Well-Being and Influencing Factors in Urban Ecological Recreation Spaces: A Human–Nature Interaction Perspective

Abstract

Urban ecological recreational space (UERS), as essential provider of ecosystem services, play a crucial role in enhancing human well-being. Nevertheless, limited research has investigated how various types of human–nature interaction influence well-being. This study takes the Hankou waterfront as the research area, using a questionnaire survey of 318 recreationists. A five-dimension well-being scale and interaction type classification were developed. Mean analysis, one-way ANOVA, and stepwise multiple regression were conducted to assess well-being and influencing factors. Results show that mental health had the highest score, while survival security scored the lowest. NDVI positively influenced all well-being dimensions, while fitness facilities and higher education levels showed negative effects. Recreationists engaged in outdoor work-oriented activities experienced higher levels of physical, mental, and self-actualization well-being than sightseeing- or socially oriented users. These findings expand the theoretical understanding of UERS by incorporating human–nature interaction patterns and offer practical guidance for sustainable urban planning.

1. Introduction

The rapid advancement of global urbanization has led to increasingly diverse demands from urban residents for an improved quality of life, emphasizing not only material needs but also heightened expectations for spiritual and cultural fulfillment, social interaction, and self-actualization. In this context, Urban Ecological Recreational Space (UERS), as a crucial component of urban green infrastructure [1], offers various ecosystem services, is essential for facilitating residents’ physical exercise and leisure activities, enhances social interaction, alleviates psychological stress [2], and has increasingly become pivotal in optimizing the living environment and enhancing residents’ well-being [3,4,5]. In recent years, the development and establishment of urban wetland parks, community green spaces, and other UERS have been markedly enhanced, demonstrating considerable improvements in both scale and quality. Do these places genuinely correspond to the diversified requirements of recreationists? What are the consequences on well-being, and what factors influence them? These inquiries merit profound contemplation.

General findings underline that contact with UERS can have various positive impacts on well-being. First of all, UERS play a pivotal role in enhancing survival security. As a core component of urban green infrastructure and natural solutions, UERS possess multiple ecological functions, such as moderating ambient temperature [6], mitigating air pollution [7], enhance air quality and living conditions, improve urban livability, and provide essential support for human survival security. Secondly, existing studies have consistently demonstrated the positive effects of UERS on mental and physical health. Engagement with nature in UERS has been shown to decrease blood pressure, mitigate obesity, alleviate anxiety, and enhance overall life satisfaction [8,9]. UERS not only provides recreationists with places for physical exercise, but also creates conditions for their psychological relaxation and emotional regulation, satisfying the dual needs of recreationists at both physical and psychological levels [10]. In addition to their impact on physical and mental health, UERS also significantly contribute to promoting social interaction and strengthening community connections. For example, the aesthetic experience, awareness of cultural heritage value, and building of a feeling of place in UERS might augment recreationists’ sense of belonging, cultural identity, and social connection [11]. Moreover, UERS’s amenities function as sites for family gatherings and community events, promoting interpersonal connections and neighborhood solidarity [12,13]. Finally, UERS also facilitate the enhancement of self-actualization. Educational experiences and nature-based learning in UERS, including exposure to biodiversity and natural aesthetics, enable individuals to comprehend life’s harmony and significance, broaden cognitive limits, and promote human growth and development [14,15].

Previous studies have demonstrated the association between UERS and well-being, providing a solid foundation for this study, and they indicate that UERS contribute to human well-being through the provision of ecosystem services [16,17,18,19]; cascade model serves as a key theoretical framework for exploring the relationship between ecosystem services and human well-being, revealing the influence path from ecosystem to human well-being through the “structure–function–value” chain [20]. This approach characterizes the services and benefits offered by the ecosystem as inherent results of biophysical structures and processes, hence neglecting human agency. In contrast, Costanza proposed the notion of co-production of ecosystem services, emphasizing the significance of human agency [21], positing that sustainable human well-being arises from the interplay of social capital, built capital, human capital, and natural capital (ecosystem services). While this framework reintroduces human agency into ecosystem service research, it overlooks how different types of human–nature interaction may produce varying well-being outcomes, especially in recreational contexts like UERS. In addition to the impact of human–nature interaction types, the well-being in UERS is also affected by spatial structure and function [12,13], the degree of spatial exposure and measurement of recreational users [14,15], and the demographic attributes of recreationists (e.g., age, occupation, income, and education) [16,17,18]. The interplay of these factors influences the efficacy in addressing the varied requirements of recreationists in UERS. Nevertheless, the majority of the existing literature concentrates on either the influence of objective physical attributes of UERS [22,23,24,25] or the effect of subjective perceived characteristics [26]. Few studies examine the interplay between these factors regarding human well-being, thereby constraining a holistic understanding of the determinants influencing the human well-being in UERS, which constrains its scientific utility in practical planning and management.

