Spatiotemporal Use Patterns and Perceived Health-Related Benefits of Pocket Parks: Evidence from Three Parks in Nanjing, China

Abstract

Rapid urban densification has intensified the scarcity of urban green space and challenged residents’ health and well-being. Pocket parks, as micro-scale infill green spaces embedded in the urban fabric, are increasingly adopted to expand everyday access to nature. Using three representative pocket parks in Nanjing, China, this study draws on self-reported data from questionnaire surveys and semi-structured interviews to characterize spatiotemporal use patterns and examine their links to perceived psychological, physiological, and social benefits through quantitative statistical analysis and modeling. Results show that pocket park use is highly routinized. Temporal patterns were evident, with weekend and autumn visits associated with improvements in emotional well-being, pain relief, and parent–child interaction. Perceived benefits were generally positive across psychological, physiological, and social domains, with psychological benefits—especially emotional relief and reduced loneliness—reported most strongly. Benefit levels varied across parks and user groups. Mechanism analysis reveals that the park supply factor, reflecting accessibility and basic facility provision, showed the most consistent direct paths to perceived benefits, whereas facility use and length of stay had no significant direct effects. These findings suggest that pocket park planning should prioritize accessibility and adequate basic provision, while strengthening activity support in ways that align with local use rhythms to enhance health-oriented performance in high-density cities.

1. Introduction

1.1. Background

Urbanization is accelerating worldwide at an unprecedented pace. According to United Nations statistics, more than half of the global population now lives in urban areas, and this proportion is expected to continue rising [1]. In developing countries such as China, large-scale rural-to-urban migration has further intensified urban expansion [2]. While rapid urban growth can generate economic vitality, it also brings major challenges. First, it can reduce per capita urban green space, limiting residents’ routine contact with nature [3]. Second, related environmental pressures, including urban heat islands and air pollution, can affect physical and psychological health as well as social and cultural life [4,5].

The ecological and health benefits of urban green spaces, especially large parks, have been widely documented. These spaces provide ecosystem services such as climate regulation, stormwater management, and air purification [6], and are associated with health outcomes including lower diabetes risk, reduced psychological stress, and obesity prevention [7]. However, the post-pandemic period has further exposed persistent shortages of accessible green spaces. Growing demand for healthier living environments and everyday activity spaces has become increasingly salient. In already built-up, high-density urban cores, creating large new parks often faces practical constraints such as scarce land, high costs, and complex stakeholder negotiations. Against this backdrop, pocket parks have emerged as a small-scale and feasible solution and have become an important instrument for urban regeneration and fine-grained governance.

The idea of pocket parks can be traced to post-World War II Europe and was later introduced in the United States. The term pocket park was associated with Robert Zion, who designed Paley Park in New York City and promoted the concept in the 1960s; one of the earliest dedicated publications appeared in the early 1980s [8]. Pocket parks convert underused parcels and irregular residual spaces into green nodes. This approach aligns with the notion of small-scale, flexible interventions emphasized in organic urban renewal theories in China [9], offering a pragmatic pathway toward healthier and more sustainable cities.

Pocket parks can be considered compact forms of neighborhood parks that directly serve nearby residents. When they are safe, well maintained, and accessible, pocket parks can provide functions similar to larger urban parks by offering convenient places for recreation, restoration, and everyday social interaction [10,11]. Through targeted design, they can support daily needs by providing seating and tables that facilitate social contact, walking paths and small play areas that enable light physical activity and children’s play [12]. Vegetation-rich green spaces in pocket parks correlate with reduced rates of depression and cardiovascular disease [13]. They enhance the microbial environment to improve the human immune function [14,15], while the aesthetic and recreational value of green spaces can also encourage physical activity and social interaction among residents, aiding health recovery [16,17]. Pocket parks may also take on location-specific roles. In Japan, they can function as emergency refuges in earthquake-prone contexts [18]. In flood-prone areas, they can contribute to risk reduction and emergency use [19]. In some cases, they have served as cultural and community identity spaces, for example through public art that reflects local heritage and strengthens community cohesion [11]. These multiple roles suggest that pocket parks can integrate health support, social interaction, community cohesion, environmental improvement, and site-specific cultural or safety functions within limited urban space. Understanding how such benefits are realized is therefore of both theoretical and practical importance for improving cities’ capacity to meet diverse resident needs and for advancing sustainable living environments.

Health-related benefits of pocket parks are closely connected to residents’ everyday activity patterns. Accessibility is often described in terms of a walkable catchment, commonly within a half-mile or about 10–15 min on foot [20]. Pocket parks typically have an even smaller service radius, often only one to four blocks, which makes them highly embedded within neighborhood fabric [21]. Such spatial embeddedness can support community interaction and social ties by providing opportunities for information exchange, casual encounters, and meeting neighbors [22]. Safe and well-maintained pocket parks may also enhance place attachment and a sense of belonging. For instance, the pocket park program in Philadelphia not only improved physical conditions but also generated positive social impacts by involving residents throughout the process, fostering participation, pride, and the mobilization of community resources [23]. Since pocket park functions depend fundamentally on how residents use these spaces in daily life, understanding when, how, and why residents use, or do not use, these highly accessible micro-spaces is essential for evaluating community benefits and informing evidence-based design and management.

