Examining Residents’ Perceptions and Usage Preferences of Urban Public Green Spaces Through the Lens of Environmental Justice

Abstract

Improving the equity of urban public green space is crucial for residents’ well-being and is a key objective in green space planning. While most existing studies focus primarily on the spatial distribution characteristics of green space resources, fewer explore the mechanisms influencing residents’ subjective perceptions and preferences. This study, based on a survey of 1419 residents in Hunan Province, constructs a structural equation model (SEM) to investigate the impact of the unequal distribution of urban public green space on residents’ happiness, mediated by social–psychological factors such as environmental perception, sense of security, and neighborhood cohesion. Additionally, a random forest (RF) algorithm is employed to identify the main factors influencing residents’ green space usage preferences. The results demonstrate that equity in green space distribution significantly influences residents’ happiness through environmental perception (path coefficient γ = 0.744, p = 0.001), security (γ = 0.664, p = 0.001), and neighborhood cohesion (γ = 0.830, p = 0.001). Key factors influencing residents’ preferences for green space use include age, housing prices, and walkability, with walkability contributing 17.5%, green space equity contributing 11.0%, and age contributing 10.2% to the frequency of green space use. These findings are critical for developing fairer and more effective urban green space policies, contributing to the creation of a more sustainable, equitable, and satisfying urban environment.

1. Introduction

Global urbanization is accelerating. According to United Nations projections, the global urban population is expected to increase by 2.5 billion by 2050, accounting for 68% of the total population [1]. While this rapid urbanization fosters economic growth and infrastructure development [2], it also exacerbates the unequal distribution of public resources, with urban green space equity being a particularly pressing issue [3]. Urban green spaces include both natural areas (e.g., parks and forests) and artificial green spaces (e.g., waterfront greenways, community green spaces, and landscaped plazas) [4,5]. These spaces not only provide essential ecosystem services such as air purification, temperature regulation, and biodiversity conservation [6] but also enhance residents’ well-being by alleviating psychological stress, promoting social interactions, and strengthening community cohesion [7,8,9]. However, the spatial distribution of urban green spaces is often inequitable, disproportionately disadvantaging low-income and marginalized communities, which tend to have limited access to green spaces of adequate size and quality [10,11,12].

Spatial inequities in urban green space not only elevate mental health risks, such as anxiety and depression [13], but also contribute to chronic conditions, including obesity [14], diabetes [15], and cardiovascular diseases [16]. Of particular concern is the issue of “latent green space inequality”, where official statistics indicate an abundance of green spaces, yet residents perceive themselves as living in a “green desert [17]”, unable to truly benefit from these areas. This disconnect may stem from limited accessibility, poor environmental quality, safety concerns, and social exclusion, all of which shape residents’ actual experiences with urban green spaces. This raises a critical question: How do structural disparities in green space distribution undermine residents’ well-being through psychosocial mechanisms?

From a theoretical perspective, environmental justice theory and biophilic urbanism provide crucial frameworks for understanding these impacts. Environmental justice theory emphasizes that environmental resources, such as urban public green spaces, should be equitably distributed to ensure that all social groups—particularly low-income populations, the elderly, and immigrant communities—have equal access to environmental benefits [18,19]. However, in practice, real estate development and urban renewal often lead to a concentration of green space resources in affluent neighborhoods, reducing accessibility and diminishing the experience of lower-income groups [20]. Biophilic urbanism, on the other hand, posits that humans have an innate need to connect with nature, which enhances psychological well-being, alleviates stress, and improves overall happiness [21]. When green space distribution is uneven, it can weaken residents’ perceived environment, sense of security, and community belonging, ultimately affecting their well-being [22]. These two theoretical perspectives are deeply intertwined in the context of urban green space allocation, providing a foundational framework for understanding how distributional inequities impact residents’ quality of life [23].

Current research on urban green space equity primarily focuses on several key areas. The first is objective measurement of green space equity. Previous studies have employed advanced technologies such as geographic information systems (GIS) and spatial analysis models to examine the physical distribution characteristics of green spaces in depth. For instance, in terms of green space coverage, national social surveys and land-use data have been used to quantify green space proportions across different regions [24,25]. Regarding spatial accessibility, transportation network analysis and spatial lag regression models have been applied to assess the ease of access to green spaces for residents [26,27,28]. In the context of service radius, buffer analysis has been utilized to delineate the effective service areas of various types of green spaces [29,30]. In urban areas globally, a substantial body of community-level research has underscored systemic disparities in the size, accessibility, and quality of green spaces [31,32,33]. For example, Zhang’s regional study in developed countries found that while most low-income communities tend to have greater green space availability, their actual accessibility remains limited. Additionally, neighborhoods with a higher proportion of Asian people had more UGS availability but less accessibility [25].

Another strand of research examines how sociodemographic characteristics influence residents’ actual access to and use of green spaces, addressing issues of social inequality and resource distribution [34,35,36]. For example, Lu et al. investigated the impact of socioeconomic factors on spatial equity across 263 cities in China. Their findings indicate that increases in GDP and population growth rates improve spatial equity in terms of community green spaces and street greenery. However, economic development and migration inflows exacerbate inequalities in public green space accessibility [20]. Similarly, Silviya Korpilo utilized Public Participation Geographic Information Systems (PPGIS) to assess environmental justice in relation to green and blue spaces (e.g., rivers and lakes) on Amager Island, Copenhagen, Denmark. The study authors found that gender significantly influences residents’ perceptions of green space equity [37]. In addition, some studies have focused on the profound impact of inequalities in green space utilization on public health [38,39,40]. For example, Gao et al. found that during the COVID-19 pandemic, patterns of park usage in the United States varied significantly based on socioeconomic status, urban–rural location, and population mobility. Due to the unequal distribution of green spaces and safe social distancing policies, pre-existing disparities in green space accessibility may have been further exacerbated, negatively affecting residents’ mental health and social interactions [41]. In Europe, studies conducted in the United Kingdom and Germany indicate that policymakers have increasingly recognized the impact of green space equity on public health and have begun integrating equity-based strategies into urban planning to optimize the distribution of green infrastructure [42,43].

