibility and convenience. This classification helps to clarify the structural roles and service boundaries of different types of parks within the urban spatial system.

In terms of methodology, this study is based on a GIS platform and integrates POI data from GaoDe map, road network data from OpenStreetMap, and population census data of Xi'an. Spatial kernel density analysis, standard deviation ellipse analysis, and a multi-scale segment model of space syntax are employed to construct a multidimensional evaluation system for the spatial structure of urban parks. By defining service radii for walking, cycling, and driving modes, the study simulates accessibility differences under varying lifestyles. In addition, by applying the syntactic metrics of integration, choice, and synergy, the potential functions and organisational characteristics of the urban green space system are assessed from a spatial-structural perspective.

The findings reveal that the spatial layout of Xi'an's urban park system suffers from fragmentation, imbalance in the service hierarchy, and service blind spots in peripheral areas. A coherent green infrastructure network has yet to be established, which constrains both the ecological functions and public service capacity of urban green spaces. Therefore, it is urgent to propose targeted optimisation strategies for urban parks, grounded in spatial structural analysis and multi-scale accessibility evaluation.

In summary, this study pursues three main objectives:

(1) to clarify the spatial distribution characteristics and structural coupling mechanisms of the three-level urban park system in Xi'an from a multi-scale perspective.

(2) to evaluate the service accessibility of different types of parks under diverse travel modes, revealing their equity and structural interrelationships.

(3) to identify service gaps and spatial fractures, and to propose optimisation pathways for urban parks oriented towards “structural equity” and “ecological continuity.”

The study adopts the 15-min leisure circle as a key reference for evaluating equity, aiming to provide theoretical support and decision-making guidance for urban public space planning under the principle of “green equity.”

2 Literature review

With the acceleration of urbanisation, increasing land development intensity and rising population density, urban parks, as a key component of the urban green space system, play an increasingly vital role in improving the urban ecological environment and enhancing residents' quality of life (Wolch et al., 2014; Yigitcanlar and Teriman, 2015). Existing studies indicate that urban parks not only provide public spaces for leisure, exercise, and social interaction, but also deliver significant benefits in regulating the urban heat island effect, improving air quality, and supporting the health of urban ecosystems (Gunawardena et al., 2017; Yan et al., 2018; Yao et al., 2022). At the same time, green spaces serve as important regulators of mental health, helping to alleviate stress and improve emotional well-being (Burls, 2007; Gunawardena et al., 2017; van den Berg et al., 2010; Yan et al., 2018; Yao et al., 2022). Furthermore, Wolf et al. (2020) emphasise that increasing urban tree canopy coverage can substantially enhance residents' resilience, thereby providing theoretical support for public health interventions.

In the field of urban planning and design, the spatial layout of urban parks has gradually become a research focus. From Howard's concept of the “Garden City,” which emphasised the integration of urban and rural areas (Howard, 1965), to Wright's theory of “Broadacre City,” which advocated the unity of architecture and nature (Howard, 1965), and further to contemporary concepts such as the “sustainable city,” “green ecological city,” “low-carbon city,” and “smart city” (Batty et al., 2012; Beatley, 2012; Tzoulas et al., 2007), urban parks have evolved from being treated as isolated greening projects to becoming critical nodes in the integration of ecological networks and social systems. These theories have continuously enriched the strategies of urban park layout in both concept and practice, while also providing theoretical support for the development of a composite green infrastructure system.

In recent years, the accessibility of urban parks has been widely employed as a key indicator for assessing spatial equity and service efficiency (Sanesi and Chiarello, 2006). Related studies have found that, with the improvement of living standards, residents' concerns regarding parks have shifted from mere availability to ease of access, making accessibility a critical factor influencing the frequency of park use (Erkip, 1997; Jim and Chen, 2006). Moreover, urban streets, as an essential medium for achieving accessibility, not only accommodate traffic flows but also embody social attributes and cultural symbolism (Audirac, 2008; Qi et al., 2024). Accessibility, which describes the cost and ease of movement from an origin to a destination, has been extensively applied in evaluating the equity of urban service facilities (Mavoa et al., 2012).

With the support of GIS platforms and their capacity to integrate multi-source geographic data, scholars have gradually developed a variety of accessibility analysis methods. These include the buffer zone method (Khahro et al., 2023; Li et al., 2017), the travel distance method (García-Palomares et al., 2013; Park, 2012; Wang et al., 2021), the minimum proximity method (Black et al., 2004), the gravity model method (Mangmang and Yuncai, 2021), and the cost-weighted distance method (Kun et al., 2012; Liao et al., 2022). These approaches are of significant value in quantifying the service coverage of parks and identifying service blind spots (Fan and Cheng, 2022), thereby providing decision-making support for urban park layout. Talen (1998) was the first to propose the GIS-based concept of the “equity map,” which offered planners an operational pathway for evaluating the spatial equity of park resource allocation.

Although GIS has achieved significant progress in the analysis of physical accessibility, there remains a lack of indepth exploration of the cognitive and behavioural mechanisms embedded in the urban spatial structure itself. Space syntax, as an analytical approach based on spatial topology, can reveal the intrinsic logical relationship between urban form and human behaviour (Bafna, 2003; Hillier, 1998, 2012). Through indicators such as integration, choice, and synergy, the theory quantifies spatial accessibility and potential connectivity at both global and local levels, and has been widely applied to the structural optimisation of urban green networks and pedestrian flow prediction (Long et al., 2023). Unlike conventional GIS-based methods, space syntax places stronger emphasis on relational structures and cognitive pathways between spaces. It uncovers the potential interactive mechanisms between street systems and park use, thereby offering a new perspective for understanding residents' accessibility patterns and the attractiveness of urban parks.

In recent years, the integrated application of space syntax and GIS has gradually emerged in urban park research, providing technical support for more precise identification of the coupling relationship between spatial structure and the balance of service supply and demand (He et al., 2023; Verdú-Vázquez et al., 2021). While GIS platforms can efficiently process geographic information (Talen and Anselin, 1998), space syntax reveals the underlying logic of spatial structure. The combination of these approaches helps to examine accessibility differences across social groups and mechanisms of spatial exclusion from the perspective of structural equity.

Building on this research trend, the present study attempts to integrate GIS-based visual analysis with the topological modelling of space syntax to construct a multi-scale analytical framework for park services (Daniels et al., 2018; Li et al., 2017). On this basis, it systematically evaluates the interactive mechanisms of urban parks across three dimensions: physical distance, spatial structure, and social accessibility. This not only enhances the overall service efficiency of the urban green space system but also provides theoretical and methodological support for achieving the “15-min living circle” and for promoting a balanced layout of green infrastructure (Barthel et al., 2015). By combining quantitative modelling with structural analysis, this study contributes to enriching the research paradigm of spatial equity in urban parks and supports the construction of a more inclusive, equitable, and sustainable urban ecological network (Niu et al., 2018).

3 Datasets and methods

3.1 Study area

Xi'an is in Northwest China and serves as the capital of Shaanxi Province as well as a nationally recognised historical and cultural city. It was the capital of thirteen dynasties and the eastern starting point of the ancient Silk Road (Jia et al., 2023; Peters, 2021). The city covers a total area of approximately 10,096.89 km ² and administers nine districts and four counties. This study focuses on the built-up area of Xi'an as the study area, including the developed parts of Xincheng District, Beilin District, Lianhu District, Yanta District, Chang'an District, Weiyang District, and Baqiao District, with a total area of about 680 km ² (see Fig. 1). The built-up area has a permanent population of 5.77 million and a park density of 1.91 parks per km ². Compared with other parts of the city, the built-up area features higher spatial complexity, denser population distribution, and stronger demand for public services, making it both representative and data accessible.