1.2. Literature Review

International work on pocket parks gained momentum during the 1960s alongside small-park programs for high-density city centers. A foundational synthesis is Small Urban Spaces, edited by Whitney North Seymour, which examined the philosophy, design, sociological dimensions, and politics of pocket parks and other micro-scale urban open spaces [24]. Early practice was further consolidated in the Planning Advisory Service report Pocket Parks (Report 229), which emphasized that park size was largely constrained by land availability and summarized planning and design considerations for implementing such spaces [23]. Paley Park in New York City, opened in 1967 and designed by Robert Zion, is frequently cited as a prototypical case demonstrating how micro-scale public spaces can succeed in intensely built-up contexts. Behavioral observation research subsequently strengthened the field by showing how fine-grained design and management shape everyday use and social interaction in small urban spaces [25]. Building on these foundations, design guideline texts codified miniparks and pocket parks within broader typologies of urban open space and offered research-based recommendations to improve usability [21]. As practice diversified across contexts, research expanded from concept definition and spatial design toward empirical assessments of health, environmental, and social outcomes.

As practice advanced, research topics expanded from conceptual definition and spatial design to health, environmental, and social dimensions. Regarding public health, empirical studies have shown that pocket parks can function as restorative landscapes and contribute to both mental and physical well-being [26,27,28]. Evidence also suggests that pocket parks can encourage moderate-intensity physical activity and support healthier lifestyles [29,30,31]. More recent work has examined psychological and physiological outcomes, including stress reduction and perceived improvements in well-being [32,33]. Regarding environmental benefits, studies have reported that pocket parks can contribute to mitigating urban heat island effects [34] and may support urban biodiversity in certain contexts [35]. Research has also assessed the restorative potential of specific natural elements within small parks, such as lawns and combinations of natural features, indicating variation in perceived restoration across design components [36].

Equity, safety, and governance have become increasingly prominent themes. Some studies emphasize the importance of integrating sustainability and resilience into pocket park design [19]. Others adopt sociological perspectives and report that, during the COVID-19 period, pocket parks not only supported residents’ physical condition but also contributed to improved green space equity and living conditions in disadvantaged neighborhoods [37]. Comparative assessments across communities highlight social disparities in park distribution and accessibility [38]. Quantitative evidence further indicates that increasing nearby greenery can reduce the risk of loneliness, particularly for people living alone [39].

Despite the growing body of research, several gaps remain. First, spatiotemporal patterns of pocket park use are still insufficiently characterized. Existing studies often focus on functions and design principles, but provide limited evidence on time of day, visit frequency, duration, and detailed behavioral composition, which constrains optimization of use and outcomes. Second, multidimensional health evaluations remain limited. Many studies concentrate on single indicators, such as physical activity or psychological health, without examining how psychological, physiological, and social benefits co-vary and interact. Third, heterogeneity across pocket park types is often underexplored. Differences by location, size, and facility provision may shape both use patterns and benefits, yet evidence that directly compares diverse pocket park types remains scarce, limiting guidance for targeted planning.

To address these gaps, this study addresses three core questions: the characteristics of users’ spatiotemporal use patterns in pocket parks, the variations in perceived multidimensional health benefits across different parks and user groups, and the pathways linking park provision, user behaviors, and perceived health benefits. To investigate these questions, we conducted questionnaire surveys and semi-structured interviews in three representative pocket parks in Nanjing. It characterizes users’ spatiotemporal use patterns, including visit frequency, duration, timing, activity types, and access modes, and assesses perceived health-related benefits across psychological, physiological, and social domains. Analytically, the study applies descriptive statistics to summarize use patterns, nonparametric tests to compare perceived benefits across parks and socioeconomic groups, and Spearman correlation to examine associations between behaviors and perceived benefits. It further uses exploratory factor analysis and structural equation modeling to explore potential pathways linking park provision, user behaviors, and perceived outcomes. The findings provide empirical evidence to inform pocket park planning, facility provision, and management strategies in high-density urban contexts.

2. Materials and Methods

As a megacity in eastern China, Nanjing has experienced rapid urban expansion in recent decades, accompanied by an increasing imbalance between limited green space supply and residents’ growing ecological and health-related needs. To address this challenge, Nanjing initiated a citywide pocket park program in 2020, positioning pocket parks as flexible supplements to the urban green space system. Guided by planning targets that emphasize access to greenery within 300 m and access to a park within 500 m, the program aims to ensure that residents can reach a public green space within a 5–10 min walk, thereby substantially improving the spatial accessibility of urban public health resources.

Considering park themes, functional attributes, and surrounding built environments, this study selected three typical pocket parks located in the central urban area of Nanjing, China as case studies: Huowa Lane Red-themed Pocket Park (hereafter Park H) in Qinhuai District, Yuanmei Plaza (hereafter Park Y) in Gulou District, and Jingming Jiayuan Pocket Park (hereafter Park J) in Yuhuatai District (Figure 1). These three parks exhibit clear heterogeneity and cover the major types of pocket parks commonly found in downtown Nanjing. To ensure representativeness and social recognition, and to provide a sufficient user base for the subsequent questionnaire survey, all three parks were selected from the finalists of the 2023 “Most Popular Pocket Parks” list in Nanjing, which is a publicly recognized and authoritative list reflecting typical high-quality pocket parks in the city.

As summarized in

Table 1. Characteristics of the three pocket parks.

Name	Area (m2)	Size	Functional Type	Theme	Key Facilities
Huowa Lane Red-themed Pocket Park	1800	Small	Residential-oriented	Cultural	Corridor bridge, reading pavilion, fitness equipment
Yuanmei Plaza	3389	Medium	Public service-oriented	Leisure	Book bar, sculptural landscapes
Jingming Jiayuan Pocket Park	9998	Large	Transportation-oriented	Nature	Landscape bridge, children’s slide

, Park H is the smallest of the three and is classified as a residential-oriented pocket park. The “Wutongyu” initiative is a livelihood project led by the Municipal Urban and Rural Construction Commission, aimed at enhancing urban spatial quality through culturally enriched micro-service stations. As one of Nanjing’s early Wutongyu mini urban living rooms, Park H combines community services with cultural display functions and includes facilities such as a covered corridor bridge and a public reading pavilion. Park Y is a medium-sized, public service-oriented pocket park themed around leisure, featuring an indoor book bar and sculptural landscapes that together create a multifunctional open space. Park J, covering nearly 10,000 m², is a large transportation-oriented pocket park with a nature theme. With elements such as landscape bridges, a dry creek, and children’s play facilities, it offers a diverse, multi-purpose activity space and is a key project within Yuhuatai District’s enjoyable park initiative. As a provincial-level urban green space development program in Jiangsu Province, “Enjoyable Landscapes” has been listed among the provincial government’s livelihood projects every year since 2022. By developing high-quality public spaces such as pocket parks, the program focuses on enhancing residents’ experience of “seeing greenery when opening windows and entering parks when stepping out”.