However, while these studies have revealed the role of green space equity in physical accessibility, sociodemographic influences, and public health, relatively few have systematically analyzed residents’ subjective perceptions and behavioral preferences. In particular, there is a lack of systematic exploration regarding how psychological and social mechanisms—such as environmental perception, sense of security, and neighborhood cohesion—mediate the relationship between green space equity and well-being. To fill this research gap, this study focuses on the equity of public green spaces at the community scale and investigates how it affects residents’ well-being through psychological and social mechanisms. Specifically, this study involves two core analytical steps: (1) constructing a structural equation model (SEM) to verify the impact of perceived public green space equity on well-being from the residents’ perspective, revealing how multi-level factors interact and shape the perception of green space equity, and (2) applying the random forest (RF) algorithm to identify key variables influencing residents’ preferences for public green space use in Hunan Province and analyze the effects of various factors on residents’ green space usage behavior. By integrating biophilic urbanism theory and the environmental justice framework, this study enhances the understanding of public green space equity and its social impacts, providing scientific support for urban planning and policy development and contributing to the fair and efficient allocation of urban green space resources.

2. Conceptual Framework and Assumptions

The conceptual framework of this study is inspired by environmental justice theory and biophilic cities theory, aligning closely with the research hypotheses and model. The core hypothesis posits that green space equity indirectly influences residents’ well-being through psychological mechanisms (i.e., environmental perception and perceived safety) and social mechanisms (i.e., neighborhood cohesion), while sociodemographic factors play a critical role in shaping perceptions of green space equity.

2.1. Theoretical Foundation and Key Variables

2.1.1. Environmental Justice, Biophilic Cities Theory, and Green Space Equity

The concept of environmental justice emerged in the early 1980s during environmental protests in North Carolina, USA. It emphasizes the necessity of equitable distribution of environmental benefits, ensuring that marginalized or low-income communities do not disproportionately bear environmental burdens or face exclusion from essential resources [44,45]. Over time, the scope of environmental justice has expanded beyond pollution-related concerns to encompass broader access to environmental resources, including air quality, green spaces, and parks [46]. Green space equity, situated within the broader environmental justice framework, refers to equal access to green space resources among different social groups, particularly low-income populations, ethnic minorities, and other socially disadvantaged groups [20,47]. It advocates for fair access to environmental benefits and equitable participation in environmental decision-making [21,48]. However, green space distribution remains highly unequal, with low-income and marginalized communities facing substantial barriers to accessing these resources.

Biophilic cities theory, rooted in the biophilia hypothesis, posits that humans have an innate affinity for nature and derive psychological and physiological benefits from natural environments [49]. This theory has been widely applied in urban planning and environmental psychology to explain how green spaces influence mental health, social cohesion, and overall well-being [49]. In highly urbanized settings, biophilic cities theory underscores that green spaces serve not only as ecological assets but also as crucial elements in promoting social equity and enhancing residents’ well-being [23].

2.1.2. Perceived Safety

Perceived safety refers to residents’ subjective sense of security in their living environment, shaped by the interaction between social and physical attributes. Research indicates that the equitable distribution of green spaces contributes to crime reduction, strengthens community cohesion, and enhances social relationships [50,51,52]. Conversely, unequal green space distribution may lead to increased feelings of insecurity. For instance, green spaces in low-income areas are often of lower quality, lack sufficient lighting and visibility, and may become hotspots for criminal activities, thereby heightening residents’ sense of insecurity [52,53]. Thus, ensuring the fair distribution of green spaces not only provides a safer environment but also fosters a greater perception of safety among residents.

2.1.3. Neighborhood Cohesion

Neighborhood cohesion refers to the social bonds and sense of community developed among residents [54,55]. It encompasses social interactions, a sense of belonging, and cooperative spirit, all of which are key aspects of social identity and support networks. Strong neighborhood relationships help reduce social isolation, foster a sense of belonging, and provide both emotional and practical support, ultimately enhancing overall well-being [56,57]. Equitably distributed green spaces serve as communal spaces that facilitate social interactions and strengthen neighborhood cohesion. In contrast, unequal access to green spaces can lead to social fragmentation and mistrust among community members, thereby weakening neighborhood cohesion.

2.1.4. Environmental Perception

Environmental perception refers to individuals’ recognition and subjective evaluation of various environmental attributes and their associated meanings, including green space quality, maintenance conditions, and landscape aesthetics [58,59]. The quality and accessibility of green spaces are closely linked to residents’ environmental perceptions [60,61]. In low-income areas, the scarcity or poor maintenance of green spaces may contribute to negative environmental perceptions, potentially affecting residents’ mental health [62]. Conversely, high-quality green spaces—characterized by extensive vegetation coverage and proper maintenance—can enhance residents’ environmental perceptions, improve life satisfaction [63], and promote psychological well-being [64].

2.1.5. Well-Being

Well-being is commonly defined as a subjective experience associated with personal happiness and the pursuit of a fulfilling life [65]. The biophilia hypothesis supports the strong connection between green spaces, psychological restoration, and emotional regulation, suggesting that humans have an evolutionary psychological and physiological dependence on natural environments. As a result, natural spaces provide psychological comfort and physical relaxation, thereby enhancing well-being [66,67]. Extensive research has demonstrated a significant association between green space accessibility, usage frequency, and residents’ subjective well-being. Adequate green spaces not only provide a safe and comfortable environment but also promote physical health and alleviate psychological stress [68,69].