Building on this theoretical foundation, this study aims to bridge these gaps by analyzing how different forms of human–nature interaction, social capital, built capital, human capital, and natural capital collectively influence recreationists’ well-being in UERS. Specifically, it aims to address the following scientific questions: (1) Evaluate the overall well-being and its multifaceted dimensions among recreationists at Hankou waterfront. (2) Analyze the well-being differences of recreationists with different types of human–nature interaction and the influencing factors of various types of well-being in Hankou waterfront. This study holds significant theoretical and practical implications for maximizing the ecological benefits of UERS and promoting the high-quality development of UERS.

2. Research Hypotheses

Ecosystem services serve as the critical bridge connecting UERS ecosystems with human well-being. Haines proposed the ecosystem services cascade framework [20], which emphasizes ecosystem services as the foundation supporting human well-being, and recognizes human well-being as both the ultimate goal and core component of sustainability [27]. The cascade framework effectively elucidates the interactive stress effects between ecosystem services and well-being through its “impact chain”, successfully addressing the transmission process from ecosystem services to human well-being. However, this framework overlooks human agency [28]. In response to this limitation, Costanza introduced a capital co-production framework, which emphasizes that natural capital must be activated through its interactions with human, social, and built capital to produce sustainable well-being [21,29] (Figure 1, left). Liu also highlights that ecological services and human well-being are varied, multilayered, and scale-dependent, characterized by intricate interactions [30]. While Costanza et al. have demonstrated the necessity of interactions among four capital types (natural, human, social, and built) for well-being generation [29], their framework contains a critical gap, which assumes that “capital interactions automatically produce well-being” without elucidating whether differentiated human–nature interactions (in both type and degree) might lead to well-being heterogeneity under equivalent capital combinations (Figure 1, right) [11,31].

Building upon these frameworks, this study therefore integrates human–nature interaction typologies into the ”ecosystem services and human wellbeing” analytical framework as an additional dimension under given capital combinations. Specifically, we hypothesize that even under identical capital configurations (e.g., same green infrastructure and access conditions), different types of human–nature interactions, such as passive appreciation vs. active ecological engagement, will lead to qualitatively distinct well-being. We transcend the traditional supply-side focus in ecosystem services research, effectively upgrading Costanza’s static capital interaction framework into a dynamic “capital interaction–-behavior choice–well-being” model, which introduces agency and behavioral heterogeneity into the “Ecosystem services–well-being” pathway, thereby offering a more robust framework for analyzing how UERS function not merely as service providers, but also as interactional arenas where individual behavior shapes the translation of ecosystem potential into realized well-being. This extension enables a better understanding of spatial and individual variation in well-being, and provides a behavioral foundation for future planning and policy design in urban ecological systems.

Consequently, drawing from Costanza’s research, we specifically analyze the possible effects of natural capital, human capital, social capital, built capital, and human–nature interaction types on human well-being. Each capital dimension operates based on empirical relevance, theoretical rationality, and data availability. While our classification is largely consistent with the original capital framework, we have adopted specific situational agents commonly used in urban behavior and environmental research, to reflect the multifaceted nature of well-being determinants.

2.1. Natural Capital

The significant influence of natural capital on human well-being has been empirically substantiated by various researchers [3,5,15]. In UERS, high vegetation coverage significantly improves air quality and regulates microclimates, thereby providing recreationists with a safer living environment. Moreover, exposure to natural environments reduces negative emotions such as stress and anxiety and enhances physical health by lowering blood pressure and decreasing obesity risks. Natural capital within UERS provides recreational spaces and fosters social interaction and self-actualization through aesthetic experiences and ecological functions. Consequently, vegetation quality in UERS is crucial for recreationists’ multidimensional well-being. Therefore, we hypothesize the following:

H1. 

NDVI have positive effects on recreationists’ well-being.

