Construction of Green Volume Quantity and Equity Indicators for Urban Areas at Both Regional and Neighborhood Scales: A Case Study of Major Cities in China ^†

Abstract

Current urban green volume quantity and equity evaluations primarily rely on two-dimensional (2D) indicators that capture the planar distribution characteristics but overlook vertical structure variations. This study constructed a three-dimensional (3D) evaluation system for green volume quantity and equity by introducing Lorenz curves and Gini coefficients. Using multi-source data, including a 10 m global vegetation canopy height dataset, land cover, and population distribution data, an automated calculation workflow was established in ArcGIS Model Builder. Focusing on regional and neighborhood scales, this study calculates and analyzes two-dimensional green volume (2DGV) and three-dimensional green volume (3DGV) indicators, along with the spatial equity for 413 Chinese cities and residential and commercial areas of Wuhan, Suzhou, and Bazhong. Meanwhile, a green volume quantity and equity type classification method was established. The results indicated that 3DGV exhibits regional variations, while Low 2DGV–Low 3DGV cities have the highest proportion. Green volume in built-up areas showed a balanced distribution, while park green spaces exhibited 2DGV Equitable Only. At the neighborhood scale, residential areas demonstrated higher green volume equity than commercial areas, but most neighborhood areas’ indicators showed low and imbalanced distribution. The proposed 2DGV and 3DGV evaluation method could provide a reference framework for optimizing urban space.

1. Introduction

Urban green space evaluation indicators, including green volume, ecosystem quality, and structure, serve as critical tools for measuring the level of urban greening and assessing the development of urban green infrastructure [1,2,3]. Green space assessment requires not only calculating green space area but also quantifying vegetation’s ecological performance indicators, including vegetation biomass and ecosystem services. This forms the foundation for establishing green space configuration standards, optimizing landscape performances, and achieving climate-resilient urban planning. Multiple international policies also underscore the importance of urban vegetation structure and the development of green infrastructure. For instance, the United Nations’ Sustainable Development Goals (SDG 11 and SDG 15) call for enhancing access to high-quality green spaces in cities and protecting terrestrial ecosystems [4]. The World Health Organization’s Guidelines on Urban Green Spaces and Health indicate that the ecological and health benefits of green spaces depend not only on their area but also on the hierarchical structure and spatial configuration of vegetation [5]. Both UN-Habitat’s New Urban Agenda and the European Union’s Green Infrastructure Strategy emphasize the construction of multi-layered, structurally robust green networks to enhance urban ecosystem services and climate resilience [6,7]. These international policies collectively demonstrate that urban green space evaluation should transcend two-dimensional boundaries to enable more comprehensive quantification of vegetation’s vertical structure.

Among different evaluation indicators, green volume indicators represent the volume quantities of green space and serve as a fundamental green space indicator from macro to micro scales. The vertical heterogeneity and canopy structure of the urban green spaces directly determine a city’s capacity for carbon sequestration, oxygen release, cooling, humidity enhancement, and dust absorption and retention. However, current green space evaluation frameworks in China primarily rely on two-dimensional green volume (2DGV) indicators (

Table 1. 2DGV indicators and minimum requirements in China’s urban green space planning and evaluation standards.

Green Space Indicator	Scope	Definition	Requirements of the Standards
Urban Green Space Ratio	Built-up area	Proportion of total urban green space area to the city’s built-up area	Overall: ≥30% [8]
			National Garden City: ≥40% (minimum 28% in each district) National Ecological Garden City: ≥40% (minimum 25% in each district) [9]
Urban Greening Coverage	Built-up area	Percentage of built-up area covered by the vertical projection of all vegetation	Overall: ≥35% [8]
			National Garden City: ≥43%, with shrubs and trees accounting for ≥70%; National Ecological Garden City: ≥41%, with shrubs and trees accounting for ≥60% [9]
Urban Green Space Area Per Capita	Built-up area	Average area per capita of urban green space	≥10 m2/person [10]
Urban Park Green Space Per Capita	Built-up area	Average area per capita of publicly accessible urban park green space	National Garden City: ≥14.8 m2/person (minimum 5.5 m2/person in each district); National Ecological Garden City: ≥12 m2/person(minimum 5.0 m2/person in each district) [9]
			≥8 m2/person [10]
Urban and Rural Green Space Ratio	City proper area	Proportion of total green space area to the city’s proper area	/
Green Space Ratio	Residential district	Proportion of green space area to the area of residential district	≥30% (for new construction) [11]
			≥30% (for new construction in Wuhan) ≥25% (for renovation projects in Wuhan) [12]
			≥37% (for new construction in Suzhou) [13]
Green Space Area Per Capita	Residential district	Average per capita area of residential green space	≥2.0 m2/person (within 15 min living circle) ≥1.0 m2/person (within 10 min living circle) ≥1.0 m2/person (within 5 min living circle) [11]
			≥1.5 m2/person (for new construction in Wuhan) ≥1.05 m2/person (for renovation projects in Wuhan) [12]
Green Space Ratio	Commercial district	Proportion of green space area to the area of commercial district	≥20% (Wuhan) [12]
			≥25% (Suzhou) [13]
			≥20% (Bazhong) [14]
Green Space Area Per Capita	Commercial district	Average per capita area of commercial green space	Requirement varies by cities

). These metrics only reflect green spaces’ planar distribution, failing to express the vertical structure and ecological functional heterogeneity within urban green spaces.

Meanwhile, the equity assessment for urban green space distribution has emerged as a crucial topic in related research, including urban nature accessibility, social equity, and environmental justice. Current studies mainly use the Gini coefficient and the Lorenz curve to assess the equity of 2D green space [15,16]. Lorenz curves illustrate the distribution patterns by plotting cumulative population percentages against cumulative green space percentages, while the Gini coefficient quantifies green space equity by measuring the area between the Lorenz curve and the absolute equality line [17,18]. For instance, Nero et al. employed the Gini coefficient to analyze the spatiotemporal evolution and equity of urban green spaces in Kumasi, Ghana [19]. Kabish et al. utilized the Lorenz curve to quantitatively assess the equity of urban green space distribution, thereby revealing spatial inequalities in Berlin’s green space allocation across diverse population groups [20]. Duan et al. applied the Gini coefficient and Lorenz curve to examine the temporal changes in green space distribution equity across Guangzhou under varying regional development conditions [21]. However, 2DGV quantity and equity assessments can only capture planar distributions and disparities while failing to express the vertical layers of urban green spaces [22,23,24,25] as the ecological benefits and residents’ perception of green space are influenced not only by planar coverage but also by the multi-layered greenery structure and spatial distribution [26]. Moreover, existing research indicates that compared to 2DGV, 3DGV exhibits a stronger correlation with the recreational and ecological service performance of the green space, and the 3DGV metrics could reflect the ecosystem service levels of green spaces more closely [27]. So, it is essential to incorporate 3D green space indicators to reflect the spatial distribution patterns of green spaces more comprehensively.

The emergence of three-dimensional green volume (3DGV) calculation methods provides a new dimension for evaluating the quantity and equity of green spaces. Defined as the spatial volume occupied by plant foliage and stems, 3DGV is also referred to as the 3D vegetation volume [28,29]. By expressing the 3D spatial distribution of green spaces, 3DGV can reflect the green volume quantity varied by canopy heights and spatial distribution in the vertical dimension, which is closely related to multiple ecosystem service performance [30,31]. 3DGV evaluations are primarily based on calculating 2D leaf areas or the 3D spatial volume occupied by plants. The leaf area calculation method treats the total leaf area as a metric for the 3DGV calculation by using indicators such as leaf area index (LAI), leaf area density (LAD), or vertical leaf profile (VLP) [32,33]. The spatial volume method establishes 3D plant geometric models by remote sensing imagery, 3D point clouds, or oblique photography to estimate the vegetation volume [34,35,36,37,38,39]. This approach focuses on morphological geometries while overlooking the structural heterogeneity. It offers straightforward data acquisition, intuitive computational processes, and applicability at various scales.