The study area inherits the urban morphology of the Tang, Ming, and Qing dynasties. Centred on the Bell Tower, it has developed a cross-shaped axis of “North-South Avenue and East-West Avenue.” Within the Second Ring Road, the city presents a typical grid-like block structure, while beyond the Second Ring Road, it evolves into a hybrid form combining grid and radial-ring patterns. This has shaped a road network framework characterised as “two axes, three rings, and eight radials” (Jia et al., 2025; Jia et al., 2023), providing a solid foundation for examining the interaction between urban park layout and spatial structure.

3.2 Research framework

This study establishes a research framework for systematically assessing the spatial pattern and optimisation pathways of urban parks, based on multi-source spatial data integration and multi-dimensional spatial analysis methods (see Fig. 2). The research process is divided into four main stages: data acquisition – spatial analysis – structural cognition – planning recommendations, with details as follows:

Firstly, in the data acquisition stage, three authoritative sources of urban spatial data were collected: (1) POI (Point of Interest) data of urban parks in Xi'an were extracted via the DaoDe map Open Application Programming Interface (API) to obtain the spatial distribution of different park levels; (2) the latest population census data were obtained from the Xi'an Statistical Bureau to capture the spatial distribution of population density; (3) a complete road network dataset was retrieved from the OpenStreetMap platform to support spatial structural analysis. These three categories of data respectively correspond to the spatial layout of urban parks, the population service base, and the street topological structure, together forming the fundamental data support system for this study.

Subsequently, relying on multi-scale spatial analysis methods, the research conducts empirical analysis at the following three levels:

Secondly, this study constructs a research framework for urban parks based on multi-source data integration and multi-dimensional spatial analysis methods, with the aim of systematically examining the distribution characteristics and spatial planning pathways of urban parks in Xi'an. The research design follows a four-level logic of data acquisition – spatial analysis – structural cognition – planning discussion, with specific steps as follows:

Step 1: Analysis of urban park distribution characteristics. Based on POI data of urban parks, Kernel Density Estimation (KDE) and Standard Deviation Ellipse (SDE) methods are applied to reveal spatial clustering trends, main expansion directions, and structural deviations of park distribution. By overlaying urban park density with population density, a dual kernel density analysis is conducted to identify the “people–park ratio” across different areas, thereby evaluating population coverage and service equity of urban parks.

Step 2: Multi-scale spatial structural analysis. Using open street network data, a segment model is constructed and combined with space syntax theory to conduct multi-scale structural analysis under walking, cycling, and driving radii. Through this, the integration, choice, and synergy of urban parks are examined across different scales, and an accessibility model of the “living circle” is developed.

Step 3: Strategies for spatial configuration of urban parks. In line with the 15-min living circle, a theoretical discussion is undertaken regarding the distribution characteristics of urban parks and multi-level planning strategies in Xi'an. Optimisation pathways are proposed that combine service balance, structural integration, and ecological network construction, thereby providing research support for the development of an urban park system oriented towards “green equity.”

Finally, Comprehensive discussion. Drawing upon the above analyses, the study extracts pathways for structural optimisation, mechanisms for enhancing equity, and approaches to ecological network construction within the urban park system. It aims to provide a transferable theoretical framework and practical guidance for urban park planning in Xi'an as well as other large cities.

3.3 Data processing

3.3.1 POI data of urban parks

Point of Interest (POI) refers to a geographic entity with specific functions and spatial significance, containing location coordinates and semantic attributes such as name, type, and rank. In urban spatial research, POI data are typically obtained through mapping platforms and are widely applied to analyse the distribution of urban functions, spatial structures, and human activities. In this study, POI data are employed to extract the distribution and hierarchy of urban parks in Xi'an, thereby providing a foundation for spatial pattern and service accessibility analysis.

The urban park data derive from POI records of Xi'an in 2023. Under the “Scenic Spots” category, the keyword “park” was used to conduct a search, yielding a total of 3028 records, including parks, botanical gardens, zoos, memorial sites, religious places, and urban squares. After restricting the geographic scope and cross-checking with the 2019 Xi'an Territorial Spatial Survey, the Third National Land Survey, and official records of urban park construction, a total of 159 urban parks within the study area were identified (see Fig. 3).

Considering park type, target population, service radius, and scale, the parks were categorised into three levels: community parks, district parks, and municipal parks (see Table 1). Among these, community parks cover an area of 0.4–20 ha, serving residents within walking or cycling distance; district parks range from 20 to 40 ha, serving multiple neighbourhoods and providing more comprehensive facilities; and municipal parks exceed 40 ha, usually including large-scale facilities such as sports grounds or ecological green spaces, serving the entire city. This classification forms the basis for subsequent analyses of spatial distribution and layout strategies.

3.3.2 Data of population distribution

In this study, population density is adopted as a key social attribute variable to measure the equity of urban green space services and the responsiveness of spatial demand. To ensure both timeliness and reliability, the 2020 Seventh National Population Census data released by the National Bureau of Statistics were selected. This dataset covers all street and community units in Xi'an, with high statistical precision and strong representativeness. It provides the number of permanent residents at the administrative-unit level, thereby forming the basis for revealing the spatial distribution of population across different areas.

During the data processing stage, population statistics were first spatially matched with street- and community-level administrative boundaries using GIS software, thus linking attributes with spatial units. Population density (unit: persons/km ²) was then calculated based on the area of each unit, and a choropleth visualisation method was employed to illustrate spatial patterns of high- and low-density areas. These results not only provide essential support for the analysis of accessibility and equity of green space services but also supply reliable input data for subsequent spatial models such as kernel density estimation.

3.3.3 Data of urban road networks

The data of urban road networks were obtained from the open mapping platform OpenStreetMap (OSM), including information on road alignments, relational structures, and geographic coordinates, ensuring full coverage of the study area. The research scope spans 34.03º–34.57º N and 108.65º–109.22º E, covering the main urban districts and surrounding suburban areas of Xi'an. For space syntax analysis, the OSM data were first imported into AutoCAD to generate a complete axial map of the street system. This was then converted into a segment model using the “Convert Axial Map to Segments” function in Depthmap. In this process, continuous axial lines are split at turning nodes, producing a network model in which each segment serves as the minimum unit of analysis. Compared with the traditional axial model, the segment model more accurately captures turning relationships and connectivity between streets, making it better suited to represent actual walking routes and to support multi-scale spatial structural analysis.

In addition, to enhance comprehension, Fig. 4 presents the road network structure of Xi'an. It illustrates the main street structure and analytical boundary of the built-up area, providing readers with a more intuitive understanding of the spatial data basis employed in this study.