2.2. Human Capital

Human capital refers to individuals’ cognitive, affective, and motivational attributes related to nature, primarily their environmental knowledge, ecological awareness, and nature preferences. In this study, we also incorporate spatial exposure, which is defined as the frequency and duration of individuals’ engagement with UERS as an observable proxy of environmental affinity, reflecting the activation and application of human capital in recreational contexts.

Specifically, nature preference indirectly enhances survival security by raising ecological risk awareness and promoting sustainable lifestyles [24]. Stronger nature preference correlates with heightened perception of aesthetic, educational, and spiritual. Individuals with higher cognitive abilities can transform nature experiences into ecological identity and responsibility, triggering positive emotions, reducing stress, and directly improving mental health and self-actualization [30,32]. Furthermore, a high preference for nature promotes common interests and social discussions, while sustained engagement in UERS enhances community belonging, reduces social isolation and loneliness, and improves overall social interaction well-being.

H2. 

Nature preference has positive effects on recreationists’ well-being.

Spatial exposure refers to recreationists’ activities in UERS, including both exposure intensity (i.e., activity frequency) and exposure dosage (i.e., activity duration), which are key factors influencing recreationists’ well-being [33,34]. Studies show that frequent visitors to UERS report higher self-rated perceptions of survival security compared to infrequent visitors. White [35] found that increased frequency of visits to natural environments is linked to greater happiness. White [36] further demonstrated that spending at least 120 cumulative minutes in nature over 7 days is associated with higher levels of well-being. Therefore, we hypothesize the following:

H3. 

Activity frequency has positive effects on recreationists’ well-being.

H4. 

Activity duration has positive effects on recreationists’ well-being.

2.3. Social Capital

Social capital refers to the familiarity of individuals with the community and the basic form of the social support network, and it influences individual well-being in UERS through community familiarity and community support. First, social support (measured by whether individuals are accompanied, i.e., companions) captures the immediate companionship and emotional support obtained during activities. Companions provide support and assistance in emergencies, potentially increasing survival security. Their presence also encourages more physical activities like walking or exercising, thereby improving physical health. Peer interactions foster a sense of belonging, reduce loneliness, and offer emotional support during group activities, which helps relieve stress and enhances both mental health and social interaction. Furthermore, the sharing of knowledge and skills among peers, specifically in ecological contexts, enhances self-efficacy. Therefore, we hypothesize the following:

H5. 

Having companions has positive effects on recreationists’ well-being.

Secondly, community familiarity (measured by length of residence) reflects the level of understanding of community public resources, geographical environment, and interpersonal networks. Long-term residents possess a familiarity with their environment that enhances their ability to manage potential environmental risks, resulting in an increased perception of survival security. Furthermore, an extended duration of residence correlates with a more intimate social network, which enhances mental health and social connections through frequent interactions. The sense of place and cultural identity cultivated through prolonged residence significantly contributes to self-actualization and overall well-being. Consequently, we hypothesize the following:

H6. 

Length of residence has positive effects on recreationists’ well-being.

2.4. Built Capital

Built capital influences leisure experience through infrastructure adequacy (measured by infrastructure provision) and facility diversity (measured by fitness facilities and recreational amenities provision). These elements constitute the physical environment that enables well-being-oriented behaviors and experiences, shapes spatial practices, and facilitates the multidimensional realization of human well-being. Studies have found that the quantity of recreational facilities and fitness equipment in UERS is likely to extend the duration in UERS [37] and is significantly associated with higher levels of walking or physical activity, thereby indirectly affecting perceived survival security and physical health [38]. Recent evidence shows that built environments have pronounced mental health benefits for high-frequency users, such as university students, particularly in post-pandemic recovery contexts [39]. Moreover, sufficient seating, properly maintained lighting, clear signage, and environments free from graffiti and litter enhance perceptions of safety and convenience in UERS. Enhanced environmental conditions promote more efficient stress relief and attention restoration [25]. However, facilities, if inadequately maintained or of substandard quality, can result in adverse outcomes. Overall, built capital positively influences human well-being. Consequently, we hypothesize the following:

H7. 

Infrastructure has positive effects on recreationists’ well-being.

H8. 

Fitness facilities have positive effects on recreationists’ well-being.

H9. 

Recreational amenities have positive effects on recreationists’ well-being.