In the practice of urban green space assessment, 3DGV has been gradually applied to urban microclimate evaluation, thermal environment optimization, quantitative assessment of ecological services, and visual landscape perception [40]. However, the research has predominantly emphasized precise measurements of 3DGV, with studies confined to technical experiments at the local site scale, while there is a lack of attention to its role as an urban green space indicator for examining overall spatial distribution patterns at medium or large scales, inter-city variations, and its relationship with 2DGV indicators [41]. In recent 3DGV research, Zhou et al. utilized 3DGV to estimate vegetation carbon sequestration and conducted further analysis and optimization [28]. Hong et al. employed 3DGV as a supplementary indicator alongside the green visibility ratio to evaluate greenery visibility in street canyons and the psychological perception benefits [42]. Yue Feng et al. used remote sensing to develop predictive models for 3DGV and AOD, proposing corresponding pollution mitigation strategies and demonstrating that 3DGV better reflects differences in vegetation structural hierarchy and ecological benefits compared to 2D metrics [43]. However, most current research and practice in 3DGV evaluation are confined to theoretical frameworks and indicator systems at a single scale, lacking a standardized, multi-scale evaluation protocol. In the context of urban green space equity, field surveys for measuring 3DGV are constrained by limited spatiotemporal ranges and high labor costs, which impede large-scale applications [44]. Recent studies have attempted to address this by utilizing high-resolution satellite imagery and LiDAR scanning [45]. For instance, Zhou et al. analyzed the evolution of 3D green accessibility equity in Nanjing’s urban core area using remote sensing and machine learning [46]. Nevertheless, the dependence on high-quality training data of such methods limits their transferability, making it difficult to establish a universal framework applicable across different locations and scales. Furthermore, there remains a lack of systematic comparisons regarding the spatial distribution characteristics and correlations of 3DGV across multiple scales.

Therefore, to address the limitations of traditional 2D planar evaluation methods and resolve the lack of unified standards for 3DGV assessment, this study constructed a generalized multi-scale evaluation framework for urban green space quantity and equity that integrates both 2DGV and 3DGV indicators. The introduction of 3DGV indicators could facilitate a paradigm shift in the evaluation of green space quantity and equity, from a focus on planar space to 3D space, and reveal the long-overlooked structural imbalances in greenery. This study also established a standardized quantification method based on indicator normalization, which could be adaptable across different land uses and scales. Furthermore, leveraging multi-source spatial data and volumetric analysis methods, an automated calculation workflow for 2DGV and 3DGV indicators was developed by using ArcGIS Model Builder at both regional and neighborhood scales. The accuracy of these results was verified through comparative analysis and field verification. This study analyzes the green volume quantity characteristics and equity disparities across 413 major cities at the regional scale and 3 representative cities at the neighborhood scale in China. Based on the analysis, a categorization method for urban green volume distributions was proposed to provide a quantitative basis for identifying areas with green volume deficits and allocation imbalances. Finally, the typical issues in green volume distribution among the cities were summarized, and planning strategies were proposed to optimize the spatial patterns and balance the allocation of green resources.

2. Materials and Methods

2.1. Research Area

This study evaluated the distribution of urban green volume quantity and equity at both regional and neighborhood scales.

At the regional scale, this study selected 413 cities in mainland China with populations exceeding 150,000 (excluding Hong Kong, Macau, and Taiwan). These cities represent diverse geographical distributions, climate zones, and economic development levels. At the neighborhood scale, three cities—Suzhou, Wuhan, and Bazhong—were selected for the analysis. The three cities are located in eastern, central, and western China, respectively, exhibiting significant regional differences. Among the three cities, Suzhou possesses multiple classical gardens among the green spaces and a highly integrated urban green space network, representing the typical characteristics of mature, high-quality green space layouts in developed eastern regions. Wuhan, located in central China, features high built-up area density, complex functional zones, and interwoven blue-green spaces. Bazhong, characterized by mountainous terrain and a green space pattern dominated by natural vegetation, reflects the spatial structure of green volume constrained by topography in medium-sized western cities. The three cities exhibit distinct differences in urbanization levels, topographical features, and green space structures. Meanwhile, they cover diverse development stages and green space configurations across China, enabling the verification of the universal applicability of the 3DGV indicators at the neighborhood scale [47,48,49,50,51].

Among the land use areas at the neighborhood, residential areas serve as the long-term living environments, while commercial areas function as major employment and social cohesion hubs. These two types represent the most frequently used spaces for urban dwellers. Green spaces in residential and commercial areas are closely related to the shade and cooling effects, air flow conditions, and environmental quality for recreational activities [52,53]. And their three-dimensional greenery structure not only influences urban vitality but also directly relates to environmental quality and public comfort. Therefore, we selected residential and commercial districts in each city to measure and compare the green volume quantity and equity (see Figure 1).

2.2. Data Sources and Data Processing

The data sources utilized for green volume assessment and equity analysis in this study fall into three main categories: (1) boundary data, which are used to define the spatial boundaries of various analysis units, including urban built-up area boundaries, administrative divisions, and functional zoning extents; (2) 3DGV calculation data, which include the canopy height dataset, land cover, and green space type distribution to derive the 2DGV and 3DGV indicators; and (3) population data, which are extracted from the Chinese Census Bureau and relevant online platforms, and are used to calculate per capita metrics to measure the green volume equity (

Table 2. Sources of research data.

Data Type	Dataset Name	Data Format	Data Source
Boundary extent data	China Administrative Boundary Data	vector data (shp)	National Geomatics Center of China (NGCC) www.ngcc.cn/dlxxzy/gjjcdlxxsjk/ (accessed on 10 December 2025)
	China Urban 2020 Built-up Area Dataset	vector data (shp)	Developed by Sun et al. [54]. (CAS) based on urban impervious surface data www.doi.org/10.11922/sciencedb.j00001.00332 (accessed on 10 December 2025)
	NPP/VIIRS Nighttime Light Data (2020, 500 m)	raster data (tiff)	Full-sequence global annual HD nighttime light product from CAS based on Suomi-NPP satellite (2012–2020)
	Gaode Maps Urban Residential and Commercial Area Boundaries AOI	vector data (shp)	Acquired via Gaode Open Platform API lbs.amap.com (accessed on 10 December 2025)
3DGV calculation data	Global Canopy Height Data (10 m)	raster data (tiff)	Global Canopy Height Data developed by Lang et al. (ETH Zurich, Yale University) [55] langnico.github.io/globalcanopyheight/ (accessed on 10 December 2025)
	Global Canopy Height Data (1 m)	raster data (tiff)	High-resolution 1 m global canopy height data developed by Meta and World Resources Institute using AI models [56].
	Land Cover Type Data, (10 m)	raster data (tiff)	World Cover dataset (ESA) viewer.esa-worldcover.org/worldcover/ (accessed on 10 December 2025)
	Park Green Space Extent	vector data (shp)	OpenStreetMap
Population distribution data	City Statistical Yearbooks	numerical data	National Bureau of Statistics of China (NBS) www.stats.gov.cn/sj/ndsj/2020/indexch.htm (accessed on 10 December 2025)
	Anjuke Platform Residential Area Population Data	numerical data	Anjuke Platform www.anjuke.com (accessed on 10 December 2025)
	China’s Seventh National Population Census	numerical data	China’s 7th National Population Census (NBS) www.stats.gov.cn/sj/ndsj/2020/indexch.htm (accessed on 10 December 2025)
	Commercial Building Personnel Density Modeling Mean	numerical data	Mean personnel density modeled for commercial buildings by Feng et al. [57].