3.4 Date analysis

3.4.1 Kernel Density Estimation (KDE) method

In this study, the Kernel Density Estimation (KDE) method within the Geographic Information System (GIS) platform was applied, abstracting each park in the study area as a point feature. For small parks, the central point was selected to represent their location, while for linear or large parks, multiple representative points were chosen. The geographic coordinates of each urban park were extracted from the “Planning Cloud” Point of Interest (POI) information platform and imported into ArcGIS for KDE analysis. The results generated a continuous density surface, which identified high- and low-density areas and revealed the distribution patterns and spatial clustering trends of geographical phenomena (Fotheringham and Rogerson, 2013). To optimise the analytical outcome, three different search radii were tested on the citywide POI dataset of urban parks to determine the most appropriate scale.

\begin{array}{}& O\left(\text{x}\right)=\frac{1}{n{h}^2}{\displaystyle \sum _{i=1}^{n}k\left(\frac{x-x_i}{h}\right).}\end{array}

O(x) denotes the kernel density, h represents the search radius, k is the kernel function, n refers to the number of samples falling within the area, and x – x_i indicates the distance between the estimated point and the measured point.

3.4.2 Standard Deviation Ellipse (SDE) analysis

The Standard Deviation Ellipse (SDE) is a statistical method commonly applied in the analysis of spatial point data, designed to extract the distributional centre, directional trend, and degree of dispersion of spatial objects (Gong, 2002). The method takes the mean centre of all sample points as the centroid and constructs the ellipse's major and minor axes by calculating the standard deviations of the points along the X- and Y-axes. The major axis reflects the spatial extension and distribution trend of the point set in the principal direction, while the minor axis indicates the lateral dispersion perpendicular to this direction. The orientation angle of the ellipse is determined by the covariance matrix, thereby revealing the overall shift in direction, which is often closely related to the alignment of urban roads or axes of functional development. In this study, the SDE method was applied separately to municipal parks, district parks, and community parks in Xi'an, to identify their spatial clustering centres, dominant orientations, and coverage ranges. This provides both quantitative support and geometric evidence for exploring the coupling relationship between the hierarchical layout of parks and the urban spatial structure.

3.4.3 Multi-scale spatial analysis method

This study, grounded in space syntax theory, constructed a segment model covering the street network of the study area and adopted a multi-scale angular analysis approach, selecting three core syntactic measures: integration, choice, and synergy. Integration measures the overall accessibility of spatial units; choice reflects the frequency with which a segment acts as a mediator in shortest paths; and synergy evaluates the consistency between local structures and the overall spatial system (Hillier, 2008, 2009). To eliminate the influence of segment quantity on results, Normalised Angular Integration (NAIN) and Normalised Angular Choice (NACH) were employed for multi-scale calculations. NACH effectively addresses the bias of traditional choice in the process of scale extension, making it suitable for comparative analysis across different spatial morphologies and enabling a clearer understanding of hierarchical road structures. When combined with NAIN and synergy, it enhances the explanatory power for urban traffic and pedestrian flow patterns, thereby providing more comprehensive and precise technical support for analysing the spatial layout and structural accessibility of urban parks (Hillier et al., 2012). The following are their calculation formulae:

\begin{array}{}& \begin{array}{l}NACH\_r=\log \left(ACH\_r+1\right)/\log \left(ATD\_r+3\right);\hfill \\ NAIN\_r=ANC\_\widehat{r}1.2/ATD\_r.\hfill \end{array}\end{array}

In these equations, ACH refers to Angular Choice, ATD denotes the total depth, and ANC represents the number of segments.

Travel speeds vary across modes and environmental conditions: walking averages 5–7 km/h, cycling 15–20 km/h, and driving around 20–30 km/h in congested urban contexts, reaching 80–120 km/h on highways or open roads. These speeds are further influenced by factors such as age, health status, and terrain. To support the construction of the 15-min leisure circle, this study calculated the accessible distances within 5, 10, and 15 min (see Table 2), which were then used as the radius thresholds for the multi-scale spatial analysis.

4 Results

4.1 Distribution characteristics of urban parks in Xi'an

4.1.1 Overall distribution characteristics of urban parks in Xi'an

As shown in the figure, a structural relationship exists between the spatial distribution of urban parks and population density in Xi'an. Overall, the spatial distribution of urban parks demonstrates a “central agglomeration and peripheral sparsity” pattern, with the highest KDE values concentrated around Beilin District, Yanta District, and Lianhu District (see Fig. 5). The SDE indicates that the distribution centroid is in the urban core, with a slightly southwest–northeast orientation, reflecting the overlap between green space resources and the historical city centre. In contrast, the population density heat map in Fig. 6 shows a high degree of overlap with the KDE results, particularly in Beilin, Yanta, and Lianhu districts, which are both high-density population areas and zones of concentrated urban parks. This spatial consistency suggests that the urban core, owing to its earlier development and higher population density, has become the priority area for planning the urban green space system.

However, peripheral areas such as Chang'an District and the Chanba Ecological District, despite their large size, exhibit relatively low population density and sparse park distribution, indicating substantial room for improvement in green space provision. Against the backdrop of gradual population expansion towards the urban fringe, potential spatio-temporal mismatches between green space resources and population growth should be carefully addressed. Therefore, future planning of Xi'an's urban park system should, while maintaining the quality of green space in the core area, place greater emphasis on rapidly growing peripheral zones. This would facilitate a transition of park allocation from a “dense core” to “balanced coverage,” thereby enhancing the spatial equity and service efficiency of the overall green space system and supporting a people-oriented goal of sustainable urban development.

4.1.2 Distribution characteristics of municipal parks

In this study, municipal parks are defined according to their scale, generally characterised by larger areas, wider service radii, and higher-level ecological and recreational functions. The KDE results indicate that municipal parks are unevenly distributed, with a concentration in Yanta District in the south and in the central urban area, while the northern part of the city is relatively sparse. As shown in Fig. 7, the high-density zone exhibits a “single-core agglomeration” pattern.

The overlaid SDE demonstrates that their distribution is oriented towards the southern core area, which corresponds closely with the city's principal development axis. This concentrated pattern reflects the spatial inertia of land resource allocation, planning orientation, and site selection for large green spaces during urban development. The elongated elliptical form reveals that Xi'an's urban growth has long been oriented along the north–south axis, with relatively lagging expansion in the east–west direction, where park provision also remains insufficient. Although the 2011 International Horticultural Exposition stimulated an increase in the number of municipal parks in the northern area, land scarcity in high-density districts such as Beilin, Lianhu, and Xincheng has made the addition of large parks increasingly difficult. For example, the Daming Palace Heritage Park, despite its large size, is constrained in its service range by the railway station, heritage conservation requirements, and surrounding development. Looking ahead, greater attention should be paid to new districts in the north and south, while also addressing the shortage of green spaces in the east–west corridor, to achieve a more balanced spatial configuration of municipal parks.

However, a mismatch between the number of municipal parks and their service capacity has gradually emerged, particularly in the northern and western areas, where residential communities are dense but large green spaces are scarce, intensifying the imbalance between green space supply and demand. As municipal parks require stringent land conditions, their locations are often concentrated in areas with well-consolidated land and intensive redevelopment, which has, to some extent, reinforced their southward clustering. Therefore, future urban green space planning should, while safeguarding service functions in the core area, progressively expand towards the northern fringe. This would promote a more balanced spatial configuration of municipal parks and enhance the equity and accessibility of green space public services.

4.1.3 Distribution characteristics of district parks

In this study, district parks are defined according to their functional hierarchy and service radius, positioned between municipal parks and community parks, and primarily serving the functions of regional regulation and supplementary provision. The KDE results show that district parks are more dispersed in spatial distribution than municipal parks, forming a polycentric pattern. Hotspot areas are concentrated in the central, southwestern, and eastern parts of the city (see Fig. 8), without a single dominant core. This polycentric configuration reflects the balanced embedding of district parks across different urban functional zones, embodying, to a certain extent, the principle of “proximity-based service.”