2.5. Human–Nature Interaction Classification and Its Effects on Well-Being

Within the same ecosystem, different recreational activities undertaken by visitors may lead to varying services and benefits obtained. In other words, human well-being perceptions in UERS differ according to the type of interaction with nature [40]. Huynh [11] suggests that the connection between cultural ecosystem services (CES) and human well-being can be observed through commonalities in terms of (i) the ways people consciously and unconsciously engage with ecosystems and experience benefits (i.e., channels of interaction) and (ii) the processes through which these interactions contribute to human well-being (i.e., mechanisms). He also identifies five assemblages that contain interacting pathways and relate to (i) sensory affection, (ii) learning and development, (iii) health and leisure fulfillment, (iv) social vibrancy, and (v) spiritual and heritage resources, which provide a crucial theoretical foundation for deconstructing human–nature interaction typologies. Li categorized the functions of urban green spaces into four types: socially-oriented, outdoor work-oriented, culturally-oriented, and multifunctional [41]. Nath et al. noted that urban green spaces serve as spatial carriers of human–nature interactions, systematically facilitating a variety of activities including sightseeing, social interaction, and physical exercise [42].

Building on Huynh’s conceptual framework of CES, this study conceptualizes human–nature interaction as a multidimensional construct encompassing both perceptual channels and behavioral mechanisms through which individuals engage with nature. Theoretical foundations are drawn from Huynh’s five interaction pathways—sensory affection, cognitive development, physical and leisure fulfillment, social connection, and spiritual engagement [11]—which capture the experiential dimensions of nature contact. Supported by field observations from the Hankou waterfront, we developed a typology of three clustered interaction types that reflect increasing levels of physical involvement and cognitive engagement:

(1)

Sightseeing-oriented, low human–nature interaction engagement, dominated by sensory appreciation and short-term emotional benefits;

(2)

Social activities-oriented, emphasizing social connections with moderate interaction depth, collective use, and emotional–social well-being;

(3)

Outdoor work-oriented, high human–nature interaction engagement, integrating nature-based learning, learning and development, health and leisure fulfillment, and spiritual/heritage connection activities to foster profound human–environment bonds through the deepest level of interaction.

These clusters are derived by mapping the five interaction pathways into broader activity-based categories (see

Table 1. Core interaction pathways and their associated human–nature interaction types.

Core Interaction Pathway	Description	Human–Nature Interaction Type
Sensory affection	Passive observation, aesthetic enjoyment	Sightseeing-oriented
Social vibrancy	Interaction with others, group leisure	Social activities-oriented
Learning and development	Environmental knowledge acquisition	Outdoor work-oriented
Health and leisure fulfillment	Physical activity, stress reduction	Outdoor work-oriented
Spiritual/heritage connection	Cultural identity, place attachment	Outdoor work-oriented

), reflecting differences in interaction intensity, intentionality, and experiential depth. This classification enables a more nuanced understanding of how recreationists translate nature contact into well-being via differentiated interaction type. Specifically, sightseeing-oriented interaction tends to support short-term emotional relief; social activities-oriented interaction enhances social cohesion and belonging; and outdoor work-oriented interaction are linked to longer-term well-being such as identity formation, self-efficacy, and eudaimonic well-being. This gradient of engagement intensity aligns with findings by Ward [43] and Bratman et al. [40], who found that deep interactions between humans and nature enhance self-efficacy, resulting in more enduring benefits for well-being. We propose the following hypothesis:

H10. 

The types of human–nature interaction influence well-being, with well-being perceived in outdoor work-oriented interactions being higher than in sightseeing-oriented and social activities-oriented interactions.

In summary, we construct an influential factor model of well-being in UERS (Figure 2). The CES in UERS are co-created by ecological systems and human social systems. By perceiving, utilizing, and experiencing CES in UERS, individuals are influenced by natural capital, human capital, social capital, and built capital, thereby obtaining well-being that includes survival security, mental health, physical health, social interaction, and self-actualization, either directly or indirectly. This well-being may vary depending on the type and intensity of human–nature interactions. This mechanism shows a significant advancement from conventional CES research, which primarily concentrates on the supply-side of services, by integrating “human agency” into the analytical framework for assessing ecosystem services and human well-being.