).

Given that no unified urban built-up area dataset exists for Chinese cities, this study obtained the built-up area extent for 413 cities by applying the regression analysis and empirical threshold methods to the 2020 NPP/VIIRS night light data. As the variations in urban development and spatial patterns across cities limit the accuracy of built-up areas derived solely from night-time light data, the results were manually cross-checked against the remote sensing imagery and the urban built-up area dataset published by Sun et al. [54]. This correction process was further refined by utilizing the “Eliminate Polygon Part” tool in ArcGIS Pro (v3.3.1). By testing different thresholds, a threshold of region area ≥ 60% of total area was selected to fill voids within contiguous areas and filter out small, non-real discontinuities, which were typically generated by image vectoring errors. This process enhanced the connectivity and integrity of the built-up areas while preserving their overall boundary morphology.

The regression validation of the adjusted results was conducted against the 2020 urban built-up area dataset from the Statistical Yearbook, yielding a goodness-of-fit value (${\mathrm{R}}^2=0.8167$). The discrepancy between the two-dimensional green coverage data for some cities and the Statistical Yearbook is attributable to the differences in the delineation methods for built-up areas by various local municipalities. The Statistical Yearbook’s definition of built-up areas encompasses a broad scope, including contiguously developed zones within urban administrative boundaries that possess complete municipal infrastructure [58], which further categorizes these areas into types such as core districts—multi-cluster zones. Consequently, the dataset includes central urban districts, peripheral new towns, satellite towns, and scattered contiguous development zones [59]. Taking Chongqing as an example, the statistical yearbook’s built-up area covers all developed zones across 26 districts, 8 counties, and 4 autonomous counties. In contrast, the research scope of this study confined the data extraction to contiguous urban built-up areas within the central urban districts, thus including only nine main urban districts [60]. For cities like Shanghai, Foshan, and Beijing, rapid urban boundary expansion over the past decade has led to high integration between central urban districts and peripheral areas [61,62]. The blurred rural and urban boundaries results in spatial discrepancies between the built-up area ranges in statistical yearbooks and those extracted via remote sensing. Based on residual analysis showing no structural patterns and the validity of linear relationships, this study retains these differences rather than excluding them, thereby reflecting the diversity of urban structures across Chinese cities.

By calling the Baidu Map APIs, the AOI datasets for corresponding cities were downloaded to extract the boundaries of each residential and commercial district.

3DGV calculations require data on green space extent and corresponding canopy height data of the different study areas. The canopy height data was extracted from the global canopy height data (vertical resolution 1 m) produced by Lang et al. using GEDI and Sentinel-2 remote sensing imagery to ensure comparability of vegetation height extraction across different cities [55]. The land cover data originated from the WorldCover 10 m land cover dataset released by the European Space Agency. This dataset demonstrates global consistency and high classification accuracy, meeting the requirements for determining green space boundaries across cities [63]. For the green space boundaries in the built-up areas, the canopy height model was clipped by the built-up area boundary for each city. Areas with non-zero values were extracted as valid areas for 3DGV calculation, thereby avoiding inclusion of image noise or highly reflective non-green surfaces. The boundaries of urban parks for each city were calculated from the area intersection of urban park vector boundaries from OpenStreetMap and built-up area boundaries, ensuring that only relevant green spaces within built-up areas were evaluated. For suburban green space extents, land cover data was clipped based on the difference between each city’s municipal coverage and built-up area boundary. The coverages of trees, shrubs, grasslands, and meadow wetlands located within the municipal boundary but outside the built-up area are extracted to define the extent of regional green space [64]. For green spaces within residential and commercial districts across cities, corresponding canopy height areas were extracted based on the areas of interest (AOI) dataset, with non-zero values selected as the valid areas.

The population data for each city was obtained from the 2020 Statistical Yearbook published by the Chinese Census Bureau. The population within each commercial district was calculated by using the average personnel density model for commercial buildings (0.13 people/m²) developed by Feng et al. [57]. This model, constructed based on field surveys, reflects typical population density levels in commercial spaces. The population estimation was obtained through a two-step calculation process: (1) multiplying the single-floor area of each commercial building by its total number of floors to determine the total floor area; and (2) multiplying this total area by the average occupancy density. The population within the residential areas was calculated using the number of households data from the Anjuke platform and the average population per household (2.47 persons) released in China’s Seventh National Population Census Report [65].

2.3. Analytical Methods

2.3.1. 3DGV Quantity Assessment Indicators

Referring to the 2DGV indicators, this study developed a 3DGV quantity and equity evaluation system at both regional and neighborhood levels [66,67]. Automated workflows for calculating 2DGV and 3DGV indicators were also created to support comparative analysis using ArcGIS Model Builder.

The regional-scale green volume quantity indicators incorporated four key indicators of each city: (1) 3DGV per unit area (built-up area), representing the city’s average green volume level and reflecting the spatial density and vertical scale of urban greening. It serves as the 3D equivalent of the urban green space ratio or urban greening coverage rate in 2D. (2) 3DGV per capita (built-up area), demonstrating residents’ allocation of green resources. It corresponds to the urban green space area per capita in 2D. (3) 3DGV per capita of urban parks (built-up area), specifically reflecting green volume supply disparities of the urban park of each city. It is the 3D equivalent of green space area per capita of the urban park. (4) 3DGV per unit area (city proper area), reflecting the average 3DGV quantity within the city proper limits, including the development and non-development land. It corresponds to the urban and rural green space ratio in 2D [68].

At the neighborhood scale, the green volume assessment employed two types of indicators: 3DGV per unit area and 3DGV per capita [69]. The first revealed spatial differences in greening intensity within residential and commercial areas, while the second showed disparities in 3DGV equity among residents in different areas (see Figure 2).

Due to the data availability for multi-scale analysis, the spatial volume occupied by plants was employed as the 3DGV representation in this study. This method has been widely used for 3DGV estimation based on LiDAR data, demonstrating high efficiency and accuracy in large-scale calculations [70,71]. As the canopy height values from the dataset are integers, trees and shrubs with 1 m or above in height were selected as valid areas for calculation. This threshold setting excludes ground vegetation coverage in unstable growth conditions, such as grasslands, thereby enhancing the accuracy of 3DGV calculations [72,73] (see Figure 3). To further validate the threshold parameters, this study employed a sensitivity analysis with progressively increasing thresholds. Since the canopy height model dataset is in integer format, thresholds were sequentially set at 0 m, 1 m and 2 m. Changes in spatial coverage and accumulated 3DGV across different height intervals were then compared.

When the threshold increased from 0 m to 1 m, the 3DGV data coverage area decreased by 22.12%. This characteristic of high area proportion and no height contribution indicates that the 0–1 m range is primarily composed of extensive ground-level herbaceous cover and LiDAR near-ground detection noise. When the threshold increased from 1 m to 2 m, the average 3DGV value decreased by 7.12%, and the data coverage area decreased by 8.12%. From 2 m to 3 m, the average 3DGV value decreased by 7.92%, and the data coverage area decreased by 8.42%. Unlike the widespread distribution observed in the 0–1 m range, data loss in this interval was primarily concentrated within specific spatial clusters, corresponding to shrub layers and low-growing trees within urban green spaces. Setting the threshold to 2 m would result in the loss of low-level vegetation information.