2.1. Data Sources

Data for this study were collected primarily through an on-site questionnaire survey and semi-structured interviews. The questionnaire was designed to capture two key domains: spatiotemporal use behaviors and perceived health-related benefits. It comprised three sections: (1) sociodemographic information (including residential location); (2) spatiotemporal activity patterns (including visit frequency, length of stay, visit timing, and activity types); and (3) perceived health-related benefits, measured using a five-point Likert scale across psychological, physiological, and social dimensions. These data provided the foundation for characterizing use patterns and examining their associations with perceived benefits in subsequent analyses.

To ensure data quality and facilitate subsequent statistical analysis, the questionnaire predominantly used closed-ended questions. Perceived health benefits were measured using a five-point Likert scale, enabling a quantitative assessment of respondents’ self-reported benefits after park use. The wording of all items was kept concise and easy to understand to minimize potential misinterpretation by participants. The full questionnaire is provided in the Supplementary Materials (Questionnaires S1–S3).

The survey adopted an on-site convenience intercept sampling strategy. From October to November 2024, the research team administered questionnaires at the three parks on both weekdays and weekends, covering two daily time windows (morning and afternoon). A total of 362 questionnaires were distributed (Park H: 118, Park Y: 115, Park J: 129), and 310 valid responses were retained, including 102 from Park H, 101 from Park Y, and 107 from Park J. This resulted in a valid response rate of 85.6%.

2.2. Analytical Strategy

This study adopted a mixed methods approach, drawing on data derived from participants’ self-reported questionnaire responses, to integrated quantitative and qualitative evidence to examine spatiotemporal use patterns of pocket park users and their associations with perceived health-related benefits (Figure 2).

Quantitative analyses were conducted using SPSS 26.0 and AMOS 26.0. All variables used in the quantitative analyses were derived from self-reported questionnaire responses and therefore reflect perceived rather than objectively measured conditions. In particular, residential proximity, access convenience, and facility provision were measured as perceived assessments or categorical self-reports and were coded so that higher values indicate greater proximity, convenience, or provision. First, descriptive statistics were used to summarize participants’ characteristics and spatiotemporal behaviors. Because perceived health benefit variables were ordinal and did not satisfy normality assumptions, nonparametric methods were used for inferential analyses. The Kruskal–Wallis test is a nonparametric statistical technique that evaluates differences among three or more independently sampled groups [40], while the Mann–Whitney U test is used to determine significant differences between the distributions of two independent sample populations. Specifically, the Kruskal–Wallis H test and the Mann–Whitney U test were applied to examine whether perceived health benefits differed significantly across the three case parks and among groups with different socioeconomic characteristics. Spearman rank correlation was then used to explore the presence and strength of associations between spatiotemporal behavioral indicators and different categories of perceived health-related benefits.

To further examine the mechanisms linking park environment, user behaviors, and health-related outcomes, structural equation modeling was employed. Structural Equation Modelling (SEM) is a statistical technique used to explore the causal relationships among both observed (“manifest”) and unobserved (“latent”) variables [41]. Prior to model specification, exploratory factor analysis was conducted to reduce dimensionality across multiple environment and behavior variables and to extract a smaller set of latent factors. Based on these factors, a structural equation model was constructed to test direct pathways from the latent factors to three benefit dimensions, psychological, physiological, and social. Model fit indices were used to evaluate the consistency between the hypothesized model and the observed data.

2.3. Socioeconomic Characteristics of the Sample

As shown in

Table 2. Socioeconomic characteristics of the sample.

Category	Option	Total (n = 310)		Park H (n = 102)		Park Y (n = 101)		Park J (n = 107)
		Count	%	Count	%	Count	%	Count	%
Gender	Male	122	39.4%	44	43.1%	43	42.6%	35	32.7%
	Female	188	60.6%	58	56.9%	58	57.4%	72	67.3%
Age	Under 16	13	4.2%	8	7.8%	2	2.0%	3	2.8%
	16–29	74	23.9%	23	22.5%	36	35.6%	15	14.0%
	30–59	158	51.0%	41	40.2%	51	50.5%	66	61.7%
	60 and above	65	21.0%	30	29.4%	12	11.9%	23	21.5%
Education	Junior high or below	60	19.4%	16	15.7%	10	9.9%	34	31.8%
	High school/vocational secondary	86	27.7%	36	35.3%	28	27.7%	22	20.6%
	College/university	150	48.4%	46	45.1%	59	58.4%	45	42.1%
	Master’s degree or above	14	4.5%	4	3.9%	4	4.0%	6	5.6%
Employment status	Flexible employment	38	12.3%	11	10.8%	8	7.9%	19	17.8%
	Stable employment	104	33.5%	19	18.6%	58	57.4%	27	25.2%
	Not employed	66	21.3%	33	32.4%	19	18.8%	14	13.1%
	Retired	102	32.9%	39	38.2%	16	15.8%	47	43.9%
Residential location	Local residents (within 5 km)	239	77.1%	91	89.2%	63	62.4%	85	79.4%
	Visitors (beyond 5 km)	71	22.9%	11	10.8%	38	37.6%	22	20.6%

, the socioeconomic characteristics of the sample are presented below. Overall, females accounted for 60.6% of respondents, and middle-aged adults constituted a substantial share of the sample. Educational attainment was relatively high, with respondents holding college or university qualifications forming the largest group, followed by those with high school or vocational secondary education. Employment status was diverse, with sizable proportions of respondents in stable employment and those who were retired. According to the distance calculated based on home address, this study uses 5 km as the threshold to distinguish local residents from visitors. It can be seen that most respondents lived within approximately 5 km of the parks, indicating that the primary users of pocket parks are residents from surrounding communities. Regarding self-rated health, most respondents considered their health status to be good, with responses concentrated in the very healthy and moderately healthy categories. The prevalence of self-reported chronic disease in the sample was 19%.