The SDE exhibits a regular elliptical form, oriented slightly towards the northeast, indicating that the distribution centre lies along the eastern section of the city's central axis. Compared with municipal parks, district parks play a more evident balancing role in spatial coverage, being mainly located in transitional areas between the urban core and the periphery, thereby alleviating, to some extent, the service pressure caused by the insufficient provision of municipal parks.

Nevertheless, low-density zones of district parks remain evident in the northern and north-western parts of the city, where the green space service system is still incomplete and cannot fully meet the growing residential demand. Particularly under accelerated urbanisation, population growth and land development in peripheral areas highlight the urgent need for planning interventions to supplement regional green space facilities and achieve spatial equity in service provision. Therefore, future optimisation of the urban green space system should focus on addressing spatial gaps in district parks in peripheral areas. Through refined planning, the balance of green space services can be improved, while strengthening the complementary linkage among municipal, district, and community parks to construct a more hierarchically integrated network of urban green open spaces.

4.1.4 Distribution characteristics of community parks

As the smallest service-radius and most finely covered unit within the urban green space system, community parks primarily serve residential neighbourhoods by providing spaces for leisure and social interaction among residents. Owing to their proximity to housing areas and their far greater numbers compared with municipal and district parks, community parks exhibit strong spatial permeability and flexibility in site selection. As shown in Fig. 9, their overall distribution is relatively dense, displaying a distinct polycentric structure. High-density cores are mainly concentrated in the north-western and southern parts of the urban centre, with certain areas showing notable clustering phenomena.

The SDE results indicate a discernible north–south distribution trend, with the centroid shifting towards the southern part of the city. This pattern is closely associated with population distribution and the layout of high-density residential areas, reflecting that the siting of community parks is strongly influenced by residential concentration and the intensity of urban development. Particularly in the southern part of the city, the KDE values of community parks are significantly higher than in other areas, suggesting a strong daily demand for green spaces among residents in this zone.

However, the distribution of community parks remains sparse in the north-eastern and north-western fringe areas, where KDE values are generally low. This may be attributed to limited development intensity or lagging planning, and a complete community-level green space system has not yet been established. Priority should therefore be given to supplementing these areas in the future, to achieve spatial equity in green space services and enhance residents' well-being.

Compared with municipal and district parks, community parks provide broader coverage and more precise services, yet their layout is significantly shaped by the stage of urban development and the distribution of residential neigh-bourhoods. It is recommended that, in future urban renewal and peripheral expansion processes, the layout, accessibility, and balance of community parks be further optimised by taking into account permanent population density, the renovation of ageing communities, and the objectives of the 15-min leisure circle, thereby contributing to the creation of a more resilient and people-oriented network of green open spaces.

4.2 Spatial structural characteristics of urban parks in Xi'an

4.2.1 Topological accessibility―Integration

Figure 10 presents the results of two core syntactic measures at the global scale (Rn) of space syntax. Normalised Angular Integration (NAIN) is a key metric in space syntax for assessing the accessibility and structural cohesion of spatial units within the overall system. It is calculated as the sum of the shortest angular paths between a given axial line and all other lines in the system, thereby quantifying the degree of “integration” of that line within the spatial network. A higher integration value indicates that the space can be more easily reached from other areas, placing it in a more “central” position within the spatial structure, with stronger connectivity and organisational capacity. The use of normalised integration (NAIN) removes the bias caused by variations in network size, thus allowing comparability across different scales and cities.

In Fig. 10(a), areas with high NAIN values are primarily concentrated in Xi'an's traditional urban core, represented by extensive continuous red and orange segments. This demonstrates the strong integration capacity of this area within the overall urban spatial network, serving as a “spatial centre” that other areas can easily access via short paths. High-integration zones often correspond to commercial hubs, historical districts, transport nodes, or major pedestrian gathering spaces, functioning as the spatial foundation of urban vitality. For urban park planning, high-NAIN areas are suitable for locating centralised parks or multifunctional public spaces to increase usage frequency and social interaction. Furthermore, integration reflects the “legibility” and “spatial clarity” of the street system, carrying theoretical implications for spatial cognition and wayfinding.

Normalised Angular Choice (NACH), in contrast, measures the frequency with which a street segment is selected as part of the shortest angular paths across the entire system. It reflects the through-movement potential of a given path, quantifying how often it is “passed through” within the city's network. Normalisation enhances comparability, enabling NACH to reveal the structural value of a space as a connecting node within the system.

As shown in Fig. 10(b), the NACH results display a “cross-shaped plus radial” distribution, with main corridors high-lighted in blue, forming the backbone that connects the north–south and east–west directions of the city. These represent the primary thoroughfares most likely to be chosen as shortest paths. In practice, streets with high choice values usually correspond to trunk roads, express-ways, cross-town corridors, or transport axes, whose spatial form and functional attributes define their critical role in organising flows of people and goods. For the urban green space system, high-choice corridors serve as the foundation for developing greenways and slow-mobility systems, thereby enhancing the accessibility and continuity of parks. Moreover, choice reflects the “potential flow pressure” within urban space, providing valuable data for transport interventions and urban mitigation strategies.

In summary, combining the NAIN and NACH indicators reveals that Xi'an's spatial structure exhibits the pattern of “high integration in the core, high choice in the periphery.” While the urban core demonstrates strong cohesion and centrality, the peripheral areas rely on high-choice corridors to build connectivity frameworks and functional link-ages. This structure highlights the co-evolution of the transport network and functional layout, offering a structural orientation for urban green space planning. Park sitting should prioritise key nodes where high integration and high choice intersect, to improve the embeddedness and accessibility of green spaces within the urban network and to achieve an organic integration of public open space with urban spatial structure.

Figure 11(a) illustrates the scatterplot correlation between integration (NAIN) and choice (NACH) at the global scale (Rn), with the regression result of R² = 0.1266 indicating a weak to moderate positive correlation. This relatively weak coupling reflects the tension and disjunction in Xi'an's spatial structure between “aggregation” (overall accessibility) and “flow” (path mediation). In other words, some areas possess strong spatial integration capacity but are not major through-movement routes, and vice versa. This phenomenon is often observed in the spatial misalignment between historical districts and modern transport corridors, reflecting the asynchronous evolution of different functional layers of urban space. The value does not reach the threshold typically associated with highly legible cities, suggesting that the continuity of perceiving the whole urban structure from the local street network remains suboptimal. When considered alongside the distribution of urban parks, such structural imbalance poses challenges for the spatial accessibility and service equity of the public green space system.

As shown in Fig. 11(b), the star model further reveals structural differences between the “foreground” and “background” networks. While the foreground system (the trunk road network) exhibits pronounced peaks in choice values, forming a clear traffic backbone, the background network shows relatively low mean integration and choice values. This highlights a structural tension between “core and periphery,” exposing weak connectivity outside the main axes. This is also reflected in the relatively high standard deviation of NACH (0.4013), suggesting that urban movement potential is heavily concentrated along a small number of routes, while the spatial guiding capacity of most areas remains weak. Although this pattern facilitates the organisation of intensive traffic flows, it undermines spatial equity and accessibility in peripheral areas.