2.1.6. Sociodemographic Indicators

Sociodemographic indicators, including housing prices, monthly income, length of residence, health status, walkability, and rent forecasts, significantly influence individuals’ perceptions of green space equity and environmental conditions at both personal and community levels. These factors reflect disparities in socioeconomic status and residential environments. Regarding housing prices, monthly income, and rent, education level, income, race, and rising housing costs are commonly used to assess middle-class status [70,71]. Well-equipped urban parks and community green spaces are often concentrated in high land-value areas, making them more accessible to high-income groups [45,72]. Previous studies indicate that urban low-income populations visit parks significantly less frequently than affluent groups [73]. Length of residence is also closely linked to green space availability [73]. While affluent residents are generally less affected by nearby green space availability, individuals with lower socioeconomic status may experience reduced park visitation due to limited access [74]. Health status plays a critical role in green space usage, as contact with nature fulfills various psychological and physiological needs [49]. Studies show that residents who perceive themselves as healthy are more likely to visit parks [75], and increased green space accessibility is significantly correlated with reduced incidence of chronic diseases [76]. In terms of walkability, research suggests that strategically planning green spaces within a reasonable walking distance can yield substantial health benefits. A case study in Germany found that older adults with abundant, easily accessible nearby green spaces were more likely to visit them daily [75].

2.2. Hypothesis Development

This framework highlights the complex relationship between green space equity and residents’ psychological well-being (Figure 1). It emphasizes not only the physical presence and distribution of green spaces but also the integration of residents’ social and environmental perspectives. By incorporating socioeconomic factors into this model, the study underscores the necessity of considering equity and inequality in green space accessibility when formulating urban planning decisions. Understanding these interactions can provide valuable insights for equitable urban green space allocation, promoting social justice and enhancing residents’ well-being. Based on this framework, we propose the following hypotheses:

H1: 

Green space justice has a significant positive impact on residents’ environmental perceptions.

H2: 

Residents’ environmental perceptions have a significant positive impact on their happiness.

H3: 

Green space justice has a significant positive effect on neighborhood cohesion.

H4: 

Neighborhood cohesion has a significant positive effect on residents’ happiness.

H5: 

Green space justice has a significant positive impact on residents’ perception of safety.

H6: 

Residents’ perceptions of safety have a significant positive impact on their happiness.

H7: 

Green space justice has a significant positive effect on residents’ happiness.

H8: 

Sociodemographic indicators (housing price, monthly income, years of residence, health status, walkability, and rent prediction) have a significant positive impact on green space justice.

3. Methodology

3.1. Research Area

The study area is located in Hunan Province, in the central and southern part of China (longitude 108°47′ to 114°15′; latitude 24°38′ to 30°08′) (Figure 2). Hunan has diverse geological and topographical features, with mountains surrounding the east, south, and west, and flat terrain in the central and northern regions. The province experiences a subtropical monsoon climate, with distinct seasons, abundant sunlight, and rainfall and rich vegetation with a high forest coverage rate. Although the green space area in built-up areas exceeds 125,000 hectares, the distribution of green spaces remains uneven. This imbalance is mainly evident in the significant differences in green space resources across urban areas. Some older districts, due to the lack of forward-thinking planning, have limited green space, and the per capita green space is far below the urban average. In contrast, newly developed areas, while emphasizing green space construction in their planning, face issues with unreasonable layouts, and some large green spaces are sparsely populated, failing to fully serve local residents. Additionally, there are flaws in the distribution of green space types. Park green spaces are mostly concentrated in city centers, meeting the leisure needs of residents in these areas. However, peripheral neighborhoods lack sufficient community green spaces, which fail to meet the daily recreational and fitness needs of local residents, hindering the balanced improvement of the overall quality of life in the city.

3.2. Data Sources and Data Collection Methods

The questionnaire data were collected through in-depth field visits and online surveys conducted by the research team in various communities within the built-up areas of Hunan Province, from December 2023 to December 2024. The sample was obtained using a random sampling method. Questionnaires were distributed randomly in several typical public green parks and communities adjacent to public green spaces. Additionally, a number of online questionnaires were distributed via an online survey platform to supplement the field survey data. A total of 1510 questionnaires were distributed, and 1419 valid responses were received. The sample size meets the standard statistical requirements.

The questionnaire was designed in two sections. The first section covered sociodemographic characteristics, green space use, and willingness to pay for green space. This section included questions on gender, age, housing status, education level, income status, marital status, health status, perceptions of urban parks and neighborhood green spaces, frequency of visits, and accessibility of green space. The second section comprised the perceived safety scale, the green space justice scale, the well-being scale, the neighborhood cohesion scale, and the perceived environment scale (

Table 1. Scale description.