Sample composition differed across the three parks. Female respondents were the majority at all sites, with the highest proportion observed in Park H at 67.3%. In each park, the 30 to 59 age group was dominant, although the age structure varied. Park H showed a more even distribution across age groups, while Park Y had the lowest share of older adults at 11.9%. Park J exhibited the strongest concentration of middle-aged users.

Educational attainment also varied by site. Respondents at Park H and Park J tended to report lower education levels, whereas Park Y was characterized by higher educational attainment. Correspondingly, Park Y had a higher proportion of respondents reporting very good health and a lower prevalence of chronic disease relative to commonly reported adult estimates. In terms of spatial reach, Park Y appeared to draw users from a wider area, whereas Park H showed a more strongly localized user base.

3. Results

3.1. Spatiotemporal Behavioral Characteristics of Park Users

3.1.1. Temporal Behavioral Characteristics

As shown in Figure 3, most respondents visited the parks at least once per week, indicating that pocket parks were frequently used and had a high level of participation in daily life. Among the three sites, Park H had the highest proportion of frequent visitors, suggesting a stronger neighborhood and community function. By contrast, Park Y had the largest share of low frequency visitors, which is likely attributable to its surrounding context being less residential and more oriented toward commercial and transit-related activities.

With respect to visit days, respondents most commonly reported visiting on weekdays and weekends, while visits during public holidays were less frequent. This pattern suggests that pocket parks primarily serve everyday needs rather than occasional holiday recreation. Park H showed a stronger weekday preference, which may be associated with its proximity to a school and routine after school visits. Park Y and Park J showed a more pronounced weekend preference.

The study also examined within-day timing preferences using two-hour intervals. Overall, park use exhibited a clear two peak pattern, with the highest number of visits occurring between 10:00 and 12:00 and between 15:00 and 17:00, while other time periods were less commonly reported. All three parks peaked between 15:00 and 17:00, indicating that users tended to visit more in the afternoon. For Park Y and Park J, reported visits reached a local minimum between 13:00 and 15:00, possibly reflecting constraints such as work schedules, meal times, or thermal discomfort during early afternoon hours.

In terms of average length of stay, more than 70% of respondents reported spending more than 30 min in the park, and most visits were within 1 h, indicating that pocket parks mainly support short duration use. Park H had the highest proportion of respondents staying 1 to 2 h, suggesting a stronger capacity to retain users compared with the other two parks. Park Y had the highest proportion of stays shorter than 30 min, and almost no respondents reported staying longer than 2 h, which may be related to its smaller size and more limited functional diversity.

Seasonal preferences showed a consistent pattern across the three parks. Spring and autumn were the most preferred seasons, with spring being particularly prominent, whereas summer and winter were generally less favored. This pattern likely reflects more comfortable climatic conditions and more attractive outdoor scenery during spring and autumn.

3.1.2. Spatial Behavioral Characteristics

Spatial behavioral characteristics of pocket park users were examined from two perspectives, accessibility and service catchment, and activity space needs. The results are shown in Figure 4.

Travel mode to the park reflects both the park’s effective service radius and the accessibility of the surrounding transport environment. Most respondents reported walking to the park, indicating that the service catchment was mainly concentrated within nearby walkable areas. This pattern was particularly evident in Park H and Park J, suggesting stronger reliance by nearby residents and a closer connection between the parks and surrounding residential areas. Overall, electric mopeds and bicycles together accounted for 21%, which further extends the service catchment to roughly 1 to 3 km and supports short distance access. This mode mix was especially relevant in contexts such as Park H, which serves an aging community, and Park J, which is embedded in a well-equipped residential neighborhood. By contrast, Park Y had the highest reported car use, indicating that it may attract users from longer distances and that its catchment potentially extends toward more distant commercial centers or industrial areas.

In terms of activity types, the core activities across the three parks were highly concentrated in two categories, fitness and social interaction. Fitness-related activities mainly include walking, square dancing, rope skipping, and so on, while social interaction primarily consists of gatherings and parent–child accompanying. This pattern reflects the fundamental and widely shared role of urban parks in supporting routine exercise and everyday social contact. Activity profiles nevertheless differed across sites. In Park H, fitness-related activities accounted for the largest share, such as jogging, fitness equipment training, and playing badminton, likely associated with its relatively comprehensive fitness facilities and a more spacious layout. In Park Y, the proportions of different activity types were relatively balanced. Supported by cultural spaces such as sculptural settings and an indoor book bar, Park Y had the highest share of culture and entertainment activities among the three parks, including instrument playing and small-scale cultural experiences. Leisure activities rank second. Visitors to Park Y usually enjoy the scenery, walk their pets or take photos. Park J showed a distinct pattern in which interaction-oriented activities were dominant, with examples like neighborly chatting, family resting, and children’s playing. Fitness-related activities also accounted for a certain share, while entertainment and leisure activities were less common.