The vegetation volume within each pixel was approximated as a cuboid geometry: the unit pixel size defined the X and Y axes, and the canopy height data within the pixel served as the Z-axis dimension. The total 3DGV was then calculated by multiplying the crown height by the occupied area and cumulating the results using the following formula:

$$S_A=x·y$$

(1)

$$GV=S_A·H$$

(2)

$$V_{TGV}=\sum _{i=1}^n{GV}_i$$

(3)

where $S_A$ is the area of a unit pixel, H represents the average canopy height data within the unit pixel, ${GV}_{i }$ is the 3DGV within the unit pixel, $V_{TGV}$ is the total 3DGV, and $n$ is the number of unit pixels. Subsequently, the calculation formulas for the four 3DGV quantity evaluation indicators were derived (

Table 3. Calculation methods for 3DGV evaluation indicators.

Indicator	Scope	Formula	Description
3DGV per capita (m3/pp)	Built-up area *	${\lambda }_{Gm}={\displaystyle \frac{V_{TGV}}{N_p}}$	$V_{TGV}$ is the total 3DGV in built-up area $N_p$ is the total regional population
	Residential area	${\lambda }_{Rm}={\displaystyle \frac{V_{Gr}}{N_{rp}}}$	$V_{Gr}$ is the total 3DGV of residential area $N_{rp}$ is the population of residential area
	Commercial area	${\lambda }_C={\displaystyle \frac{V_{Gc}}{N_{cr}}}$	$V_{Gc}$ is the total 3DGV of commercial area $N_{cr}$ is the population of commercial area
3DGV per capita of urban parks (m3/pp)	Built-up area	${\lambda }_{G1}={\displaystyle \frac{V_{G1}}{N_P}}$	$V_{G1}$is the total 3DGV of the region urban parks
3DGV per unit area (m3/m2)	Built-up area	${\lambda }_u={\displaystyle \frac{V_{TGV}}{V_c}}$	$V_{c }$ is the total regional land area
	City proper area **	${\lambda }_L={\displaystyle \frac{V_{TGV}+V_{EG}}{A_c}}$	$V_{TGV}$ is the total 3DGV in built-up area $V_{EG}$ is the total 3DGV of suburban green space $A_c$ is the city proper area
	Residential area	${\lambda }_{Rm}={\displaystyle \frac{V_{Gr}}{A_{rp}}}$	$A_{rp}$ is the residential area
	Commercial area	${\lambda }_{Cm}={\displaystyle \frac{V_{Gc}}{A_{cp}}}$	$A_{cp}$ is the commercial area

* Built-up area is the existing development land of the city. ** City proper area is the area contained within city limits, including the development land and non-development land.

).

Wuhan City was selected as the verification site to validate the reliability of the 3DGV estimation method and the accuracy of canopy height data. Its highly heterogeneous urban built environment and rich vertical vegetation structure could be used to test the model algorithm’s performance in complex scenarios. A field-verification approach was employed using five representative sample areas. Data collection involved multi-angle ground photographs from field surveys and oblique aerial imagery from UAVs to obtain vegetation height, average canopy width, and other metrics. The sampling areas encompassed representative urban parks such as Beitaizi Lake Park, Dai Jia Lake Park, Hankou Riverside Park, Huanglongshan Park, and the green spaces of Huazhong Agricultural University, ensuring diversity in vegetation types and spatial distribution. After screening, 1412 valid high-resolution images of the 5 sample sites were obtained. Three-dimensional model reconstruction was performed using the open-source photogrammetry platform OpenDroneMap (v3.5.6), which automatically generated high-density point clouds and orthoimages for the sample sites.

A regression analysis was conducted to compare crown height data derived from field measurements with remote sensing imagery. The validation results demonstrated a high correlation between the two datasets (${\mathrm{R}}^2 = 0$.933). The spatial distribution of 3DGV generated by the model aligns with field observations of canopy density and structural layers. The model-derived crown heights showed a close match with field-measured heights, indicating that the crown-height-based volume estimation model achieves high accuracy at the plot scale (Figure 4).

2.3.2. Green Volume Equity Evaluation Method

The Gini coefficient and the Lorenz curve were introduced to quantify the equity of 2DGV and 3DGV across different areas. The curvature of the Lorenz curve reflects the cumulative distribution of green volume against the population, which exhibits a negative correlation: a greater curvature indicates a lower level of equity [74]. The Gini coefficient is derived from the Lorenz curve. It quantifies the deviation from the ideal equitable distribution by calculating the area ratio between the Lorenz curve and the Line of Perfect Equality to the total area beneath the equality line [75]. The formula is as follows:

$$G=1-\sum _{i=1}^n\left(P_i-P_{i-1}\right)\left(S_i+S_{i-1}\right)$$

(4)

where $G$ represents the Gini coefficient, $P_i$ is the cumulative population, and $S_i$ is the cumulative proportion of the green volume indicator. A higher value indicates more inequity in green volume distribution within the study area. Population size and 3DGV data for study units were calculated by using the Zonal Statistics tool in ArcGIS Pro. Then, a batch processing Python (v3.1.3) script was used to calculate the coefficients and plot the curves.

2.3.3. Classification of Green Volume Quantity and Equity Types

To facilitate the comparative analysis of 2DGV and 3DGV indicators in regional and neighborhood scales, this study proposed a classification method for urban green volume types based on the calculated indicators. The evaluation at the regional scale aimed to compare green volume capacity within the built-up areas among different cities. Before integrating 2DGV and 3DGV indicators, it is essential to ensure comparability across various scales, administrative boundaries, and spatiotemporal granularities. Therefore, this study employed a unified normalization method to process urban-scale and community-scale data by selecting green volume per unit area, green volume per capita, and green volume per capita of urban parks to reflect the spatial distribution of green volume within urban built-up areas. The green volume per unit city proper area indicator reflects the overall green volume situation within the city’s administrative boundary; so, it was not selected in this study’s analysis [76]. At the neighborhood scale, green volume per unit area and green volume per capita were used as the indicators for classification. All the selected green volume indicators underwent Z-score normalization to ensure the comparability and consistency across different scales during the calculation:

$$\mathrm{Z}\mathrm{i}={\displaystyle \frac{Xi-\overline{X}}{s_X}}$$

(5)

where $X_i$ is the original indicator value for a city or a site, and $\overline{X}$ and $s_X$, respectively, represent the mean and standard deviation of that indicator in the sample.

For weight determination, this study employed a variance contribution rate weighting method to calculate the 2DGV performance index (P₂D) and 3DGV performance index (P₃D) due to limited indicator types and their independence among factors. This method assigns weights based on a variable’s proportion of total variance, reflecting its contribution to the variability in and differentiation of composite indicators. And it is widely used in principal component analysis (PCA), indicator synthesis, environmental and ecological assessments, and other research areas to objectively assign weights within comprehensive evaluation systems. Weights are directly determined by variance contribution rates. An indicator will be assigned a higher weight if its variance is larger, thereby indicating a greater difference between samples and a stronger influence on the overall result. This avoids the bias introduced by subjective weighting assignment, making the overall result more consistent with the inherent structural characteristics of the data [77,78]. The calculation formula is as follows:

$$a_i={\displaystyle \frac{s_i^2}{{\sum }_{j=1}^n{s}_j^2}}$$

(6)

$$\mathrm{P}=\sum _{i=1}^n{a}_i{Z}_i$$

(7)

where $a_i$ is the weight of the $\mathrm{i}$-th indicator, and $s_i^2$ is the sample variance in that indicator, $n$ is the total number of indicators. P represents the comprehensive index, and $Z_i$ represents the standardized value of the indicator.