Aspect	Number	Scale Description
Green space justice [37]	GSJ1	I understand the various ways and opportunities available to be involved in the environmental decision-making process for public green spaces.
	GSJ2	I believe I have sufficient opportunities to express my views prior to environmental decisions regarding public green spaces.
	GSJ3	I think there is an effective communication and negotiation mechanism in the decision-making process for urban public green spaces.
	GSJ4	I am satisfied with the transparency of environmental decision-making for the city’s public green spaces.
	GSJ5	I believe that the process of making green environmental decisions in our city is fair
	GSJ6	In urban public green space environmental decision-making, the needs and interests of different groups are balanced and considered
Perceived safety [77]	PS1	There are no unsafe people hanging around public green spaces in the city
	PS2	There are no feared animals near public green spaces in the city (e.g., stray dogs, stray cats, dogs without a leash, etc.)
	PS3	Do you feel safe walking alone after dark near a public green space in a city?
	PS4	There are no unsafe small laneways near urban public green spaces
Well-being [79]	WB1	My everyday life is full of interesting things
	WB2	I feel peaceful and relaxed
	WB3	I feel happy and relaxed
	WB4	I feel clear and energetic
Perceived environment [78]	PE1	I am satisfied with the aesthetic design and overall landscape of the urban green space near my community
	PE2	I think the city green is close to where I live and can be easily accessed
	PE3	The urban green space near my community is well maintained
	PE4	In general, I have a high degree of overall satisfaction with the urban green spaces I usually visit
	PE5	I think my neighborhood is green in the city
	PE6	I think the urban green space near my community is spacious enough for all kinds of activities
	PE7	I think the urban green space near my neighborhood has improved the air quality around me
	PE8	The urban green space near my community provides a quiet environment and reduces the interference of urban noise
Neighborhood cohesion [80]	NC1	The people around me are willing to help their neighbors
	NC2	I recently met with one of my neighbors
	NC3	My community is a close-knit community
	NC4	I am free to express my personal views in my community

). The perceived safety scale measured community safety and security perceptions using the four-point neighborhood environment walkability scale (NEWS) [77], focusing on pedestrian and personal safety. The green space justice scale assessed respondents’ involvement in neighborhood and green space decisions [37]. The perceived environment scale included aspects of aesthetic design, noise levels, temperature regulation, and recreational space [78]. The well-being scale was derived from the C-WEMWBS Warwick–Edinburgh Mental Well-being Scale [79]. The neighborhood cohesion scale was adapted from Sampson et al. [80]. To ensure the structural validity of the scales, confirmatory factor analysis (CFA) was performed for each scale. The analysis results show that all of the indices meet the ideal standard [80]. Each scale was rated on a 5-point Likert scale (1 for “Strongly Disagree”; 5 for “Strongly Agree”). Respondents spent an average of 2–3 min completing the questionnaire.

3.3. Data Processing and Analysis

This study focuses on green space equity, residents’ well-being, neighborhood cohesion, environmental perception, and sense of security, systematically analyzing their interrelationships using a combination of SEM and RF. SEM is employed to validate causal relationships and elucidate the structural pathways of latent variables, while RF is utilized to identify nonlinear patterns in high-dimensional data, thereby enhancing the explanatory power of SEM.

3.3.1. Structural Equation Modeling

SEM is first used to examine the impact pathways of green space equity on environmental perception, perceived safety, and neighborhood cohesion. This approach evaluates how psychosocial mechanisms influence residents’ well-being, assessing both direct and indirect effects [81]. In this study, questionnaire data were coded into numerical values to ensure consistency and standardization. Subsequently, AMOS software was employed to estimate both the measurement and structural models. In this study, green space equity and sociodemographic indicators are treated as independent variables, while well-being serves as the dependent variable. The mediating variables include environmental perception, perceived safety, and neighborhood cohesion. The measurement model equations are formulated as follows:

$$x={\Lambda }_x\xi +\sigma$$

(1)

$$\mathrm{y}={\Lambda }_y\eta +\epsilon$$

(2)

where $x$ is the vector composed of exogenous explicit variables; ${\Lambda }_x$ is the relationship between exogenous explicit variables and exogenous latent variables, i.e., the factor loading matrix of exogenous explicit variables on exogenous latent variables; $\xi$ is the vector composed of exogenous latent variables; and $\sigma$ is the error term of exogenous explicit variables $x$; $y$ is the vector composed of endogenous explicit variables; ${\Lambda }_y$ is the relationship between endogenous explicit variables and endogenous latent variables, i.e., the factor loading matrix of the endogenous explicit variables on the endogenous latent variables; $\eta$ is the vector composed of endogenous latent variables; and $\epsilon$ is the error term of the endogenous explicit variables $y$.

The regression equation for the structural model is as follows:

$$\eta =B\eta +\Gamma \xi +\zeta$$

(3)

where $\eta$ is the vector consisting of endogenous latent variables; $\xi$ is the vector consisting of exogenous latent variables; $B$ is the relationship between endogenous latent variables, i.e., the matrix of path coefficients between endogenous latent variables; $\Gamma$ is the effect of exogenous latent variables on endogenous latent variables, i.e., the matrix of path coefficients between exogenous latent variables and endogenous latent variables; and $\zeta$ is the residual term of structural equation modeling.

3.3.2. Random Forest

Secondly, the random forest algorithm was employed to evaluate and compare the relative significance of residents’ sociodemographic background, economic factors, environmental factors, and psychological factors in influencing preferences for green space usage. This analysis aimed to identify which characteristics most substantially contributed to predicting green space use preferences, thereby complementing the findings from the SEM [82]. This algorithm constructs multiple decision trees and combines their outputs—either by voting or averaging—to make a final prediction. By aggregating the predictions of individual trees, the algorithm reduces variance and increases robustness against noise and outliers, thereby enhancing stability and accuracy. Furthermore, random forests tend not to overfit the data, enabling the model to generalize well to unseen data [83].

This study employed a random forest model to elucidate the relationship between preferences for green space usage and its associated factors, as illustrated in Figure 3. The dependent variables include the frequency and duration of residents’ visits to public green spaces, while the independent variables encompass social and economic factors (such as housing prices, income, length of residence, age, etc.), environmental factors (including the walkability of green spaces and adequacy of green space availability), and psychological factors (such as perceptions of justice, sense of security, and neighborhood cohesion). By analyzing these determinants, this study bridges the gap between residents’ perceptions of fairness in green space allocation and their actual usage patterns.