3.2. Health-Related Benefits of Pocket Parks

3.2.1. Overall Health-Related Benefits Across Parks

Based on survey data collected in Park H, Park Y, and Park J, overall health-related benefits of pocket parks were examined across three dimensions, psychological, physiological, and social. The results indicate substantial variation between the three parks.

Psychological benefits. Psychological benefits were widely reported across all parks, although their magnitude varied by site (Figure 5). For mood improvement, Park H showed the strongest performance, with nearly 90% of users reporting a positive effect. Benefits related to loneliness reduction were most prominent in Park J, where about 70% of users perceived a substantial improvement and almost none reported no positive effect. Park H ranked second, while Park Y showed a comparatively weaker response in loneliness reduction. Perceived improvements in cognitive functioning, including concentration and mental agility, were highest in Park J, with nearly half of users reporting a strong effect. Park Y ranked next, whereas Park H showed more limited perceived gains. Overall, Park J exhibited the strongest combined psychological benefits, with close to 90% of users reporting substantial benefits.

Physiological benefits. Physiological benefits showed a pattern of stronger restorative effects but weaker outcomes related to chronic condition management (Figure 6). Fatigue relief, physical recovery, and sleep improvement were reported more frequently, with Park J performing best. More than 60% of users in Park J reported substantial benefits, which is consistent with its higher facility use. By contrast, Park H showed a weaker response in fatigue relief. Overall benefits related to body shaping and pain relief were limited. In Park H, around half of users reported no noticeable pain relief and insufficient improvement in body shaping.

Social benefits. Social benefits differed across parks (Figure 7). In Park H, the proportions of users reporting moderate benefits and substantial benefits were similar, suggesting that benefits were perceived by many users but were not strongly pronounced. In Park Y, most users reported moderate benefits. In Park J, 57% of users reported substantial social benefits. Across all parks, the proportion of positive responses regarding willingness to socialize was relatively high, indicating that pocket park use may support social interaction. However, effects varied across indicators and user groups. For parent–child interaction, Park H and Park Y showed stronger perceived benefits. Park J exhibited a more polarized pattern, with 25.2% reporting no effect and 29% reporting a strong effect, which may reflect heterogeneity in its user groups. Across the three parks, respondents generally reported positive effects on neighborhood relations and on enhanced motivation or positivity.

3.2.2. Differences in Perceived Health Benefits

Normality tests indicated that ratings on health benefit indicators were not normally distributed. Kruskal–Wallis H tests were therefore used to examine whether perceived benefits differed across the three parks. Results are presented in Figure 8 and

Table 3. Data for the difference test of perceived health benefits.

Perceived Health Benefit Indicators		Park		Gender		Employment Status		Residential Location		Education Level
		H	p	Z	p	H	p	Z	p	H	p
Psychological health	Improve mood	28.369	<0.001	−0.383	0.701	7.670	0.053	−2.640	0.008	5.532	0.137
	Alleviate loneliness	25.139	<0.001	−1.541	0.123	12.890	0.005	−1.403	0.161	5.963	0.113
	Enhance attention	20.172	<0.001	−1.631	0.103	1.505	0.681	−0.745	0.456	15.520	0.001
	Overall psychological health	44.628	<0.001	−2.570	0.010	6.691	0.082	−1.034	0.301	10.825	0.013
Physical health	Relieve fatigue	18.850	<0.001	−1.169	0.243	7.827	0.05	−1.006	0.314	6.994	0.072
	Boost energy	18.955	<0.001	−0.936	0.349	5.581	0.134	−2.720	0.007	12.364	0.006
	Improve physique	8.428	0.015	−1.304	0.192	5.720	0.126	−1.994	0.046	12.587	0.006
	Improve sleep	26.628	<0.001	−1.139	0.255	0.957	0.812	−3.720	<0.001	19.226	<0.001
	Relieve pain	11.765	0.003	−2.625	0.009	3.485	0.323	−2.264	0.024	10.967	0.012
	Overall physical health	26.820	<0.001	−0.610	0.542	2.249	0.522	−3.334	<0.001	10.910	0.012
Social health	Promote neighborhood harmony	0.839	0.657	−1.251	0.211	9.852	0.020	−4.940	<0.001	24.769	<0.001
	Facilitate parent–child communication	0.123	0.941	−2.898	0.004	3.584	0.310	−5.583	<0.001	10.399	0.015
	Increase social willingness	2.338	0.311	−2.316	0.021	7.801	0.050	−2.712	0.007	15.868	0.001
	Improve motivation	16.827	<0.001	−1.140	0.254	6.478	0.091	−1.996	0.046	9.887	0.020
	Overall social health	11.484	0.003	−2.288	0.022	12.790	0.005	−4.264	<0.001	7.496	0.058

.

Significant between-park differences were found for psychological benefits. For all psychological indicators and the overall psychological benefit score (H = 44.628), p values were below 0.001. Physiological benefits also differed between parks, with p values below 0.05 for individual indicators and the overall physiological benefit score (H = 26.820, p < 0.001). For social benefits, significant differences were observed for enhanced positivity in learning, work, and daily life (H = 16.827), and for the overall social benefit score (H = 11.484), with p values below 0.001 and 0.003, respectively. In contrast, differences were not significant for neighborhood relations (H = 0.839), parent–child interaction (H = 0.123), and willingness to socialize (H = 2.338), as indicated by larger p values.

To assess heterogeneity across population groups, additional nonparametric tests were conducted by gender, employment status, education level, and residential location. Results are shown in Figure 9 and Table 3. Significant gender differences were identified for pain relief (H = −2.625, p = 0.009), parent–child interaction (H = −2.898, p = 0.004), and willingness to socialize (H = −2.316, p = 0.021), with female users reporting higher scores. Residential location was associated with differences for most indicators, with p values below 0.05. Only a few indicators, including improved concentration, increased mental agility (H = −0.745, p = 0.456), and fatigue relief (H = −1.006, p = 0.314), did not differ significantly by residential location. Education level and employment status were also associated with differences for many indicators, although a small number of indicators showed no significant group differences.