Table 3 presents the statistical characteristics of these two indicators. The mean NAIN is 0.977 with a standard deviation of 0.249, indicating that integration is relatively concentrated in the urban core, which demonstrates strong spatial organisation. By contrast, the mean NACH is 0.826 with a standard deviation of 0.401, suggesting a more dispersed distribution, where a limited number of routes bear substantial movement pressure―a typical traffic concentration effect.

From a space syntax perspective, this dataset highlights both critical problems and potentials of Xi'an's spatial structure. The city's main corridors possess high levels of accessibility and strong flow-controlling capacity, yet a large portion of local street systems is insufficiently embedded in the overall network, leaving peripheral areas in a relatively “passive” position within spatial organisation. Future planning should therefore strengthen the integration and continuity of the background network, reducing the over-concentration of traffic resources on a few routes, and enhancing spatial equity, resilience, and legibility. Such improvements hold direct relevance for the configuration of green spaces and park systems, providing structural guidance for developing a more balanced and liveable urban environment.

Overall, Xi'an's spatial system demonstrates a condition of being “structurally imbalanced though framework clear.” To enhance spatial equity and overall accessibility, it is recommended that future urban design strategies reinforce the continuity of the background network, strengthen linkages between peripheral areas and main corridors, and construct a multi-centred, multi-directional street framework. This would help distribute movement potential more evenly and guide the city's morphology towards greater resilience and inclusivity.

4.2.2 Geometric accessibility―choice

In space syntax analysis, Normalised Angular Choice (NACH) is employed to measure the frequency with which a spatial segment is “traversed” within the overall system, serving as an important indicator of path through-movement potential (Hillier, 2008). Using Xi'an as a case study, this research constructed a segment model and applied multi-scale radii ranging from 400 m to 20,000 m to simulate spatial flow structures under different travel modes, including walking (5–15 min), cycling, and driving. As shown in Table 4, the average distribution of different types of parks along streets was calculated. By dividing these values by the mean NACH at each scale, a ratio was obtained. A higher ratio indicates a stronger correspondence between the spatial distribution of different park types and the intensity of accessibility.

To present these differences more intuitively, the ratio values of different park types across multiple scales were plotted as line graphs, showing how dependency varies with scale (see Fig. 12). All three curves demonstrate an upward trend as the radius increases, yet with different growth rates and turning points, reflecting a typical scale-matching effect.

From the overall trend, the dependency curves of all three park types rise gradually as the accessibility radius expands. However, the growth magnitude and turning points differ significantly by type. At smaller scales, community parks exhibit the highest dependency; at medium scales, district parks rapidly catch up and eventually surpass community parks; while municipal parks consistently remain at the lowest level, though their curves converge gradually at larger scales. At the macro-level, the three curves approach each other, indicating a scale-matching relationship between different park hierarchies and corresponding levels of the street network.

Municipal parks, by contrast, consistently display the lowest dependency. Their values increase almost linearly from approximately 0.25 at NACH400 to around 1.3 at NACH20000, without marked turning points. This suggests that municipal parks are weakly tied to accessibility advantages in the street network; instead, their distribution is guided by macro-level planning and ecological patterns (e. g., rivers, green corridors, heritage sites). Their siting often emphasises ecological and landscape resources―such as rivers, lakes, hills, or cultural landmarks―rather than local or meso-scale accessibility. While this enhances ecological and scenic value, it creates a disjunction from the street network, keeping their dependency consistently low. Residents typically rely on cross-district transport systems, rather than simple street-level walking accessibility, to reach municipal parks.

District parks start with lower dependency at small scales, but at medium scales (NACH1200–6000) their curve rises steeply, eventually overtaking community parks and becoming dominant. This indicates a close association with meso-scale road networks, allowing them to provide broader services at the district level. However, once the scale extends to large radii (NACH9000 and above), the curve plateaus, stabilising at around 1.3–1.4, without further significant increase. This outcome reflects the functional positioning of district parks and their dependence on sub-arterial or district-level road networks. As their overall number is limited and unevenly distributed across districts, some areas suffer from supply deficits. In short, district parks can meet demand within specific scales, but their scarcity generates clear spatial inequalities in coverage.

Community parks maintain high dependency at the small-scale stage (NACH400–1200), showing strong alignment with neighbourhood-level street networks and effectively meeting residents' daily needs within the 15-min leisure circle. However, as the radius extends to medium and large scales (NACH2400 and above), their growth slows significantly and, in some intervals, falls behind district parks. This suggests that, at the macro-level, their distribution lacks balance and fails to achieve cross-district coverage. The underlying reason lies in planning logic: community parks are typically embedded within residential compounds and neighbourhood streets, forming a “locally embedded” pattern. This ensures high correspondence at small scales but results in insufficient coverage at larger scales.

Overall, the line graphs reveal clear structural issues in the spatial distribution of Xi'an's parks. Municipal parks are well-sited at the macro-level but poorly integrated with the street network, lacking accessibility advantages; district parks perform well at medium scales but remain limited in number and unevenly distributed; Community parks are highly embedded locally but lack macro-level balance. The fundamental cause lies in the hierarchical planning logic and developmental trajectory of Xi'an's park system: municipal parks have been shaped by ecological, historical, and policy imperatives, prioritising landscape value over transport accessibility; district parks are too few to provide consistent district-wide coverage; community parks were designed to meet daily neighbourhood needs, producing locally dense but spatially fragmented patterns. This distribution logic results in mismatches and disjunctions across scales, hindering the formation of a coherent park system that functions seamlessly at small, medium, and large scales.

As shown in Fig. 13, the spatial correspondence between large-scale choice and municipal parks demonstrates that such parks are predominantly situated along urban arterial roads and high-choice axes. Their service radius generally extends to a 15-min driving catchment, forming a clustered distribution with strong radiating power. Beyond their ecological and landscape landmark role at the city level, they also function as inter-district ecological and social hubs.

Secondly, as showed in Fig. 14, the spatial overlap between medium-scale choice and district parks reveals that district parks are typically located at transport junctions or on the boundaries of urban districts. This siting enables them to cover the service range of a 10-min cycling catchment, thereby exerting an integrative function across neighbourhoods at the meso-scale.

Finally, as shown in Fig. 15, the coupling relationship between small-scale choice and community parks indicates that these parks are generally anchored to street networks with high local accessibility and are thus evenly distributed. This demonstrates that community parks are highly aligned with residents' everyday walking routes, forming a spatial pattern supported by micro-accessibility.

Taken together, the three-tier park system―municipal, district, and community―presents an interwoven, hierarchical accessibility structure: municipal parks maintain macro-level ecological balance and symbolic significance, district parks provide meso-level spatial integration across neighbourhoods, and community parks ensure equitable coverage at the micro-level. The interplay of these layers forms an ordered, interconnected, and complementary network of urban green open spaces.

4.2.3 Perceptual accessibility―Synergy

Synergy is employed to evaluate the correlation between global and local integration, serving as a key indicator for measuring the perceived accessibility of urban parks (Hillier, 2008; Seamon, 2015; Xing and Guo, 2022). Global integration reflects the relationship between a spatial node and all other nodes within the system, while local integration captures its connectivity within a limited topological radius, thereby indicating local accessibility.

A higher level of synergy suggests that individuals are better able to perceive the overall urban structure through local spatial configurations (Hillier, 2008). Based on a validated segment model of Xi'an's Street network, this study applied multiple radii (400 m, 1200 m, 1600 m, 2400 m, 3000 m, 4000 m, 4800 m, 6000 m, and 9000 m) to compute local integration using depthmap. Scatterplots were then produced between local and global integration, with fitted regression lines to derive the coefficient of determination (R ²) as a measure of synergy (see Table 5).