4. Results

4.1. Results of the Sociodemographic Characteristics of the Participants

The results of the basic demographics indicated (

Table 2. Basic demographic description (n = 1419).

Abridge	Variant	Count	Percentage
SEX	Sex
	Male	797	56.2
	Female	622	43.8
AGE	Age
	18–24	63	4.3
	25–29	175	12.3
	30–34	406	28.5
	34–39	439	30.9
	40–44	171	12.1
	≥45	165	11.9
HP	House price (CNY)
	<5000	227	16.0
	5000–9999	845	59.5
	10,000–19,999	332	23.4
	>19,999	15	1.1
MS	Marital status
	Unmarried or Divorced	162	11.4
	Married	1257	88.6
EL	Educational level
	Junior high school and below	48	3.4
	High school or junior college	59	4.2
	Higher vocational or college	275	19.4
	Undergraduate	800	56.4
	Graduate and above	237	16.7
MI	Monthly income (CNY)
	<3000	62	4.5
	3000–6000	255	18.0
	6000–9000	436	30.7
	9000–12,000	354	24.9
	12,000–15,000	195	13.7
	>15k	117	8.2
HC	Housing condition
	Renting a room	103	7. 3
	School or unit dormitory	27	1. 8
	Commercial/self-built housing	1142	80. 5
	Living with family	147	10. 4
DOR	Duration of residence
	<6 months	54	3.8
	6 months–1 year	221	15.6
	1–3 years	590	41.6
	3–5 years	223	15.7
	More than 5 years	331	23.3
HS	Health status
	Not so good	134	9.5
	Average	950	66.9
	Better	335	23.6

) that among the respondents, the proportion of males (56.2%) was slightly higher than that of females (43.8%). The main concentration was in the age group of 34–39 years, accounting for 30.9%. In the neighborhoods where the respondents resided, the majority of the respondents had a house price range of 5000–9999 CNY (59.5%). Regarding marital status, the majority of the respondents were married (88.6%). In terms of education, those with a bachelor’s degree were the most numerous (56.4%). In terms of the distribution of the respondents’ monthly income, it was mainly concentrated in the range of 6000–12,000 CNY, constituting 30.7% of the total number of respondents. In terms of housing status, the majority of the respondents (80.5%) owned commercial or self-built housing, demonstrating more stable living conditions. Renters accounted for 7.3%. With respect to the length of residence, the highest percentage of respondents (41.6%) had resided in a city or region for 1–3 years, indicating that a significant portion of the respondents had a relatively stable residence in a city or region. The proportion of respondents who had lived in a city or region for 6 months to 1 year was 15.6%, and the proportion of those who had lived in a city or region for 3–5 years was 15.7%, while the proportion of those who had lived in a city or region for a shorter period of time (less than 6 months) or a longer period of time (more than 5 years) was relatively small, accounting for 3.8% and 23.3%, respectively. The majority of the respondents (66.9%) considered their health condition average, suggesting that the respondents were generally neutral about their health condition.

In terms of green space usage (

Table 3. Description of green space use.

Abridge	Variant	Count	Percentage
SGS	Agree or disagree that city parks or community green spaces are sufficient
	Agree completely	430	30.3
	Partially agree	775	54.6
	Neutral	117	8.2
	Disagree	65	4.6
	Completely disagree	32	2.3
GSF	Frequency of use of nearby city parks or community green spaces
	(Almost) daily	114	8.0
	1–3 times per week	532	37.5
	1–3 times per month	476	33.5
	Rarely more than 1 time per month	297	20.9
RT	Length of stay
	<30 min	340	34.0
	30 min–1 h	730	51.4
	1–2 h	310	21.8
	More than half a day	39	2.8
ETW	Is it easy to walk to a nearby city park or community green space?
	Yes	1199	84.5
	No	220	15.5

), respondents were overwhelmingly positive about the adequacy of public green space near their neighborhoods. Among them, 30.3% of respondents fully agreed and 54.6% partially agreed, with the total of the two amounting to 94.9%. In terms of frequency of use, the proportion of respondents using the park 1–3 times per week is the highest, reaching 37.5%, showing that respondents have a certain regularity and frequency of using urban parks or community green spaces. In terms of length of stay, the proportion of respondents staying for 30 min to 1 h was the highest, at 51.4%. Among the respondents, the vast majority (84.5%) believed that the nearby urban parks or community green spaces were easy to reach on foot, indicating that the accessibility of urban parks or community green spaces is good for residents’ daily use and leisure activities.

In

Table 4. Description of willingness to pay for greenfield sites.

Abridge	Variant	Count	Percentage
PRI	Rent Premium Prediction: If a new city park is built near the neighborhood, you predict that rents will increase by
	<0.10%	196	13.8
	0.10–0.50%	519	36.6
	0.50–1.00%	354	24.9
	1.00–3.00%	210	14.8
	>3.00%	140	9.9
FHP	Home Price Premium Prediction: If a new city park is built near the neighborhood, you predict that home prices will increase by
	<0.10%	164	11.6
	0.10–0.50%	413	29.1
	0.50–1.00%	324	22.8
	1.00–3.00%	314	22.1
	>3.00%	204	14.4
WPG	Willingness To Pay: Of the two average houses in A (500 m from the greenbelt) and B (8 km from the greenbelt), how much would you be willing to increase your willingness to pay for house A in CNY?
	<100	241	17.0
	101–500	694	48.9
	501–800	348	24.5
	>800	136	9.6

, it can be clearly seen that residents’ expectations of rent increases due to the construction of a new city park near their neighborhood exhibit a skewed distribution. Nearly 40% of respondents (36.6%) predicted that rents would increase by 0.10–0.50%. Similarly, the expected increase in property prices also follows this distribution pattern, with nearly 30% of respondents (29.1%) anticipating a price rise of 0.10–0.50%. When asked about the additional budget they would be willing to pay for housing closer to green spaces, over 40% of respondents (48.9%) indicated a willingness to pay an extra 101–500 yuan, and about one-third of respondents (24.5%) were willing to pay an additional 501–800 yuan, indicating a high willingness to pay.