3.2.3. Associations Between User Behaviors and Perceived Health Benefits

Figure 10 summarizes correlations between spatiotemporal use behaviors, facility use frequency, and perceived health benefit indicators.

Visit frequency showed the most consistent pattern, with significant positive correlations across multiple benefit indicators. These correlations were particularly evident for social benefit indicators, including neighborhood relations (Spearman’s ρ = 0.291), parent–child communications (ρ = 0.222), social willingness (ρ = 0.233) and motivation (ρ = 0.169). Compared with visit frequency, timing-related variables showed more selective associations. Weekend visits were positively correlated with pain relief (ρ = 0.158) and neighborhood harmony (ρ = 0.179), whereas correlations for weekday visits were weaker and less consistent. Seasonal patterns were also differentiated. In contrast to other seasons, autumn-related visits showed broader positive correlations spanning psychological outcomes (ρ = 0.182), perceived physiological functioning (ρ = 0.177), and social interaction (ρ = 0.231).

Correlations between frequency of facility use and perceived benefits varied across parks, indicating site specific patterns. In Park H, the frequency of respondents’ use of fitness facilities showed a moderately significant positive correlation with overall perceived physical health (ρ = 0.255), and a certain positive correlation with the indicator of promoting parent–child communication (ρ = 0.209). In Park Y, correlations between facility use and perceived benefits were predominantly negative. The strongest correlation in the results was between the usage frequency of landscape facilities and increased energy (ρ = −0.469), as well as perceived social health (ρ = −0.445). In Park J, children’s play facility use showed comparatively stronger positive correlations with both physiological and social indicators, particularly with neighborhood relations (ρ = 0.626) and parent–child interaction (ρ = 0.709). Leisure facility use in Park J was also positively correlated with multiple indicators, whereas correlations for landscape features and public service facilities were generally limited to a small number of indicators.

Overall, visit frequency emerged as the most robust behavioral correlate of perceived benefits, while the strength and direction of facility benefit correlations depended strongly on the specific park context.

3.3. Mechanisms Underlying Health-Related Benefits of Pocket Parks

3.3.1. Factor Analysis of Influencing Variables

To investigate the mechanisms through which pocket parks contribute to perceived health-related benefits, reliability and validity analyses were first conducted using the 310 valid questionnaires from the main survey. Reliability was assessed using Cronbach’s alpha, one of the most widely used indices [42]. The results showed that Cronbach’s alpha values for all latent constructs exceeded 0.87, indicating good internal consistency of the measurement model.

Validity refers to the extent to which questionnaire items accurately capture the intended evaluation objectives. Higher validity indicates that survey responses better represent the characteristics of the construct being measured. A KMO value greater than 0.6 is considered acceptable for conducting factor analysis [43], and Bartlett’s test should meet the criterion of p < 0.005 [44]. The Kaiser–Meyer–Olkin test and Bartlett’s test of sphericity confirmed that the data met these prerequisites (KMO = 0.909; Bartlett’s test p < 0.001).

Principal component analysis was conducted in SPSS, and three components were retained. To facilitate interpretation, varimax orthogonal rotation was applied, and the rotated loadings are reported in

Table 4. Rotated factor loading matrix (loadings ≥ 0.5 shown).

Indicator	Factor 1	Factor 2	Factor 3	Label
Residential proximity	0.785	-	-	Park supply factor
Accessibility to the park	0.786	-	-
Perceived quantity of facilities	0.558	-	-
Perceived quality of facilities	-	0.880	-	Facility use factor
Facility use frequency	-	0.891	-
Average length of stay	-	-	0.981	Length of stay factor

. Residential proximity, accessibility to the park, and perceived quantity of facilities showed relatively high loadings on Factor 1, indicating that this factor primarily represents these variables. Factor 1 was therefore labeled the park supply factor. Among these, the index of accessibility to the park is assessed based on the travel modes selected by respondents. In this study, walking, cycling, electric mopeds, cars, and public transportation are assigned scores of 5 to 1 respectively. Lower scores intuitively reflect that visitors generally arrive at the pocket park via more convenient modes. Factor 2 was mainly determined by perceived quality of facilities and facility use frequency, and was labeled the facility use factor. Factor 3 was dominated by average length of stay in the park, and was labeled the length of stay factor. All environmental perception variables in this analysis—including residential proximity, accessibility to the park, and perceived quantity of facilities—were derived from participant self-reports.

3.3.2. Empirical Examination of Mechanisms Using Structural Equation Modeling

To further examine the associations and potential pathways linking the three extracted common factors to perceived health-related benefits of pocket parks, structural equation modeling was employed. Structural equation modeling is a multivariate statistical technique that integrates elements of factor analysis and path analysis. By specifying a theoretical model, structural equation modeling allows observed variables and latent constructs to be analyzed within a unified framework. It can estimate direct and indirect effects among variables and assess overall model fit to the data, thereby supporting the evaluation of hypothesized relationships between pocket park-related factors and perceived health benefits [45].

Following the questionnaire design, benefit indicators were grouped into three dimensions, psychological, physiological, and social. It was hypothesized that each of the three common factors affects all three benefit dimensions, and that direct associations exist between each pair of factors. Accordingly, the park supply factor (Factor 1) was hypothesized to directly influence psychological benefits (H1), physiological benefits (H2), and social benefits (H3). The facility use factor (Factor 2) was hypothesized to directly influence psychological benefits (H4), physiological benefits (H5), and social benefits (H6). The length of stay factor (Factor 3) was hypothesized to directly influence psychological benefits (H7), physiological benefits (H8), and social benefits (H9). In addition, direct associations were specified between the park supply factor and the facility use factor (H10), between the park supply factor and the length of stay factor (H11), and between the facility use factor and the length of stay factor (H12).