3. Materials and Methods

3.1. Study Area

Urban waterfront is a critical component of UERS. Since the 1950s, a wave of revitalization and reclamation of urban waterfront has emerged in developed countries, resulting in the gradual transformation of waterfront from productive spaces to recreational, consumer, and cultural spaces [44]. As the core city of the Yangtze River Economic Belt, Wuhan possesses abundant water resources and extensive waterfront. Through years of revitalization and redevelopment, Wuhan’s waterfront has achieved significant accomplishments in urban development, ecological conservation, and cultural construction, becoming an important destination for urban tourism and public recreation. This study selects Hankou waterfront as the study area, which serves as a typical representative of Wuhan’s waterfront. As the earliest developed, largest in scale, most comprehensively equipped, and most vibrant UERS in Wuhan, Hankou waterfront integrates natural ecology, historical culture, and modern recreational functions, making it the preferred destination for both residents and tourists seeking leisure, exercise, cultural activities, and sightseeing.

Moreover, Hankou waterfront may reflect broader patterns found in urban green and blue spaces. Although urban nature takes diverse forms, such as community gardens (promoting food security, social cohesion, and mental well-being) [45], rooftop and vertical greenery (linked to climate regulation and aesthetic perception) [46], and blue spaces (associated with psychological restoration and intergroup connection) [47], these environments similarly facilitate human well-being through CES-driven pathways. Such similarities may inform broader frameworks for UERS planning and assessment across different urban contexts. While structural-functional similarities, cultural and regional contexts may shape how users perceive and value CES in UERS. Studies have shown that aesthetic and recreational preferences differ between cultural groups, for example, East Asian users often emphasize harmony and symbolic meaning, whereas Western users tend to value individualistic recreation and visual openness [16,48]. However, such cultural differences are more likely to influence the strength of preferences for specific functions or activity types within a given CES pathway, rather than altering the underlying mechanisms through which CES contribute to human well-being. Therefore, selecting Hankou waterfront as the research site demonstrates both representativeness and scientific validity (Figure 3).

3.2. Survey Design

This research gathered primary data via field investigations and questionnaire surveys. The questionnaire comprised three sections: demographic characteristics, well-being, and human–nature interaction types. The demographic section encompassed gender, age, marital status, education level, average monthly income, length of residence in Wuhan, companionship status, and nature preference. In this study, we adopted a broad definition of recreationists, including all individuals engaging with UERS. Participants were not explicitly classified as residents or tourists, as we focused on actual user experience and perceived well-being. However, recognizing potential perceptual differences between tourists and residents, we introduced an additional option in the questionnaire to address this. Specifically, in the well-being section, respondents could select “did not bring me this kind of benefit,” which was assigned a value of 0, ensuring more accurate reflection of individual experiences and minimizing response bias among diverse user groups. The well-being section was developed in accordance with the well-being elements specified in the Millennium Ecosystem Assessment [49] and the methodology suggested by Jones et al. [50], creating a multidimensional framework for measuring well-being that encompasses five dimensions and thirteen indicators (see

Table 2. Measurement scale for well-being in Hankou waterfront.

Well-Being	Code	Item
Survival security (SG)	SG1	Offers fresh air, open spaces, and a healthy ecological environment.
	SG2	Contributes to the reduction of urban crime.
	SG3	Provides resources related to work, economy, and leisure, among other topics.
Mental health (MH)	MH1	Facilitates the alleviation of negative emotions, promotes mood relaxation, and enhances attention restoration.
	MH2	Improves life satisfaction.
Physical health (PH)	PH1	Offers health advantages, including the reduction of blood pressure and obesity.
	PH2	Enhances immune system.
Social interaction (SI)	SI1	Strengthened bonds with family and friends, or greater opportunities for social interactions.
	SI2	Enhances comprehension of community dynamics and fosters engagement in local activities.
	SI3	Increases attachment to Wuhan and the sense of belonging.
Self-actualization (SA)	SA1	Enriches personal experiences and enhances aesthetic Appreciation.
	SA2	Enhances cognitive capacity, linguistic articulation, and interpersonal communication competencies
	SA3	Facilitates personal development and elucidates the significance of life.

). A five-point Likert scale was used to quantified responses, ranging from “very little impact” to “very large impact”, with scores assigned from 1 to 5.