Using the average of the comprehensive 2DGV and 3DGV indicators as the threshold, cities can be subsequently classified into four green volume quantity types: High 2DGV–High 3DGV, Low 2DGV–Low 3DGV, High 2DGV–Low 3DGV, and Low 2DGV–High 3DGV.

Since the equity indicators are continuously distributed, this study categorized the calculated Gini coefficients for green volume based on the international Gini coefficient standard [79]. At the regional scale, Gini coefficients were calculated for green volume in built-up areas and urban parks. And at the neighborhood scale, Gini coefficients were computed for green volume within residential and commercial areas [80,81]. The distribution is considered equitable when $\mathrm{G}<0.4$ and inequitable when $\mathrm{G}\ge 0.4$. Integrating these thresholds with the distribution characteristics of 2DGV and 3DGV quantity, study areas were classified into four equity types: Both Equitable, 3DGV Equitable Only, 2DGV Equitable Only and Both Inequitable, reflecting the disparities in green volume equity across different dimensions (

Table 4. Classification of Green Volume Quantity and Equity Types.

Classification Metric	2D Indicator	3D Indicator	Classification Type
Green Volume Quantity	P2D ≥ Average	P3D ≥ Average	High 2DGV–High 3DGV
	P2D ≥ Average	P3D < Average	High 2DGV–Low 3DGV
	P2D < Average	P3D < Average	Low 2DGV–High 3DGV
	P2D < Average	P3D < Average	Low 2DGV–Low 3DGV
Green Volume Equity	$G<0.4$	$G<0.4$	Both Equitable
	$G<0.4$	$G\ge 0.4$	2DGV Equitable Only
	$G\ge 0.4$	$G<0.4$	3DGV Equitable Only
	$G\ge 0.4$	$G\ge 0.4$	Both Inequitable

).

3. Results

3.1. 3DGV Quantity and Equity Distributions at the Regional Scale

3.1.1. 3DGV Quantity Distributions Among Different Cities

The four 3DGV quantity indicators were calculated for 413 Chinese cities (see Figure 5). For built-up areas, 3DGV per unit area ranged from 0.25 to 7.97 m³, with an average of 2.80 m³. The distribution exhibited a left-skewed normal pattern, indicating that most cities were concentrated at the medium-to-low level, while only a few possessed high 3DGV values. 3DGV per capita ranged from 22.91 to 2203.47 m³, averaging 458.15 m³. The distribution showed pronounced clustering at lower values, suggesting residents generally have limited green spaces. 3DGV per capita of urban parks ranged from 0.75 to 218.82 m³, with an average of 24.93 m³. The distribution exhibited a near-monotonically decreasing pattern, reflecting significant differences between the urban park green spaces across cities. For city proper areas, 3DGV per unit area ranged from 0.36 to 32.53 m³, with an average of 12.89 m³. The bimodal (“M-shaped”) distribution indicated the coexistence of cities with very high and very low green volume levels, highlighting substantial spatial heterogeneity between densely built urban cores and peripheral or less-developed regions.

In terms of geographical distribution, the 3DGV of cities exhibits pronounced spatial disparities. Overall, the city built-up areas in Southwest, Northeast, and East China generally possess higher green volume per unit area than those in Northwest, North, and Central China. City built-up areas in East and Northwest China demonstrate relatively high green volume per capita, whereas those in Central and North China generally exhibit lower values. Regarding 3DGV per capita of urban parks, high-value cities are primarily concentrated in parts of East China, Central China, and South China, while low-value distributions are observed in Southwest China and Central China. High-value areas for 3DGV per unit area (city proper area) are mainly clustered in South and Northeast China, whereas the cities in North China and Northwest China generally exhibit lower 3DGV values.

3.1.2. 2DGV and 3DGV Quantity Disparities Among the Cities

To compare the data distribution between 2DGV and 3DGV indicators across different cities, linear regression analyses were performed on corresponding pairs of indicators (see Figure 6). The weakest correlation was found between the 2DGV and 3DGV per unit area. This suggests that the 2D green space indicators do not align with the proportion of vegetation distribution in 3D space. The correlation coefficient R² between the 2DGV and 3DGV per unit city proper area was higher than the correlation per built-up area. This indicates that using both 2DGV and 3DGV indicators, especially in urban built-up areas, can provide a more comprehensive view of green space characteristics.

Based on 2DGV and 3DGV indicators, urban green volume quantity types are divided into four categories derived from the established classification methods. These categories reflect green volume quantity distribution variations among different cities (see Figure 7). Cities classified as the High 2DGV–High 3DGV type account for 34.38% and are primarily found in regions in good ecological condition, including the Yangtze River Delta, Pearl River Delta, Northeast China, and Southwest China. The Low 2DGV–Low 3DGV type cities account for 49.64% and are distributed across arid or semi-arid areas of Northwest China and North China, as well as construction land-constrained regions in Central and North China. The High 2DGV–Low 3DGV type cities account for 8.23% and mainly locate in East China. Finally, the Low 2DGV–High 3DGV type cities account for 7.75%, largely located in the southwest or along the eastern coastal areas.

3.1.3. 2DGV and 3DGV Equity Disparities Among the Cities

The 2DGV and 3DGV equity distributions for the built-up areas of 413 Chinese cities were then calculated. The overall Gini coefficients for the 2DGV and 3DGV distributions were 0.31 and 0.32, respectively. These values indicated an overall balanced distribution of 2DGV and 3DGV across cities, which can be classified as the Both Equitable type with relatively limited spatial disparities.

The Gini coefficients for 2DGV and 3DGV of urban parks were 0.31 and 0.41, which indicated that the 3DGV distribution was more inequitable than the 2DGV distribution. This finding suggested that despite broad urban park green space coverage, the 3DGV distribution of urban parks remained uneven (see Figure 8).

3.2. 3DGV Quantity and Equity Distributions at the Neighborhood Scale

3.2.1. 3DGV Quantity Distribution in Residential and Commercial Areas

At the neighborhood scale, the 3DGV per unit area and 3DGV per capita indicators were calculated for the residential and commercial areas of the three cities (see Figure 9).

The results for the indicators across the three cities generally clustered at low values, indicating that the overall 3DGV quantity of residential and commercial areas was generally low. 3DGV indicators of residential areas showed higher values than those for commercial areas, suggesting a higher vegetation complexity and density in residential areas. The green spaces in commercial areas were constrained by high-density buildings and extensive impervious surfaces, resulting in an overall insufficient green volume.

Regarding 3DGV per unit area, the average value for residential areas was 0.66 m³/m², with 94.87% falling within the low-value range of 0–2 m³/m². The average value for commercial areas was 0.27 m³/m², with 87.72% falling within the low-value range of 0–0.5 m³/m². Regarding 3DGV per capita, the average value for residential areas was 9.58 m³, primarily distributed between 0 and 20 m³. The value of commercial areas averaged 0.66 m³, concentrated between 0 and 2 m³, with extremely low proportions in high-value zones (residential areas >20 m³ per person and commercial areas >2 m³ per person). Significant disparities also existed among different cities. Bazhong achieved a 3DGV per capita of 13.25 m³ in residential areas and 0.99 m³ in commercial areas. Its green volume per unit area was 0.74 m³/m² and 0.42 m³/m², respectively, exceeding Suzhou and Wuhan. Suzhou only marginally surpassed Wuhan in terms of residential 3DGV per capita.