The results indicate that when the analytical radius is below 2000 m, the correlation between local and global structures is weak, implying that the city's spatial structure is not easily perceived at the neighbourhood scale and perceived accessibility remains low. Once the radius reaches 3000 m or above, however, synergy increases markedly, revealing a stronger structural correspondence between park locations and the street network. The 4000 m radius emerges as an optimal balance point, as it preserves local structural detail while simultaneously capturing the overall characteristics of the spatial network.

Municipal-level parks fulfil dual functions of city-wide service provision and symbolic representation, relying predominantly on motorised travel and serving extensive catchment areas. When the analytical radius reaches 4800–9000 m, synergy exceeds 0.71, suggesting that the integrity and legibility of the urban structure are strongest at this scale, thereby supporting the systemic organisation and rapid recognition of public spaces. While finer-grained details of the street network become less visible at this level, the structural tension between the urban skeleton and landmark nodes is accentuated. Consequently, municipal parks should be preferentially located in high-synergy areas, enhancing their accessibility, recognisability, and symbolic value within the wider urban system.

District-level parks serve wider catchments such as subdistricts or administrative areas and are generally accessed via cycling or short-distance public transport. Their effective functioning requires a balance between local accessibility and a degree of spatial legibility. The findings reveal that synergy rises markedly within the 2400–4000 m range (from 0.43 to 0.65), with the 4000 m radius emerging as the optimal scale that preserves the detail of the street network while simultaneously reflecting the city's structural coherence (see Fig. 16). Medium-synergy areas should therefore be prioritised for the siting of district-level parks, as they demonstrate both structural adaptability and service efficiency.

Community parks primarily provide high-frequency, proximity-based services, typically accessible within a 5–10-min walking distance and are strongly dependent on the connectivity and density of the local street network. As shown in Table 5, synergy remains consistently low (<0.3) within the 400–1600 m radius, indicating that the overall urban structure cannot be perceived at this scale, although a strong sense of local identity is preserved. This feature aligns with the service logic of community parks, embedding them into the micro-scale spatial routines of everyday life. Accordingly, low-synergy areas should not be regarded as structurally disadvantaged, but rather as a critical foundation for satisfying neighbourhood-level green space needs.

Despite a degree of structural correspondence between the existing park hierarchy and the street network, space syntax analysis exposes critical mismatches. On the one hand, several high-choice arterial corridors lack municipal or district-level parks, failing to establish continuous green corridors or gateway nodes, thereby weakening ecological and spatial support along the principal urban axes. On the other hand, peripheral areas of low-choice streets―such as new development zones or low-density urban edges―are underserved by community parks, revealing structural discontinuities and service blind spots.

These findings highlight that although Xi'an has developed a multi-level park hierarchy, its spatial arrangement has not been fully informed by accessibility structures. Future planning should therefore adopt a network-based strategy: strengthening the alignment of municipal parks with high-choice corridors and ecological greenways, enhancing the capacity of medium-accessibility structures to support district-level parks, and densifying the provision of community parks in low-accessibility areas. Such a framework would establish a three-tiered green equity system―comprising primary corridors, secondary nodes, and micro-scale services―thereby promoting both spatial justice in park distribution and the integrated enhancement of the urban ecological network.

4.3 Strategies for the spatial configuration of urban parks

As an essential component of green infrastructure, the spatial configuration of urban parks not only concerns the equity of public service provision but also directly affects residents' quality of life and the resilience of urban spaces (Khatibi et al., 2024, 2025; Mahdzar, 2008). Based on the functional hierarchy and service radii of municipal, district, and community parks, this study introduces an isochronous accessibility model of the “5-min walking circle, 10-min cycling circle, and 15-min driving circle” to construct a hierarchical and responsive spatial framework.

As shown in Fig. 17, the three-tiered park system exhibits a spatial pattern of “axial linkage–green-core radiation–point-like coverage,” which closely aligns with the core axes of high integration and high choice identified through space syntax analysis. This configuration not only reflects the logic of function deployment under structural dominance but also reveals how the green space network, through multi-scalar synergy, achieves layered effectiveness―from supporting everyday micro-scale life to shaping the macro-scale image of the city.

Municipal parks (see Fig. 17(a)) are characterised by their limited number, concentrated scale, and strong radiation capacity. They are typically distributed along urban arterial roads and high-choice axial lines, forming a “linear-cluster” pattern. These parks not only provide city-level ecological services and function as landscape landmarks but also serve a dual role within the urban green space system as ecological corridor nodes and social activity hubs, supporting cross-regional spatial connectivity and ecological balance.

District-level parks (see Fig. 17(b)) exhibit a “core green space–radiating circle” pattern and are often located at traffic intersections or the boundaries of urban sectors, undertaking an integrative function across multiple communities. Internally, they connect to clusters of community parks to provide centralised support for daily activities, while externally they link hierarchically with municipal parks, acting as intermediate nodes in the system. The optimisation of district-level parks should be combined with the allocation of public service facilities such as education, healthcare, and culture, thereby enhancing both their spatial integration and functional hybridity.

Community parks (see Fig. 17(c)) are characterised by a relatively balanced point-like distribution, embedded within locally high-integration street networks to form a “micro-accessibility model.” These parks commonly take the form of corner greens, pocket parks, and linear green strips, emphasising every day, high-frequency, and short-duration use. They are particularly important in providing accessible and inclusive services for children, the elderly, and other vulnerable groups. By embedding community parks into the urban fabric, not only are the demands of the 5–10 min living circle met, but street-level vitality and the density of social interaction are also significantly enhanced.

In summary, as illustrated in Fig. 16, the three-tiered system of municipal, district, and community parks demonstrates a complementary spatial and functional relationship: municipal parks ensure cross-regional balance, district-level parks achieve sectoral integration, and community parks support micro-scale daily accessibility. Together, they constitute a hierarchically ordered, interconnected, and mutually supportive network of urban green open spaces.

5 Discussion

With the widespread promotion of the “15-min living circle” concept in global urban governance, urban parks, as key carriers of green public resources, have increasingly become critical indicators of urban liveability and spatial equity (Chang and Liao, 2011; Zhang et al., 2022). By integrating GIS and space syntax, this study systematically analyses the spatial distribution patterns, hierarchical structural differences, and service accessibility of Xi'an's three-tier park system from a multi-scalar perspective, aiming to explore the dynamic balance between structural logic and social equity, and to provide theoretical support for the optimisation of green infrastructure.

Firstly, kernel density analysis based on POI and population density reveals that Xi'an's urban park system demonstrates a pattern of “multi-core clustering with weak peripheries”: municipal parks are distributed in a linear belt along high-level corridors; district-level parks are concentrated in sectoral core nodes; and community parks are embedded in high-density residential areas to meet daily needs. Space syntax further uncovers the structural logic underlying this pattern: integration (NAIN) reflects the organisational advantage of park hierarchies, choice (NACH) indicates traffic potential, and synergy reveals the cognitive structure of space. Different categories of parks correspond significantly to these syntactic indicators.

However, the analysis also highlights structural imbalances: some high-choice axes lack the support of municipal parks, while low-integration peripheral zones exhibit insufficient community green spaces, resulting in service blind spots and structural fragmentation. The introduction of synergy further expands the cognitive dimension of accessibility assessment: high-synergy areas demonstrate stronger structural legibility and are therefore more suitable for municipal park allocation, whereas low-synergy areas reveal risks of cognitive inaccessibility.