Based on the hypotheses and the factor labels derived from Table 4, a structural equation model was specified to represent the influencing mechanism of pocket park-related factors on perceived health benefits. Model fit was assessed, and the model was refined using modification indices in AMOS. After adding selected correlated error terms with relatively large modification indices, model fit improved and met commonly used thresholds [46,47,48] (

Table 5. Model fit indices before and after model modification.

Index	Common Threshold	Before Modification	After Modification
PCMIN/DF	≤3	4.168	2.920
GFI	>0.85	0.829	0.892
NFI	>0.85	0.829	0.887
CFI	>0.90	0.863	0.922
RMSEA	<0.08	0.101	0.079

), suggesting that the final model adequately represented the data structure (Figure 11). Estimated path coefficients are reported in

Table 6. Model parameter tests and hypothesis verification.

Hypothesis	Path	Estimate	S.E.	C.R.	p-Value	Result
H1	Park supply → Psychological health	11.229	3.711	3.026	0.002	Supported
H2	Park supply → Physical health	10.834	3.589	3.019	0.003	Supported
H3	Park supply → Social health	11.116	3.697	3.007	0.003	Supported
H4	Facility use → Psychological health	−0.246	0.283	−0.868	0.386	Not supported
H5	Facility use → Physical health	−0.216	0.253	−0.853	0.394	Not supported
H6	Facility use → Social health	−0.311	0.349	−0.890	0.373	Not supported
H7	Length of stay → Psychological health	0.317	0.216	1.471	0.141	Not supported
H8	Length of stay → Physical health	0.268	0.208	1.288	0.198	Not supported
H9	Length of stay → Social health	0.382	0.215	1.779	0.075	Not Supported
H10	Park supply ↔ Facility use	0.025	0.010	2.523	0.012	Supported
H11	Park supply ↔ Length of stay	−0.015	0.014	−1.092	0.275	Not supported
H12	Facility use ↔ Length of stay	−0.020	0.024	−0.813	0.416	Not supported

.

3.3.3. Effect Analysis

Based on the structural equation modeling results, psychological health, physical health, and social health were specified as dependent variables, while park supply, facility use, and length of stay were specified as independent variables. An effect analysis was conducted to further clarify the relationships among variables and to quantify the magnitude of the estimated effects.

Using park supply as the predictor, the estimated direct effects were: psychological health (11.229), physical health (10.834), and social health (11.116). Using facility use as the predictor, the estimated direct effects were: psychological health (−0.246), physical health (−0.216), and social health (−0.311). Using length of stay as the predictor, the estimated direct effects were: psychological health (0.317), physical health (0.268), and social health (0.382). Overall, park supply showed significant effects on perceived health benefits across all three dimensions, while the direct effects of facility use and length of stay on health benefits were not supported. Among the three predictors, park supply exhibited by far the largest effect magnitude, indicating that it is the most central and direct driver in the model.

The park supply factor was constructed from residential proximity, accessibility to the park, and perceived quantity of facilities. In practical terms, when a pocket park is located closer to residential areas, is easier to access, and provides a more sufficient baseline supply of facilities, users are more likely to report broader benefits across psychological, physiological, and social dimensions. Park supply was also positively associated with the facility use factor, indicating that a convenient park with adequate provision tends to encourage more frequent and higher quality facility use.

In contrast, the facility use factor, captured by perceived quality of facility and use frequency, showed no significant direct effects on psychological, physiological, or social benefits. This suggests that facilities may not generate perceived health gains through a simple direct pathway. Instead, facility use is more likely to operate as a contextual enabler, with its contribution expressed indirectly by supporting meaningful activities within a park that is accessible and adequately supplied. When basic provision is limited or access is inconvenient, improvements in facility quality alone may be insufficient to translate into measurable benefits.

Similarly, the direct effects of length of stay on perceived health benefits were not statistically significant, indicating that spending more time in the park, by itself, is not a sufficient driver of health-related gains. The value of longer stays may lie in the opportunities they create for higher-quality experiences, particularly the type and quality of activities undertaken during the visit. Without engaging activities or restorative experiences, the potential influence of duration is likely to be attenuated.

4. Discussion

4.1. Policy Implications and Practical Relevance

Under the combined momentum of urban regeneration and the Healthy China agenda, a national strategy launched in 2016 that integrates health into all policies with strong implications for urban planning and public space design, pocket parks have increasingly been positioned as a key micro-scale intervention in urban governance. Recent policy emphasis has shifted from merely increasing green space quantity to improving usability and health relevance, aiming to address gaps in central districts through infill greening while responding to everyday needs for recreation and health-supportive activity. In practice, however, pocket park projects may still be misaligned with residents’ needs, facility provision can prioritize form over function, and health benefits may remain limited. This underscores the need for evidence-based guidance that links planning and design decisions to observable use patterns and perceived outcomes.

Findings from this research highlight a consistent pathway in which proximity, accessibility, and adequate baseline provision drive use and support perceived benefits. From a planning perspective, the first priority is to strengthen accessibility and convenience. Strategic site selection should ensure that service catchments effectively cover high-density residential areas, while walking and cycling networks and public transport connections reduce barriers to reaching the park. Within pocket parks, sufficient basic service facilities, such as seating, lighting, and fitness equipment, help meet common needs and provide a material foundation for benefit delivery.