4.2. Reliability and Validity Analysis

The reliability and validity of the measurement model were assessed based on the valid data obtained from the questionnaire (see

Table 5. Results of the reliability and validity analysis.

	Variant	Cronbach’s α	CR	AVE
Green space justice	GSJ1	0.866	0.871	0.574
	GSJ2
	GSJ3
	GSJ4
	GSJ5
	GSJ6
Perceived environment	PE1	0.772	0.773	0.531
	PE2
	PE3
	PE4
	PE5
	PE6
	PE7
	PE8
Well-being	WB1	0.805	0.804	0.510
	WB2
	WB3
	WB4
Perceived safety	PS1	0.753	0.759	0.514
	PS2
	PS3
	PS4
	PS5
Neighborhood cohesion	NC1	0.784	0.702	0.516
	NC2
	NC3
	NC4

). The results indicated that Cronbach’s α coefficients for the five potential variables—green space justice, perceived environment, well-being, perceived safety, and neighborhood cohesion—ranged from 0.753 to 0.866. These values suggest acceptable reliability and meet conventional criteria (with α > 0.7 generally considered indicative of good reliability) [84]. Furthermore, the Composite Reliability (CR) for each potential variable ranged from 0.702 to 0.853, all exceeding the threshold of 0.70 and thus reaching an acceptable level [84]. Additionally, the Average Variance Extracted (AVE) values for each potential variable were greater than 0.500 [84], demonstrating that each construct within the scale possesses strong reliability and validity and can effectively measure the intended constructs. In summary, the questionnaire exhibited robust performance in terms of both reliability and validity, making it suitable for further analysis.

4.3. SEM Results

The hypothetical model’s fit was verified in this study using the maximum likelihood method. The fit indices of the final model following fitting correction are presented in

Table 6. Overall fit coefficients.

Index	X2	X2/df	RMSEA	IFI	CFI	TLI
Recommended value	The smaller, the better	The smaller, the better	<0.08	>0.9	>0.9	>0.9
Measured value	1943.236	4.299	0.048	0.915	0.915	0.901

. Based on the data in Table 5, it can be observed that the primary fit indices of this study fall within the recommended acceptable range, suggesting a good model fit [85].

The hypotheses of the research model were tested using AMOS. The standardized path coefficients between latent variables and their significance are presented in

Table 7. Structural modeling results.

Hypothesis	Estimate	p-Value	S.E.	Z
Green space justice → Perceived environment	0.744	***	0.027	19.606
Green space justice → Neighborhood cohesion	0.830	***	0.028	18.859
Green space justice → Perceived safety	0.664	***	0.025	16.113
Perceived environment → Well-being	0.184	***	0.054	4.031
Neighborhood cohesion → Well-being	0.512	***	0.083	8.137
Perceived safety → Well-being	0.327	***	0.057	7.646
Green space justice → Well-being	−0.105	0.098	0.053	−1.657
House prices → Green space justice	−0.326	***	0	−12.067
Monthly income → Green space justice	0.217	***	0.016	8.296
Residence duration → Green space justice	−0.106	***	0.018	−4.132
Health status → Green space justice	0.173	***	0.035	6.662
Walkability → Green space justice	−0.127	***	0.017	−4.895
Expected rent increase → Green space justice	−0.218	***	0.058	−8.256

*** indicates p < 0.001.

and Figure 4. Both the structural model, representing the hypothesized relationships, and the measurement model, defining the relationship between latent variables and their indicators, were estimated. As shown in

Table 8. Measurement model results.

	Variant	Estimate	p-Value
Green space justice	GSJ1	0.728	***
	GSJ2	0.720	***
	GSJ3	0.728	***
	GSJ4	0.682	***
	GSJ5	0.646	***
	GSJ6	0.660	***
Perceived environment	PE1	0.700	***
	PE2	0.524	***
	PE3	0.715	***
	PE4	0.641	***
	PE5	0.679	***
	PE6	0.647	***
	PE7	0.673	***
	PE8	0.659	***
Well-being	WB1	0.766	***
	WB2	0.719	***
	WB3	0.705	***
	WB4	0.649	***
Perceived safety	PS1	0.600	***
	PS2	0.657	***
	PS3	0.535	***
	PS4	0.638	***
	PS5	0.596	***
Neighborhood cohesion	NC1	0.654	***
	NC2	0.530	***
	NC3	0.672	***
	NC4	0.526	***

*** indicates p < 0.001.

, the factor loadings of all measured indicators exceed the acceptable value of 0.5 [86]. The path coefficients (γ) and their significance levels can be interpreted as follows:

When the path coefficient (γ) is positive, it indicates a positive relationship or direct effect between the variables, meaning that an increase in one variable leads to an increase in another. When the path coefficient (γ) is negative, it indicates a negative relationship or inverse effect, where an increase in one variable leads to a decrease in another. A path coefficient with a p-value greater than 0.05 suggests that the relationship between the variables is not statistically significant.