3.2.2. 2DGV and 3DGV Quantity Disparities in Residential and Commercial Areas

The linear regression of 2DGV and 3DGV indicators in residential and commercial areas across the three cities revealed the correlation differences (see Figure 10). Among them, indicators of Bazhong exhibited the strongest correlation, with R² = 0.88 for commercial areas and R² = 0.92 for residential areas. Indicators of Suzhou followed, with R² values of 0.88 and 0.74 for residential and commercial areas, respectively, and 0.91 and 0.81 for per capita indicators. The correlation in Wuhan was relatively low. Also, among the three cities, the 2DGV and 3DGV in residential areas showed higher consistency than those in commercial areas. In contrast, the correlation between 2DGV and 3DGV was not significant in commercial areas, indicating the importance of adopting the two types of indicators for evaluation.

Based on the 2DGV and 3DGV indicator values, the green volume quantity structures were categorized as four types (see Figure 11). Overall, the proportion of residential and commercial areas classified as the Low 2DGV–Low 3DGV type was the highest, indicating that most of the areas are still experiencing a shortage of green space. The proportion of the High 2DGV–High 3DGV type was the second highest, concentrated in districts with abundant green volume. Meanwhile, the proportion of the High 2DGV–Low 3DGV type was higher than that of the Low 2DGV–High 3DGV type, indicating that current urban green space development tends to focus more on planar greening, which might be influenced by current greening requirements, which mainly concentrate on 2D indicators.

Among the three cities, residential areas in Wuhan and Suzhou had proportions of the High 2DGV–High 3DGV type at 31% and 36%, respectively, which were higher than the corresponding commercial area proportions in these cities. As there is a lack of local green space standards for commercial areas, many commercial areas are dominated by hardened surfaces, which limits the potential for 3D vegetation planting. In Bazhong city, the proportion of High 2DGV–High 3DGV types in commercial districts was 36%, slightly exceeding the proportion of residential areas. This disparity may result from Bazhong’s lower urban density and mountainous terrain, which allow commercial development to integrate more extensively with natural green spaces. The High 2DGV–Low 3DGV pattern outranked the Low 2DGV–High 3DGV pattern in Wuhan and Suzhou, indicating insufficient multi-layered greening spaces in these cities. Conversely, Bazhong exhibited a slightly higher proportion of Low 2DGV–High 3DGV green spaces than other cities, reflecting its abundant multi-layered vegetation space derived from distinct geographical characteristics.

3.2.3. 2DGV and 3DGV Equity Disparities in Residential and Commercial Areas

This study calculated Gini coefficients and Lorenz curves for 2DGV and 3DGV of different functional areas in the three cities (see Figure 12).

This study revealed that among the three evaluated cities, only Suzhou’s residential areas were categorized as the 2DGV Equitable Only type. In contrast, all other residential and commercial areas were classified as the Both Inequitable type. Furthermore, the imbalance in the distribution of 2DGV and 3DGV was more pronounced in commercial areas than in residential areas. In Wuhan, the spatial distribution of 2DGV was significantly more uneven than that of 3DGV, which indicated a concentration of green space resources in certain areas. In Bazhong, the 2DGV distribution in residential and commercial areas was more equitable than the 3DGV distribution, indicating that the concentration of 3DGV was higher than that of 2DGV (

Table 5. Green volume equity type classification for residential and commercial areas.

Area	City	2DGV Gini Coefficient	3DGV Gini Coefficient	Green Volume Equity Type
Commercial Areas	Wuhan	0.76	0.67	Both Inequitable
Residential Areas	Wuhan	0.59	0.50	Both Inequitable
Commercial Areas	Suzhou	0.64	0.71	Both Inequitable
Residential Areas	Suzhou	0.39	0.47	2DGV Equitable Only
Commercial Areas	Bazhong	0.68	0.70	Both Inequitable
Residential Areas	Bazhong	0.49	0.58	Both Inequitable

).

4. Discussion

This study developed a 3DGV quantity and equity evaluation framework to comprehensively capture disparities in urban green space across regional and neighborhood scales in major Chinese cities. Compared with traditional 3DGV metrics, the proposed framework offers several innovations in its indicator system, spatial resolution, and analytical depth. First, by integrating 2DGV and 3DGV indicators, it simultaneously evaluates horizontal and vertical green coverage, enabling multi-level and multi-scale assessments of urban green space equity. Second, through an environmental equity lens combined with socioeconomic factors, this study systematically reveals the spatial heterogeneity of urban green volume and identifies key areas with quality imbalances and spatial mismatches. Third, the methodology integrates automated data processing with field validation, enhancing the accuracy and operational feasibility of the analysis and providing planners with measurable, comparable 3D green space performance indicators. Overall, this framework not only addresses the limitations of 2D metrics in capturing vertical and multi-layered urban green structures but also offers a practical basis for designing differentiated green space optimization strategies and policies. The following discussion focuses on spatial distribution patterns, primary influencing factors, and actionable classification strategies.

4.1. Potential Influential Factors of 3DGV Spatial Distribution

The 3DGV indicators showed that the distribution of urban green volume was strongly influenced by a city’s greening policies, natural conditions, economic development level, and historical layout.

At the regional scale, the 3DGV distribution showed a significant geographical gradient, with higher values in southern and coastal regions than in northern and inland areas. This demonstrated the foundational constraints imposed by topographic conditions on vegetation 3D structure while also reflecting the reshaping effects of socio-economic factors and urban development phases [82]. Correlation analyses were conducted between the total 3DGV in built-up areas and indicators of economic development (GDP per capita), topographic features (elevation data), construction intensity (built-up area building density), and climatic conditions (annual precipitation data) (

Table 6. Correlation between influential factors and the total 3DGV in built-up areas.

Influential Factor	Data Source	Indicator	Correlation Direction	Correlation Coefficient (r)
Level of Economic Development	China Statistical Yearbooks www.stats.gov.cn/sj/ndsj/2020/indexch.htm (accessed on 10 December 2025)	GDP per Capita	Positive	+0.68
Topographic Feature	China Elevation Data (SRTM90m) [83]	Elevation	Positive	+0.47
Construction Intensity	Third National Land Survey	Percentage of Built-up Land	Negative	−0.56
Climatic Condition	Building Volume Panel of 106 Chinese Cities 2018–2023 [84]	Regional Building Density	Negative	−0.61
Influencing Factor	China Statistical Yearbooks www.stats.gov.cn/sj/ndsj/2020/indexch.htm (accessed on 10 December 2025)	Mean Annual Precipitation	Positive	+0.34

).

The results revealed a moderate positive correlation between 3DGV and urban GDP per capita (r = 0.68). This finding aligned with the “Luxury Effect” hypothesis in urban ecology, suggesting that wealthier cities have a greater capacity to adopt complex plant communities rather than lawns [85]. Some cities in Eastern China and Northwestern China have implemented multi-layered greening and ecological corridors during urban renewal, resulting in higher 3DGV levels, indicating the conversion of economic capital into ecological capital [86]. In contrast, cities in Central and Northern China showed notably lower 3DGV due to the expansion of built-up areas and the shrinkage of green spaces, reflecting the green space quantity compliance orientation driven by land finance during the rapid urbanization process, which often neglects high-ecological-value multi-layered greening to reduce construction and maintenance costs [87]. Built-up area density significantly suppresses 3DGV (r = −0.56), with public green space availability often lagging population growth, and the potential to improve the 3DGV in commercial areas is highly constrained. This quantitative study confirmed the core principle of the “Compact City” theory: while high-density development improves land-use efficiency, it also compresses ecological green space [88,89]. However, some mountainous cities in Southwestern and Northeastern China retain substantial vegetation patches due to high native vegetation coverage and terrain constraints that necessitate terrain-adaptive land development with reduced building density [90]. This passive green space preservation offers insights for ecological planning in high-density cities, demonstrating that using non-construction areas constrained by terrain as natural areas can help improve the quantity of 3DGV [91].