Among the various ecosystem services provided by UERS, recreationists, government managers, and scholars typically prioritize CES [18], which not only attract recreationists for leisure activities but also serve as a crucial strategy for administrators to foster livable environments and encourage sustainable urban renewal [6]. Therefore, based on Maslow’s hierarchy of needs, we assert that recreationists obtain perceptions of survival security, mental health, physical health, social interaction, and self-actualization by engaging with different forms of CES. It is worth noting that the dimensions of social interaction and self-actualization are relatively subjective. We aimed to ensure the content validity of these dimensions through careful scale design and to capture their multidimensionality as much as possible during measurement. For social interaction (SI), we adopted three indicators (see Table 2, SI1~SI3) that reflect the well-being associated with social interaction from multiple perspectives, including interpersonal relationships, community engagement, and emotional belonging, highlighting the influence of social networks and emotional support on individual well-being. For self-actualization (SA), we selected the indicators (see Table 2, SA1~SA3) that reflect the core characteristics of self-actualization by covering cognitive, emotional, and personal growth dimensions.

The survey on types of human–nature interaction was meticulously developed by incorporating feedback from preliminary visitor surveys and building upon the theoretical framework proposed by Huynh et al. [11]. The questionnaire was organized into five interaction pathways: sensory affection, health and leisure fulfillment, learning and development, social vibrancy, and spiritual and heritage resources (

Table 3. Survey of human–nature interaction types.

Interacting Pathways	Specific Activities	Item
Sensory affection	sightseeing, photography, flower appreciation, strolling, etc.	1 = Yes; 0 = No
Health and leisure fulfillment	using fitness equipment, practicing Tai Chi or martial arts, flying kites, playing ball games, skateboarding, etc.	1 = Yes; 0 = No
Learning and development	walking pets, reading or writing, camping, fishing, singing, playing musical instruments, dancing, etc.	1 = Yes; 0 = No
Social vibrancy	parenting activities, dating, socializing, etc.	1 = Yes; 0 = No
Spiritual and heritage resources	reading or writing, visiting historical sites, learning about local culture, etc.	1 = Yes; 0 = No

). A binary coding scheme was utilized, with participation in a designated activity at Hankou waterfront represented as “1” and non-participation as “0”. In field surveys, each activity within these interaction pathways was given as a distinct question item to guarantee data accuracy and operational viability. For example, within the sensory affection pathway, various activities such as ’sightseeing’, ’photography’, ’flower appreciation’, and ’strolling’ were individually examined to thoroughly understand recreationalists’ engagement patterns with the waterfront.

3.3. Data Collection

To assess the well-being of recreationalists at Hankou waterfront, this study employed questionnaire surveys as the primary method for data collection. Specifically, multiple field surveys were conducted during November–December 2022, followed by a three-day preliminary survey on 8–10 January 2023. Based on feedback from recreationalists, the questionnaire was refined before the formal survey was implemented on 25–26 February 2023. The poll produced 331 completed questionnaires, yielding 318 valid replies after the exclusion of invalid inputs, resulting in a high validity percentage of 96.07%. The sample demonstrated an equitable gender distribution (52.5% female, 47.5% male) and was primarily comprised of long-term inhabitants of Wuhan (34.9% having resided there for over 20 years). Significantly, 82.1% of participants attended with companions, while the duration of activities predominantly ranged from 1 to 3 h (53.8%). The frequency of visits exhibited a bi-modal distribution, with the highest occurrence at 1–2 weekly visits, followed by less than 1–2 monthly visits as the second most prevalent category (see

Table A1. Demographic profile of Hankou waterfront recreationalists (N = 318).

Category	Item	Frequency	Percentage (%)	Category	Item	Frequency	Percentage (%)
Gender	Female	167	52.5	Length of residence	Below 1 year	15	4.7
	Male	151	47.5		1–5 years	38	11.9
Age (years old)	<20	41	12.9		6–10 years	87	27.4
	21–30	95	29.9		11–20 years	67	21.1
	31–40	80	25.2		Above 20 years	111	34.9
	41–50	25	7.9	Marital status	Unmarried	109	34.3
	51–60	40	12.6		Married	198	62.3
	≥61	37	11.5		Other	11	3.4
Education	Primary education or below	15	4.7	Companions status	Accompanied	261	82.1
	Junior secondary education	40	12.6		Unaccompanied	57	17.9
	Secondary vocational/ General high school	67	21.1	Frequency	Daily	58	18.2
	Associate/bachelor’s degree	138	43.4		5–6 times/week	14	4.4
	Master’s degree and above	58	18.2		3–4 times/week	22	6.9
Average monthly income (yuan)	Below 1000	60	18.9		1–2 times/week	85	26.7
	1001–2500	12	3.8		1–2 times/ month	63	19.9
	2501–5000	83	26.1		Rarely	76	23.9
	5001–10,000	79	24.8	Single duration	<60 min	70	22
	10,001–15,000	60	18.9		60–180 min	171	53.8
	Above 15,000	24	7.5		>180 min	77	24.2

, Appendix A).