In the results, green space justice had significant positive effects on perceived safety (γ = 0.664, p = 0.001), perceived environment (γ = 0.744, p = 0.001), and neighborhood cohesion (γ = 0.830, p = 0.001). These factors, in turn, significantly influenced happiness, with perceived safety (γ = 0.327, p = 0.001), perceived environment (γ = 0.184, p = 0.001), and neighborhood cohesion (γ = 0.512, p = 0.001) all showing positive effects on happiness. However, green space justice itself did not significantly impact happiness. Additionally, housing price (γ = −0.326, p = 0.001), length of residence (γ = −0.106, p = 0.001), and projected rent premium (γ = −0.218, p = 0.001) had significant negative effects on the fairness of green space, while monthly income (γ = 0.217, p = 0.001) and personal health status (γ = 0.173, p = 0.001) had significant positive effects on green space justice.

4.4. Mediating Effect

The bootstrap method was used to evaluate the mediating effect, and the estimates were derived based on the bias-corrected interval. A statistically significant mediating effect is indicated if the 95% confidence interval (95% CI) for the indirect effect does not include zero. The model produces both direct and aggregate indirect effects (see

Table 9. Mediating effect results.

Mediation Path	Indirect Effect	Coefficient	p-Value (Two-Tailed)	95% CI Estimate
Green space justice → Environmental perception → Happiness	0.188	0.625	0.000 ***	[0.117, 0.256]
Green space justice → Security perception → Happiness	0.120	0.504	0.000 ***	[0.07, 0.171]
Green space Justice → Neighborhood cohesion → Happiness	0.144	0.771	0.000 ***	[0.085, 0.21]

*** indicates p < 0.001.

).

Table 9 shows that the 95% CI values of “green space justice → environmental perception → happiness”, “green space justice → security perception → happiness” and “green space justice → neighborhood cohesion → happiness” do not contain zero, indicating that there is a significant mediating effect in these paths.

4.5. Green Space Preferences: Determinants and Relative Importance

To explore the factors influencing residents’ preferences for park usage, “the frequency of residents’ use of nearby urban parks or community green spaces” and “the duration of residents’ stay in these areas” were selected as indicators of their preference for public green spaces. A random forest model was employed to assess the impact of these factors on park usage preferences and identify the most significant features contributing to the predictions. The dataset was divided into a training set and a test set in an 80:20 ratio; 80% of the samples were used for training, while 20% were reserved for validation. The model’s overall accuracy for predicting “the frequency of residents’ use of nearby urban parks or community green spaces” was 67.5%, with a prediction accuracy of 83.2%. For predicting “the duration of residents’ stay in nearby urban parks or community green spaces”, the overall accuracy was 60.2%, with a prediction accuracy of 77.0% (Figure 5).

The random forest model also assessed the importance of each variable. For predicting “the frequency of residents’ use of nearby city parks or community green spaces”, the most influential variables were ease of walking (17.5%), green space justice (11.0%), and age (10.2%). In the model predicting “the duration of residents’ stay in nearby city parks or community green spaces”, the top factors were housing price (11.5%), environmental perception (10.9%), and age (9.4%). The analysis reveals that, among social background factors, age is the most significant determinant, while housing price is the key factor among residential environmental influences, and anticipated rent increases play a notable role in market factors.

5. Discussion

5.1. Discussion of the Findings

This study employs a mediation model and random forest algorithms to assess the impact of green space equity on subjective well-being and explores the mediating roles of environmental perception, perceived safety, and neighborhood cohesion. Specifically, this study validates the mediation pathway of “green space equity → perceived mediation → happiness”, which resonates with Bratman’s environmental neuroscience theory: green spaces indirectly promote mental health by reducing amygdala activation (corresponding to an increase in the sense of safety) and enhancing prefrontal cortex function (corresponding to optimized environmental cognition) [87]. The mediating role of neighborhood cohesion further supports the social capital theory—green spaces, as “third spaces”, enhance community belonging by fostering casual social interaction [88]. Notably, the insignificance of direct effects suggests that the impact of green space equity on residents’ well-being unfolds gradually through a series of long-term socio-psychological mechanisms, validating Kellert’s “hierarchical model of nature contact” [89]: when environmental attributes (green space equity) are translated into individual perceptual experiences (environmental/safety perceptions) and social capital (neighborhood cohesion), deep improvements in well-being are achieved.

The negative impact of housing prices and rent on green space equity reveals the erosion mechanism of spatial commodification on public resources. On one hand, rental prices are typically closely tied to land values and income levels in urban areas. In high-rent regions, land value often takes precedence, leading developers and governments to prioritize projects that can enhance the economic value of the area rather than investing in the construction and maintenance of public green spaces. Consequently, rising rental prices may result in a reduction in green space resources, particularly in high-demand areas where green spaces are considered “inefficient” and may be converted to commercial or residential use. This phenomenon aligns with Logan and Molotch’s “Urban Growth Machine” theory [90], which argues that developers and governments, in their pursuit of economic growth, often sacrifice public spaces to increase capital returns. As a result, green spaces that should improve residents’ well-being may be compressed due to real estate market pressures, further undermining the achievement of green space equity. On the other hand, with the expansion or improvement of green spaces, property values and rents in surrounding areas tend to rise, a phenomenon commonly referred to as “gentrification” [91]. In the gentrification process, as environmental quality improves, low-income residents who previously lived in the area are often forced to relocate due to the rising rent costs. An example of this can be seen in the Lene-Voigt-Park in East Leipzig [92], which was upgraded through green strategies and the redevelopment of surrounding brownfields. However, as housing costs increased, previous residents, including the elderly and less affluent families, had to leave the area. As a result, the benefits of green space improvements do not reach the original residents and may, in fact, exacerbate social inequality.