Design strategies should also support social use. Although the structural model did not confirm a significant direct effect of length of stay on health outcomes, longer visits can plausibly increase opportunities for interaction and community connection, and the broader results indicate that social benefits are closely tied to how spaces enable everyday encounters. Spatial configurations that allow both casual contact and comfortable co-presence can facilitate interaction without requiring prolonged or intensive use. Diverse resting niches, small open lawns or plazas, and community garden elements can accommodate different groups and activity intensities. Thermal comfort and perceived safety remain essential enabling conditions, and shade provision through planting, together with cleanliness and routine maintenance, can increase willingness to stay and interact. Where feasible, light-touch community programming may further strengthen social engagement and reinforce local attachment.

Facility investment should follow a rational and cost-effective approach. Rather than pursuing novelty or high-specification installations, practice should prioritize durability, safety, and maintainability. Results suggest that the overall convenience of reaching the park and the sufficiency of basic provision may matter more than upgrading a single facility type, because facility value is more likely to be realized through enabling activities within an accessible setting. In other words, a successful pocket park is primarily one that is easy to enter and comfortable to use. Premium facilities alone cannot compensate for poor accessibility or insufficient baseline provision.

In practical terms, these insights offer a clearer action framework for practitioners by prioritizing site selection, accessibility improvements, and basic provision, reducing the risk of resource misallocation. By aligning interventions with actual use needs and health expectations across groups, pocket parks can improve utilization and user satisfaction and contribute more effectively to healthy urban living in high-density settings.

These finding align with recent experimental evidence from a European context. Using psychometric tests to measure immediate psychological responses, Kobyłka and Korcz (2025) found that short walks in a Polish city park during autumn significantly improved young adults’ mood, vitality, and restoration, while walking on city streets produced negative effects [49]. Their objective measurements complement our subjective, perception-based findings in several ways. Both studies identify autumn as a season conducive to park-related health benefits—theirs through measured psychological improvements, ours through perceived benefits such as emotional relief. Additionally, their observation that even park edges contribute to restoration resonates with our finding that basic accessibility and facility provision, rather than premium amenities, drive perceived benefits. While their experimental approach captures immediate, short-term responses to controlled walks, our study reflects routine, everyday use patterns and accumulated perceptions over time. This cross-methodological comparison—contrasting objective experimental measures with subjective perceived benefits—strengthens the international relevance of our findings and suggests that the health-supportive functions of urban green spaces operate consistently across different cultural contexts and research approaches.

4.2. Limitations

This research has several limitations. First, the cases were limited to three pocket parks in a single city and the sample composition was not fully balanced, which may constrain generalizability across different urban and socioeconomic contexts. Second, health-related benefits were assessed using self-reported perceptions. Such measures may be affected by reporting and recall biases, and objective health outcomes were not included. In addition, key explanatory variables, including residential proximity, access convenience, and facility provision, were also derived from self-reports and represent perceived rather than objectively measured conditions, which may introduce measurement error and common method bias. Consequently, the findings of this study can only confirm the hypothesis that participants perceive parks as beneficial to their health, rather than demonstrate any objective impact of parks on individual health outcomes. Third, environmental and behavioral variables were measured at a relatively coarse level. Detailed environmental features and objective behavioral tracking were not incorporated, which may limit the precision of spatiotemporal behavior characterization. Finally, the cross-sectional design and the modeling strategy support association-based inference, and causal interpretation should be made cautiously.

Future work could expand the study scope to more cities and park types with more diverse samples, integrate subjective assessments with objective health and environmental measurements, and combine fine-grained environmental indicators with behavioral tracking to better identify activity-specific pathways linking pocket parks to health outcomes.

5. Conclusions

The central finding is that park supply is the primary and most robust driver of perceived health-related benefits in pocket parks. In the structural model, the park supply factor showed direct effects on psychological, physiological, and social benefits, whereas facility use showed no significant direct effects, and length of stay showed a positive but non-significant tendency toward social benefits. This pattern suggests that, in high-density urban settings, perceived benefits depend first on whether a pocket park is easy to reach and provides sufficient baseline provision, rather than on premium facilities or facility use alone.

Evidence from three representative pocket parks in Nanjing further indicates that pocket parks are strongly embedded in everyday life. Users tended to visit frequently, stay for relatively short periods, and display regular temporal rhythms, with most visits coming from nearby residents within walking distance. Users also showed distinct timing-related patterns, and these were selectively associated with perceived benefits: weekend visits were more strongly associated with pain relief and neighborhood harmony, while autumn visits showed broader positive associations across fatigue alleviation, social engagement, and most psychological indicators. Pocket parks therefore function less as occasional leisure destinations and more as convenient micro-spaces integrated into daily commuting, resting, and social interaction, creating conditions for perceived benefits to accumulate over time.

Perceived benefits were primarily restorative and everyday in nature, with stronger effects reported for emotional relief and fatigue recovery than for outcomes related to chronic condition management or intensive rehabilitation. Associations between behaviors, facilities, and benefits were selective and context dependent. Across parks, facility types showed a consistent pattern of benefit specificity: facilities were associated with perceived benefits mainly by shaping what people do in the park and with whom they interact, rather than producing uniform gains across all health domains. Facility–benefit links nevertheless varied by park and facility type. In Park H, landscape facilities were more strongly positively correlated with physical health, whereas in Park J, children’s play facilities aligned more strongly with social benefits, particularly neighborhood relations and parent–child interaction. In Park Y, facility measures tended to show weaker or even negative associations, reinforcing that facility value depends on context and user–park fit. Together with the structural model, these patterns suggest that facilities may function mainly as enablers whose contribution is expressed indirectly through supporting activity and interaction within an accessible and adequately supplied park environment. Overall, the findings imply that planning, design, and management should prioritize accessibility and adequate baseline provision, while adopting behavior-oriented strategies that account for when and how parks are used to match pocket park functions to neighborhood needs, thereby strengthening pocket parks as practical health assets in high-density cities.