The determinants affecting well-being were derived from questionnaires, field surveys, and remote sensing. As derived from the theoretical analysis in Section 2, human well-being in UERS is affected by natural capital, human capital, social capital, and built capital [51]. In this study, human capital was characterized by nature preference, activity frequency, and activity duration, collected via questionnaires. Social capital was measured by companion status (with/without) and length of residence, obtained through surveys. Built capital was assessed using three indicators—infrastructure, fitness facilities, and recreational amenities—based on field investigations. For systematic quantification, the waterfront was divided into three zones prior to field surveys, which are Zone A (Recreational boulevard, A1~13), Zone B (Landscape Trail, B1~B5) and Zone C (Ecological berm, C1~C7). Each zone was further subdivided into 25 sub-areas (see Figure A1, Appendix B). Facility scores were assigned and summed according to their spatial distribution. Natural capital was evaluated by mean NDVI, extracted from Sentinel-2 remote sensing imagery. Specifically, after resampling using ESA SNAP 10.0.0, NDVI was calculated for the entire study area using the Band Math tool in ENVI 5.3. Based on the activity areas of individual samples, the mean NDVI values were extracted using the Zonal Statistics tool in ArcGIS 10.8. Control variables included gender, age, education level, and average monthly income (see

Table 4. Specific indicators of influencing factors of well-being effect.

	Item
Control variables
Age (years old)	<20 = 1; 21–30 = 2; 31–40 = 3; 41–50 = 4; 51–60 = 6; ≥61 = 7
Education level	Primary education or below = 1; junior secondary education = 2; secondary vocational/general high school = 3; associate/bachelor’s degree = 4; master’s degree and above = 5
Average Monthly Income (yuan)	Below 1000 = 1; 1001–2500 = 2; 2501–5000 = 3; 5001–10,000 = 4; 10,001–15,000 = 5; above 15,000 = 6
Gender	Male =1; female = 0
Human Capital
Activity frequency	Rarely = 1; 1–2 times/ month= 2; 1–2 times/week = 3; 3–4 times/week=4; 5–6 times/week = 5; daily = 6
Activity duration	<60 min = 1; 60–180 min = 2; >180 min = 3
Nature preference	Strongly dislike = 1; moderately dislike = 2; neutral = 3; moderately like = 4; strongly like = 5
Social Capital
Length of Residence	Below 1 year = 1; 1–5 years = 2; 6–10 years = 3; 11–20 years = 4; above 20 years = 5
companion status	Accompanied = 1; Unaccompanied = 0
Built capital
Infrastructure	Bench, picnic table, gazebo, covered walkway, toilet, kiosk, etc.
Fitness facilities	Fitness equipment, fitness facility, jogging path distance, etc.
Recreational Amenities	Chess table, children’s play equipment, amusement ride, etc. Museum, science and technology museum, etc.
Natural capital
Ecological Indicators	Mean NDVI

).

3.4. Data Analysis

3.4.1. Well-Being Evaluation Method

Before evaluating well-being outcomes related to the Hankou waterfront, we conducted a reliability and validity analysis of the well-being survey data to confirm their consistency and credibility. The total well-being effect table has a Cronbach α of 0.922, and all five subdimensions had Cronbach α higher than 0.7, indicating high reliability and internal consistency. The total well-being effect table and five dimensions scale’s KMO ranged from 0.679 to 0.882, Bartlett’s sphericity test results were 0.000, and each scale’s total variance interpretation above 70%, showing good validity (Table 5). All observed variables of each scale were subjected to exploratory factor analysis to verify dimension creation. The main component analysis of the well-being effect scale yielded five common variables. The maximum variance approach is used for orthogonal rotation after identifying five common factors, and 0.5 is the crucial factor load. The findings show that no items have factor loads less than 0.5 or multiple factor loads larger than 0.5, hence none should be deleted. Overall, the scale construction effect is good and commensurate with building dimension. The overall well-being scores comprising five dimensions (subsistence security, mental health, physical health, social interaction, and self-actualization) were calculated for each respondent using the mean value method.