The impact of residential tenure on green space justice is a multi-layered issue. On one hand, long-term residents may have become accustomed to the existing green space environment, resulting in a diminished sensitivity to issues or injustices related to green spaces [93]. As their length of residence increases, residents may focus more on issues directly related to their daily lives, such as the aging of housing facilities or neighborhood relations, while their attention to macro-level public resource issues, like green space equity, decreases. In other words, long-term residents might overlook the inequities in green space allocation due to their habituation to the existing environment, which in turn affects both their perception and the realization of green space justice. On the other hand, existing research has shown that the availability of green spaces impacts different socioeconomic groups’ residential tenure differently. Specifically, for low-income groups, as their length of residence increases, the green space resources available to them may actually decrease, leading to a more negative perception of green space justice among these groups [74]. Therefore, future green space planning could implement a “participatory green space planning” mechanism [94], establishing long-term and dynamic community green space feedback systems. To address the reduced sensitivity of long-term residents to green space issues, periodic community green space surveys could be conducted to actively collect residents’ opinions.

The analysis using the random forest model further reveals the key factors influencing the frequency and duration of public green space use. Walkability and age are particularly significant. This finding corroborates the “15-min neighborhood” theory [95]. Within the framework of the “15-min neighborhood” theory, it is emphasized that residents should be able to meet all of their daily needs within a 15 min walk, with walkability being one of the core elements. As C. Barbara explained, areas with high walkability can greatly increase residents’ willingness to travel, thereby increasing the frequency of public green space use [96]. When residents can easily walk to green spaces, they are naturally more inclined to visit and stay longer. Therefore, from the perspective of urban planning and construction, creating walkable streets and communities, increasing the planting of street trees, rationally designing sidewalks and bike lanes, and ensuring “instant natural contact” during daily commutes can enhance residents’ quality of life and well-being [45,97]. Age, in relation to public green space use frequency and duration, may reflect lifecycle behavior patterns, with different age groups having varying needs and preferences for green spaces [98]. Teenagers tend to visit local green spaces for socializing and physical exercise, middle-aged individuals use green spaces more due to parenting needs, and the elderly, constrained by physical limitations, tend to prefer small, local spaces within the community. This provides a basis for designing age-friendly green spaces, which suggests the need to incorporate pocket parks with resting facilities near communities. In summary, changes in individual factors are often insufficient to significantly increase green space utilization; a more comprehensive consideration of the synergistic effects of various factors is necessary. This will promote the implementation of more inclusive and equitable urban green space planning and policies.

5.2. Theoretical Implications

This study integrates environmental justice and biophilic urbanism theories to expand the theoretical framework of green space equity in relation to residents’ well-being. It verifies that the impact of green space equity operates through psychological and social mechanisms rather than direct effects. The theoretical contributions are threefold: First, green space equity is introduced as a psychological–social experience element, complementing biophilic urbanism theory by emphasizing its social welfare value beyond physical distribution; second, the study moves beyond the traditional quantity-based approach to green space, revealing the multi-dimensional and synergistic effects of quality, safety, and community interaction on well-being; third, by employing SEM and RF analysis, this research quantifies the nonlinear effects of housing economic factors on green space equity, providing a methodological reference for interdisciplinary studies.

5.3. Policy Implications

To enhance the equity of urban public green spaces and optimize their management for improved resident well-being, several policy recommendations are proposed: First, policymakers should prioritize improving green space quality rather than solely focusing on quantity. This can be achieved by enhancing maintenance, upgrading lighting and safety facilities, and optimizing spatial distribution to strengthen residents’ environmental perception, sense of security, and community cohesion. Second, a compensation mechanism for green space accessibility should be implemented in affordable housing areas to ensure equitable access for low-income residents. Integrating rent control with green space allocation can prevent green space appropriation during gentrification, thereby preserving social equity. Third, urban planning standards should be revised to improve green space accessibility and optimize pedestrian infrastructure. Fourth, future green space planning should incorporate participatory and age-inclusive planning mechanisms, establishing a long-term and dynamic community feedback system. Although this study provides practical recommendations, further exploration is needed to examine the impact of green space equity across different spatial scales.

5.4. Limitations and Directions for Future Research

Despite its contributions, this study has several limitations. First, it primarily focuses on the equity of public green spaces at the community scale, without addressing the allocation of larger-scale resources such as urban forests. Future research could adopt a multi-scale approach to explore the synergistic equity mechanisms between urban forests and community green spaces, offering a more comprehensive understanding of green space distribution. Second, as the study relies primarily on survey data, potential social desirability bias may affect the findings. Future studies could incorporate social media data, GPS trajectory analysis, and other methods to enhance the dynamic monitoring of residents’ green space utilization [99,100,101]. Additionally, the interaction between green space equity and urban systems, such as transportation networks and commercial layouts, remains underexplored. Future research could integrate spatial econometric models to analyze the coupling mechanisms among multiple factors, providing a more comprehensive understanding of the dynamic evolution of urban green space equity.

6. Conclusions

This study combines SEM and RF to systematically examine the relationship between green space equity and residents’ subjective well-being. The following key findings are drawn:

(1)

Green space equity primarily influences residents’ well-being indirectly through psychosocial mechanisms, with no direct effect. Specifically, environmental perception, sense of safety, and neighborhood cohesion play key mediating roles in this process.

(2)

Socioeconomic factors significantly affect perceptions of green space equity. In particular, housing prices, length of residence, walkability, and predicted rent have a significantly negative impact on green space equity perceptions. Monthly income and health status show a positive correlation with perceptions of green space equity.

(3)

Green space usage patterns are influenced by both individual factors and spatial characteristics. We found that walkability, green space equity perception, and age are key variables affecting both the frequency and duration of green space use.

These findings provide empirical support for environmental justice theory and biophilic cities theory, highlighting the multi-pathway influence of green space equity on well-being. This study offers practical insights for urban planners and policymakers aiming to foster equitable and health-enhancing green spaces.