This study demonstrates that space syntax offers unique advantages in structural insight and cognitive simulation for urban park system analysis, providing methodological support for achieving spatial equity and ecological sustainability. For the future planning of Xi'an, three priorities should be emphasised: first, strengthening the layout of municipal parks along high-choice arterial axes to construct gateway nodes and ecological corridors; second, optimising the distribution of district-level parks within highly integrated areas to expand service radii and enhance shared functions; and third, embedding community parks precisely within high-density neighbourhoods to ensure walkable accessibility and equitable coverage. Furthermore, space syntax should be incorporated into the planning toolkit for green spaces, combined with socio-economic and land-use data, to establish a dynamic monitoring mechanism that fosters the long-term coordination of green infrastructure and urban spatial structure.

6 Conclusions

This study, taking Xi'an as a case, integrates GIS-based kernel density analysis with space syntax to construct a hierarchical distribution model and service circle evaluation system for urban parks. It systematically reveals the spatial distribution characteristics and locational mechanisms of municipal, district, and community parks. The findings indicate that the three tiers of parks are associated with different levels of syntactic indicators: municipal parks align with high-choice arterial corridors, district parks are organised around medium-integration junction nodes, and community parks are embedded within highly locally integrated street networks. Together, they establish a spatial coordination pattern characterised by “structural orientation–hierarchical stratification–functional matching.”

From the perspective of spatial equity, the study identifies deficiencies in green space provision in certain peripheral urban areas, where structurally weak zones often coincide with the absence of parks, creating service vacuums. This demonstrates that the layout of the urban green space system requires further optimisation to achieve a more balanced allocation and efficient use of public spatial resources. In addition, synergy analysis supplements the evaluation of the “perceived accessibility” of parks, thereby providing theoretical support for the construction of perception-friendly urban green space systems in the future.

The novelty of this study lies in integrating the multidimensional syntactic indicators with the hierarchical park system, and in proposing an optimisation framework for park spatial planning from a “structure–hierarchy–service circle” perspective. This framework offers significant policy-oriented value. It is recommended that future urban green space planning adopt a strategy combining structural guidance with multi-scalar service coordination, promoting the formation of a three-tier park network of “arterial axes–green cores–grid nodes,” thereby supporting the city's transition towards high-quality green development.

The study also recognises its limitations. The data primarily rely on POI information and vector street networks, lacking dynamic observations of residents' actual usage behaviour. Future research could incorporate mobility trajectory data, questionnaire surveys, and other approaches to further investigate the interaction between behavioural preferences and spatial structure. Additionally, the integration of terrain, land use, and socio-economic indicators would enable the construction of a more comprehensive evaluation system for green spatial equity, providing more precise data support for urban green governance.

References

[1] Audirac, I., 2008. Accessing transit as universal design. J. Plann. Lit. 23 (1), 4–16.

[2] Bafna, S., 2003. Space syntax: a brief introduction to its logic and analytical techniques. Environ. Behav. 35 (1), 17–29.

[3] Barthel, S., Parker, J., Ernstson, H., 2015. Food and green space in cities: a resilience lens on gardens and urban environmental movements. Urban Stud. 52 (7), 1321–1338.

[4] Batty, M., Axhausen, K.W., Giannotti, F., Pozdnoukhov, A., Bazzani, A., Wachowicz, M., Ouzounis, G., Portugali, Y., 2012. Smart cities of the future. Eur. Phys. J. Spec. Top. 214, 481–518.

[5] Beatley, T., 2012. Green Urbanism: Learning from European Cities. Island Press.

[6] Black, M., Ebener, S., Aguilar, P.N., Vidaurre, M., El Morjani, Z., 2004. Using GIS to Measure Physical Accessibility to Health Care. World Health Organization, pp. 3–4.

[7] Burls, A., 2007. People and green spaces: promoting public health and mental well-being through ecotherapy. J. Publ. Ment. Health 6 (3), 24–39.

[8] Chang, H.-S., Liao, C.-H., 2011. Exploring an integrated method for measuring the relative spatial equity in public facilities in the context of urban parks. Cities 28 (5), 361–371.

[9] Chiesura, A., 2004. The role of urban parks for the sustainable city. Landsc. Urban Plann. 68 (1), 129–138.

[10] Daniels, B., Zaunbrecher, B.S., Paas, B., Ottermanns, R., Ziefle, M., Roß-Nickoll, M., 2018. Assessment of urban green space structures and their quality from a multidimensional perspective. Sci. Total Environ. 615, 1364–1378.

[11] Erkip, F.B., 1997. The distribution of urban public services: the case of parks and recreational services in Ankara. Cities 14 (6), 353–361.

[12] Fan, Y., Cheng, Y., 2022. A layout optimization approach to urban park green spaces based on accessibility evaluation: a case study of the central area in Wuxi city. Local Environ. 27 (12), 1479–1498.

[13] Fotheringham, S., Rogerson, P., 2013. Spatial Analysis and GIS. Crc Press.

[14] García-Palomares, J.C., Gutiérrez, J., Cardozo, O.D., 2013. Walking accessibility to public transport: an analysis based on microdata and GIS. Environ. Plann. Plann. Des. 40 (6), 1087–1102.

[15] Gong, J., 2002. Clarifying the standard deviational ellipse. Geogr. Anal. 34 (2), 155–167.

[16] Gunawardena, K.R., Wells, M.J., Kershaw, T., 2017. Utilising green and bluespace to mitigate urban heat island intensity. Sci. Total Environ. 584, 1040–1055.

[17] He, M., Wu, Y., Liu, X., Wu, B., Fu, H., 2023. Constructing a multi-functional small urban green space network for green space equity in urban built-up areas: a case study of Harbin, China. Heliyon 9 (11), e21671.

[18] Hillier, B., 1998. A note on the intuiting of form: three issues in the theory of design. Environ. Plann. Plann. Des. 25 (7), 37–40.

[19] Hillier, B., 2008. Space and spatiality: what the built environment needs from social theory. Build. Res. Inf. 36 (3), 216–230.

[20] Hillier, B., 2009. What do we Need to Add to a Social Network to Get a Society? Answer: Something Like What we Have to add to a Spatial Network to Get a City. Royal Institute of Technology (KTH), Stockholm, Sweden.

[21] Hillier, B., 2012. Studying cities to learn about minds: some possible implications of space syntax for spatial cognition. Environ. Plann. Plann. Des. 39 (1), 12–32.

[22] Hillier, W., Yang, T., Turner, A., 2012. Normalising least angle choice in Depthmap-and how it opens up new perspectives on the global and local analysis of city space. J. Space Syntax 3 (2), 155–193.

[23] Howard, E., 1965. Garden cities of to-morrow, 23. Mit Press.

[24] Jia, Q.H., Mahdzar, S.S.S., Shaharuddin, S.N.B., 2023. Study on the spatial characteristics of urban heritage in Xi'an city's historical core area. In: IOP Conference Series: Earth and Environmental Science.

[25] Jia, Q., Mahdzar, S.S.S., Shaharuddin, S.N., 2025. Urban development and cultural integration: spatial analysis of Hui ethnic group area in Xi'an. J. Asian Architect. Build Eng. 1–21.