At the neighborhood scale, residential areas exhibited a more complex 3D green space structure than commercial areas, while the total 3DGV value negatively correlated with building density (r = −0.61). Among the study sites, residential areas generally featured more complex, multi-layered greenery and higher concentrations of green volume [92]. In contrast, commercial areas showed lower potential for multi-layered greening due to high building density and extensive hard surfaces [93,94]. The analysis of the three sample cities showed distinct characteristics: Suzhou, as a high economically developed city on the plain, exhibited high overall 3DGV levels with balanced green space distribution in residential and commercial areas, attributable to the systematic planning of ecological corridors and urban park networks in its early planning stages; Bazhong, with its rugged terrain, low building density, and abundant natural vegetation, showed the highest 3DGV per capita in residential and commercial areas, highlighting the positive role of topographic conditions on 3D greening; Wuhan, as a typical high-density city, demonstrated relatively low overall green volume levels in residential and commercial areas, with its green volume concentrated along riverfronts and large public green space patches [95,96].

4.2. Differences in Green Volume Equity at Regional and Neighborhood Scales

At the regional scale, the built-up area green space equity evaluation revealed that although the overall green volume distribution across cities tended toward equilibrium, the 3DGV of urban park green space equity of the different cities was notably low. Compared with the park accessibility evaluation method based on travel time, this study indicated that even when residents have equal park accessibility, lawns and other simple vegetation could lead to unequal supply of ecological services (e.g., limited shading areas) among residents, which may result in a more profound quality-based environmental injustice [97,98].

At the neighborhood scale, both residential and commercial districts showed an uneven distribution of green volume, with a widespread tendency toward spatial concentration within specific areas. The higher 3D equity observed in residential areas reflected a planning and design priority on multi-layer greening and green space accessibility. Conversely, the low equity in commercial districts exemplified an efficiency-concentrated planning approach. This disparity suggested that the functional segregation in urban planning may exacerbate the stratification of green space accessibility. If high-quality 3DGV primarily concentrates in enclosed, high-end residential areas, while commercial and public spaces suffer from a scarcity of multi-layered greening, this will lead to the degradation of ecological quality in the public domain, a trend often termed ecological privatization [82]. Furthermore, the results indicated that 2DGV inequity was primarily driven by development intensity, whereas 3DGV inequity was more influenced by land topography and building morphology. Planar greening is sensitive to land-use intensity and building density, whereas the constraint on 3DGV is partly due to topographical undulations and a higher proportion of low-rise buildings [99,100].

4.3. Optimizations for Different Green Volume Quantity and Equity Types

Based on differences in the distribution of quantity and equity indicators, this study proposed a classification method for urban green volume types to guide the optimization of urban green spaces.

4.3.1. Regulation Implementing Strategies for Different Green Volume Quantity and Equity Types

Among the four green volume quantity types, areas classified as High 2DGV-High 3DGV often possess high ecological values and low development pressure. They may focus on integrating the existing multi-layered vegetation regulations into urban planning, such as adopting the “Biotope Area Factor (BAF)” to set restrictions on soil layer thickness and hard surface coverage, ensuring the space for multi-layered vegetation [101]. High 2DGV-Low 3DGV areas are mainly characterized by rapid population growth, land development, and urbanization. These areas may experience insufficient ecological benefits with limited multi-layered greening, as the ecosystem service benefits of tree canopies, such as cooling and carbon sequestration, significantly surpass those of lawns [102]. Therefore, the green space regulations will need to focus on increasing tree planting and multi-layered greening. Mechanisms similar to the BAF or green plot ratio (GPR) could be introduced to establish a green volume coefficient based on the local context [103]. For Low 2DGV-High 3DGV areas with high building density or complex terrain, differentiated greening criteria could be established based on corresponding land use patterns [104]. This includes ensuring the minimum street greening width in high-density residential areas and the openness of public green spaces. In areas with complex terrain, using slope development control and greening management can help fortify the 3DGV advantage and enhance public accessibility. For Low 2DGV-Low 3DGV cities, multiple economic and environmental factors should be integrated and analyzed to set quantifiable targets for designating priority greening zones and expanding greening types (e.g., vertical and rooftop greening) [105].

The different green volume equity types could also serve as a reference for zoning regulations. For 2DGV Equitable Only areas, mandatory requirements for multi-layered greening (e.g., tree planting, vertical green walls, and green roofs) should be added to residential and commercial area zoning codes [106]. For 3DGV Equitable Only areas, improving green space accessibility may require adjusting land-use layouts and implementing green-space-prioritized retrofit for development projects [107]. For Both Inequitable areas, minimum green volume standards will need to be set based on the distribution of socially vulnerable areas (e.g., high-density residential zones or low-income communities), and compensatory planning could be implemented based on the population density distribution [97].

4.3.2. Optimization Strategies for Typical Green Space Development Patterns

By associating green space quantity and equity types across different cities, three typical green space development patterns were then identified: low-quality development, spatially constrained development, and resource misallocated development. The three patterns expose the core challenges that urban green spaces face under different urbanization models.

The low-quality development pattern, characterized by the combination of High 2DGV-Low 3DGV and Low 2DGV-Low 3DGV, is marked by superficial and single-layer greening. It is commonly observed in new urban districts, industrial parks, and areas undergoing extensive expansion (e.g., Hefei Binhu and Eastward new development zone and Zhengzhou airport economic zone), which are often driven by a land finance model prevalent in rapid urbanization. In those areas, the pursuit of short-term economic gain and low-cost land transfers may lead to minimal compliance with green space regulations, favoring low-cost, easy-to-maintain lawns while neglecting the long-term cultivation of arbors and multi-layered vegetation communities [88]. At the system scale, green space governance must transition from fragmented, incremental supply to integrated, structural-pattern management, shifting the focus from mere quantitative expansion to enhancing the structural quality and spatial connectivity. Cities should establish continuous green networks at the regional scale and incorporate binding 3DGV structure indicators (e.g., minimum canopy coverage) into new urban area planning. Concurrently, an evaluation system centered on ecological performance should be established, incorporating metrics such as thermal environment regulation and biomass enhancement into land supply and planning access mechanisms [108].

The spatially constrained pattern typically evolves from a combination of Low 2DGV-High 3DGV and High 2DGV-High 3DGV types, along with 2DGV Equitable Only distribution. It is characterized by limited green space and complex morphology. Typical areas include mountainous cities (e.g., Chongqing), ultra-high-density commercial centers (e.g., Futian in Shenzhen). This pattern arises from capital accumulation under high land prices and rapid infrastructure development. In urban commercial cores, elevated land values displace planar green space, compelling greening to seek vertical compensation. In mountainous cities, residential areas adopt low-density, passively designed layouts to navigate complex topography [109]. For this pattern, optimization strategies should begin by enhancing three-dimensional green spaces, which involves greening “pocket” spaces to fill gaps and establishing a positive correlation between 3DGV and floor area ratios to incentivize rooftop gardens and vertical green walls. Simultaneously, in mountainous cities, mountain trails and terraced greening should promote the forests’ accessibility [100]. Furthermore, a multidimensional capacity model based on urban morphology can be constructed to estimate potential green coverage. By prioritizing spatial governance, the fairness of 3DGV allocation can be enhanced. In cities on plains, green volume equity can be improved through land use adjustment, while in mountainous or high-density cities, enhancing 3DGV equity could possibly rely on the fine-grained design of micro-topography and vertical interfaces, which provides the theoretical basis for refined management [110].