[26] Jim, C.Y., Chen, W.Y., 2006. Recreation–amenity use and contingent valuation of urban greenspaces in Guangzhou, China. Landsc. Urban Plann. 75 (1–2), 81–96.

[27] Khahro, S.H., Talpur, M.A., Bhellar, M.G., Das, G., Shaikh, H., Sultan, B., 2023. GIS-based sustainable accessibility mapping of urban parks: evidence from the second largest settlement of Sindh, Pakistan. Sustainability 15 (7), 6228.

[28] Khatibi, M., Khaidzir, K.A.M., Syed Mahdzar, S.S., Sharifi, A., 2024. Measuring spatial configurations for spatial neighborhood sustainability assessment. Environ. Dev. Sustain. 26 (8), 19177–19194.

[29] Khatibi, M., Khaidzir, K.A.M., Syed Mahdzar, S.S., Sharifi, A., 2025. Revisiting the neighborhood definition in view of the 15-minute neighborhood and sustainable neighborhood concepts. Cities 162, 105986.

[30] Konijnendijk, C.C., Annerstedt, M., Nielsen, A.B., Maruthaveeran, S., 2013. Benefits of Urban Parks. A Ssystematic Review. A Report for IFPRA, Copenhagen & Alnarp, pp. 1–70.

[31] Kun, W., Hao, S., Yannan, X., Mingrui, X., Quan, Z., 2012. Accessibility analysis of urban parks based on GIS. In: 2012 Fifth International Conference on Information and Computing Science.

[32] Li, F., Liu, X., Zhang, X., Zhao, D., Liu, H., Zhou, C., Wang, R., 2017. Urban ecological infrastructure: an integrated network for ecosystem services and sustainable urban systems. J. Clean. Prod. 163, S12–S18.

[33] Liao, Z., Zhang, L., Wang, X., 2022. Spatial distribution characteristics and accessibility analysis of beautiful leisure villages in China. PLoS One 17 (10), e0276175.

[34] Long, Y., Qin, J., Wu, Y., Wang, K., 2023. Analysis of urban park accessibility based on space syntax: take the urban area of Changsha city as an example. Land 12 (5), 1061.

[35] Mahdzar, S.S.B.S., 2008. Sociability vs accessibility urban street life. Doctoral thesis. University of London, London.

[36] Mangmang, W., Yuncai, W., 2021. Measurement and optimization of supply and demand of park green space at urban block scale: a case study of Yangpu District, Shanghai. Lands. Architect. 28 (2), 22–27.

[37] Mavoa, S., Witten, K., McCreanor, T., O'sullivan, D., 2012. GIS based destination accessibility via public transit and walking in Auckland, New Zealand. J. Transport Geogr. 20 (1), 15–22.

[38] Niu, Q., Wang, Y., Xia, Y., Wu, H., Tang, X., 2018. Detailed assessment of the spatial distribution of urban parks according to day and travel mode based on web mapping API: a case study of main parks in Wuhan. Int. J. Environ. Res. Publ. Health 15 (8), 1725.

[39] Olmsted Jr, F.L., Kimball, T., 1822. Frederick law olmsted. Landsc. Architect. 1903 (2).

[40] Özgüner, H., 2011. Cultural differences in attitudes towards urban parks and green spaces. Landsc. Res. 36 (5), 599–620.

[41] Park, S.J., 2012. Measuring public library accessibility: a case study using GIS. Libr. Inf. Sci. Res. 34 (1), 13–21.

[42] Peters, K., 2010. Being together in urban parks: connecting public space, leisure, and diversity. Leisure Sci. 32 (5), 418–433.

[43] Peters, M.A., 2021. The ancient Silk Road and the birth of merchant capitalism. Educ. Philos. Theor. 53 (10), 955–961.

[44] Qi, J., Mazumdar, S., Vasconcelos, A.C., 2024. Understanding the relationship between urban public space and social cohesion: a systematic review. Int. J. Commun. Well-Being 7 (2), 155–212.

[45] Sanesi, G., Chiarello, F., 2006. Residents and urban green spaces: the case of Bari. Urban For. Urban Green. 4 (3–4), 125–134.

[46] Seamon, D., 2015. Understanding place holistically: cities, synergistic relationality, and space syntax. J. Space Syntax 6 (1), 19–33.

[47] Talen, E., 1998. Visualizing fairness: equity maps for planners. J. Am. Plann. Assoc. 64 (1), 22–38.

[48] Talen, E., Anselin, L., 1998. Assessing spatial equity: an evaluation of measures of accessibility to public playgrounds. Environ. Plann. A 30 (4), 595–613.

[49] Thompson, C.W., 2002. Urban open space in the 21st century. Landsc. Urban Plann. 60 (2), 59–72.

[50] Tzoulas, K., Korpela, K., Venn, S., Yli-Pelkonen, V., Kaźmierczak, A., Niemela, J., James, P., 2007. Promoting ecosystem and human health in urban areas using Green Infrastructure: a literature review. Landsc. Urban Plann. 81 (3), 167–178.

[51] van den Berg, A.E., Maas, J., Verheij, R.A., Groenewegen, P.P., 2010. Green space as a buffer between stressful life events and health. Soc. Sci. Med. 70 (8), 1203–1210.

[52] Verdú-Vázquez, A., Fernández-Pablos, E., Lozano-Diez, R.V., López-Zaldívar, Ó., 2021. Green space networks as natural infrastructures in PERI-URBAN areas. Urban Ecosyst. 24, 187–204.

[53] Waitt, G., Knobel, H., 2018. Embodied geographies of liveability and urban parks. Urban Stud. 55 (14), 3151–3167.

[54] Wang, S., Wang, M., Liu, Y., 2021. Access to urban parks: comparing spatial accessibility measures using three GIS-based approaches. Comput. Environ. Urban Syst. 90, 101713.

[55] Wolch, J.R., Byrne, J., Newell, J.P., 2014. Urban green space, public health, and environmental justice: the challenge of making cities “just green enough”. Landsc. Urban Plann. 125, 234–244.

[56] Wolf, K.L., Lam, S.T., McKeen, J.K., Richardson, G.R.A., van den Bosch, M., Bardekjian, A.C., 2020. Urban trees and human health: a scoping review. Int. J. Environ. Res. Publ. Health 17 (12).

[57] Xing, Z., Guo, W., 2022. A new urban space analysis method based on space syntax and geographic information system using multisource data. ISPRS Int. J. GeoInf. 11 (5), 297.

[58] Yan, H., Wu, F., Dong, L., 2018. Influence of a large urban park on the local urban thermal environment. Sci. Total Environ. 622, 882–891.

[59] Yao, X., Yu, K., Zeng, X., Lin, Y., Ye, B., Shen, X., Liu, J., 2022. How can urban parks be planned to mitigate urban heat island effect in “Furnace cities”? An accumulation perspective. J. Clean. Prod. 330, 129852.

[60] Yigitcanlar, T., Teriman, S., 2015. Rethinking sustainable urban development: towards an integrated planning and development process. Int. J. Environ. Sci. Technol. 12, 341–352.

[61] Zhang, R., Peng, S., Sun, F., Deng, L., Che, Y., 2022. Assessing the social equity of urban parks: an improved index integrating multiple quality dimensions and service accessibility. Cities 129, 103839.

RIGHTS & PERMISSIONS

2095-2635/2026 Publishing services by Elsevier B.V. on behalf of KeAi Communications Co. Ltd.