The resource misallocated development pattern typically arises from the juxtaposition of two extremes: High 2DGV-High 3DGV with low equity, and Low 2DGV-Low 3DGV with low equity. Its core issue lies not in the scale of green space, but in the imbalance between spatial differentiation and resource allocation mechanisms. Typical examples are the urban layouts of Beijing and Shenzhen, where high-end communities coexist with aging urban villages. As a resource configuration mismatch, it reflects the conflict between spatial differentiation driven by markets and historical planning deficits. High-end residential areas with more green spaces may have established privatization barriers, resulting in exclusive access to green spaces and the displacement of low-income populations. Conversely, urban villages, old residential areas, and large affordable housing blocks suffer from a long-term deficit in green space due to a historical lack of effective planning interventions and sustained fiscal investment. At the systemic level, a green space redistribution framework focusing on public benefit and equity should be established, transforming green space into an instrument for equitable urban governance [99]. Through fiscal adjustments, cross-regional compensation, and strengthened public ownership, a minimum green space baseline should be established for urban areas. Simultaneously, institutional constraints should curb the exclusivity of gated communities’ green spaces, enabling shared access to premium green spaces.

4.4. Limitations and Future Prospects

This study still contains several limitations. For the 3DGV calculation process, the simplified volume model has an approximation of the complex crown geometries. While this assumption is widely accepted at the macro-scale to balance computational efficiency and data accessibility, it overestimates the total volume by including the space under the crown, and neglects the crown’s morphological shape, which may introduce certain biases in calculations [111]. Furthermore, constrained by data sources, the 10 m resolution canopy height data used in this study is limited in its precision for identifying small fragmented green spaces, rooftops, and vertical greening elements. Although the 3DGV model calculation process, thresholds, and the results were subjected to sensitivity analysis and field validation, relying solely on remote sensing data failed to capture certain green spaces such as vertical green walls, balcony greening, and underbridge greening, potentially leading to an underestimation of green volume in built-up areas [112].

The neighborhood scale analysis in this study included only three sample cities: Bazhong, Suzhou, and Wuhan. While these cities were chosen based on urbanization levels, topographical features, and green space structures, they may still not fully reflect the diversity and complexity of neighborhood green coverage distribution in other cities [113]. Future research is still needed to expand the sample cities using multiple classification criteria to explore heterogeneous green space distribution patterns across cities under different regional contexts.

Regarding 3DGV and related indicators, current analyses are still confined to spatial patterns and linear correlations with individual influencing factors. While these approaches partially reveal the impacting factors, the current framework still lacks the establishment of a comprehensive multi-factor impact assessment. Future research can introduce spatial econometric models to advance from correlation to causal analysis. Applying MGWR models or geographic detectors could further identify the geospatial correlation and influence intensity of different driving factors at various scales [114,115].

5. Conclusions

3DGV spatial variation not only reflects ecological differences but also impacts residents’ environmental experiences and health benefits. Areas with low 3DGV are more likely to face thermal stress, insufficient shading, and reduced outdoor comfort, diminishing residents’ daily health benefits and potentially exacerbating inequalities in urban environmental gains. By referencing to 2DGV evaluation metrics, this study constructed a 3DGV evaluation and equity analysis framework at both regional and neighborhood scales. A script in ArcGIS Model Builder was used to calculate the quantity and equity evaluation indicators for 413 Chinese cities with urban built-up areas exceeding 150,000 inhabitants, as well as residential and commercial districts in three representative cities—Wuhan, Suzhou, and Bazhong. The analysis examined spatial equity and structural disparities in green space distribution across the commercial and residential districts in three sample cities and identified the key influencing factors.

At the regional scale, the calculated 3DGV indicators across different cities generally exhibited low values concentrated in specific areas and regional variations, reflecting the complex interactions between natural environments and socioeconomic factors. Cities in Southwest, Northeast, and East China exhibited relatively higher 3DGV per unit area. This volume positively correlates with the city’s per capita GDP, aligning with the luxury effect observed in urban contexts. Conversely, consistently lower values in North China, Northwest China, and Central China revealed a tendency in planning management, driven by land finance, to prioritize meeting green space ratio targets while neglecting multi-layered greening. Correlation analysis indicates divergent distributions between 2DGV and 3DGV indicators, suggesting that 2DGV metrics do not accurately reflect the actual spatial distribution of urban vegetation. The results also showed that 3DGV was constrained by urban spatial morphology, where construction land density exerted an inhibitory effect that quantified the compression of vertical ecological space by high-density development. At the neighborhood scale, every 3DGV indicator clustered within low-value ranges. Residential areas exhibited higher 3DGV per unit area and per capita compared to commercial districts, reflecting functional disparities in green space distribution. However, in Bazhong, the mountainous terrain necessitated slope-adaptive construction and low-density development, which preserved substantial native vegetation. Consequently, its 3DGV metrics surpassed those of other cities, demonstrating the positive influence of topography on multi-layered greening. The regional disparities not only reveal the correlation mechanism among economic level, development priority, and ecological structure, but also highlight divergent pathways for cities to enhance ecological well-being.

Simultaneously, 3DGV is constrained by urban spatial form, with construction land density exhibiting a negative impact factor, quantifying the compression of ecological space by high-density development. At the community scale, all 3DGV indicators clustered in low-value ranges, while residential areas exhibited higher 3DGV per unit area and per capita than commercial areas. Among the three cities, Bazhong’s mountainous terrain necessitated building construction that follows natural contours, and its low-density development preserved substantial native vegetation. Consequently, Bazhong’s 3DGV metrics outperformed other cities, demonstrating the terrain’s positive influence on vertical green volume. Based on the 2DGV-3DGV analysis, the results of the established urban green volume quantity and equity classification indicated that at both the city and neighborhood scales, High 2DGV–High 3D GV green volume type accounted for the highest proportion, with both planar green spaces and spatial green volume remaining relatively insufficient. The High 2DGV–Low 3DGV pattern was prevalent in densely developed areas where urban greening has yet to transition from simple to multi-layered vegetation. As urban terrain positively influences three-dimensional green coverage, the Low 2DGV–High 3DGV pattern occurred primarily in mountainous or highly vegetated urban areas. While urban green coverage was generally evenly distributed across cities, disparities persist in park green space allocation and in neighborhood-scale green volume distribution. As green spaces exhibit structural variations across various regions, typological analysis helps identify quantity gaps that affect residents’ experiences, thereby driving a shift from an “area-oriented” to a “structure-oriented” green space planning and design approach.

Based on the classification results, enhancing urban green space requires a differentiated management framework. The proposed three typical green space patterns (low-quality development, spatially constrained development, and resource misallocated development) indicate that urban green space planning will need to shift from horizontal expansion to performance and equity-oriented regulation. Low-quality development areas should prioritize strengthening vegetation structure through mandatory canopy targets and multi-layer indicators. Spatially constrained areas shaped by high land prices or complex terrain should link 3DGV requirements with floor area ratios and leverage vertical interfaces or micro-topographies as compensatory spaces. In resource-mismatch zones, green resources will need to be redistributed by enhancing publicness, adopting equity-based allocation, and applying anti-gentrification instruments. The strategies could not only increase green volume in cities but also improve the delivery of environmental and ecological benefits. By linking green volume structures to human well-being, this study provides a pathway for integrating ecosystem service performance and equity into urban green space planning and management.