3.4.2. Analysis Method for Influencing Factors on Well-Being

(1) Hierarchical Clustering. This study employed hierarchical cluster analysis based on the theoretical framework of human–nature interaction types detailed in Section 2.5, utilizing five typological pathways as clustering variables: sensory affection, health and leisure fulfillment, learning and development, social vibrancy, and spiritual and heritage resources. Due to the binary characteristics of all research variables, Jaccard distance was utilized as the similarity metric, while the Ward.D2 approach was applied for clustering. The cluster labeling adhered to a mean-maximization concept, in which each cluster’s defining characteristic was established by determining the variable with the greatest mean value within the group. The hierarchical analysis grouped the 318 samples into three distinct clusters. The results indicate that Cluster 1 demonstrated the highest engagement rate in sensory affection activities (mean = 0.99), and is therefore classified as the sightseeing-oriented group (n = 56). Cluster 2 exhibited significant engagement in both health and leisure satisfaction (mean = 1) and learning and development activities (mean = 0.7), justifying its designation as the outdoor work-oriented group (n = 126). Cluster 3 exhibited elevated participation levels in social vibrancy (mean = 0.85), alongside moderate engagement in sensory affection (mean = 0.79), indicating a composite pattern primarily characterized by social activities with additional sightseeing elements, thus classified as the social activities-oriented group (n = 136). This classification corresponds effectively with the theoretical framework outlined in Section 2.5 and establishes a robust basis for further analysis.

(2) One-Way ANOVA. One-way ANOVA was conducted using human–nature interaction types as the independent variable and overall well-being, along with its sub-dimensions, as dependent variables, evaluating inter-group differences by mean comparisons and statistical significance (p < 0.05).

(3) Stepwise Regression. The stepwise regression analysis was performed in R with packages such as dplyr, MASS, and stats. Data pretreatment included dummy variable encoding for human–place interaction typologies, utilizing the “social activities-oriented group” as the reference category to reduce multicollinearity, and eliminating incomplete cases through na.omit(). A multiple linear regression approach was utilized to analyze the impact of predictors (natural capital, human capital, social capital, constructed capital, and interaction types) on well-being (overall well-being and its sub-dimensions). Model optimization employed stepwise selection guided by the Akaike Information Criterion (AIC), with final coefficients assessed for statistical significance to identify principal factors influencing variations in well-being. To enhance the methodological robustness of variable selection, LASSO regression was further employed as a validation procedure. LASSO is a regularized regression technique that performs both variable selection and shrinkage, effectively addressing multicollinearity and overfitting concerns. The glmnet package in R was used to perform LASSO regression with 10-fold cross-validation. All predictor variables, including dummy-coded categorical factors, were standardized prior to modeling.

4. Results

4.1. Well-Being Evaluation of Hankou Waterfront

This study employed mean-value analysis to conduct descriptive statistical analysis (Table 5) of the well-being among visitors at Hankou waterfront.

4.1.1. Overall Well-Being at Moderately Elevated Levels

The average overall well-being effect of the Hankou waterfront was 3.33 (SD = 0.785). The total well-being effect experienced by recreationists was at a middle-upper level, with a more centered distribution.

4.1.2. Mental Health Was Most Prominent with the Highest Mean Score

The mental health dimension demonstrated the highest mean score (mean = 3.89) with a relatively low standard deviation, signifying consistently robust psychological benefits among visitors at Hankou waterfront. This underscores the recognition of Hankou waterfront’s role in mitigating negative emotions and restoring attention, closely linked to its function as an Urban Ecosystem Restoration Site.

4.1.3. Social Interaction Effect Ranks Second

Social interaction exhibited a mean score of 3.50, placing it second among all sub-dimensions. Significantly, specific indicators such as SI1 (mean = 3.68) and SI3 (mean = 3.62) attained upper moderate scores, demonstrating Hankou waterfront’s considerable efficacy in creating interpersonal connections and enhancing place attachment among recreationists.

4.1.4. Physical Health and Self-Actualization Exhibited Comparatively Lower Ranks

The mean score for the physical health is 3.42, ranking third, and this is accompanied by a significant standard deviation, indicating varied perceptions among visitors. The mean value for the self-actualization dimension is 3.19, ranking fourth, indicating that recreationists experienced a rather moderate self-actualization effect at Hankou waterfront. However, within this dimension, they reported a stronger effect in enriching personal experience (SA1), with a mean score of 3.57.

4.1.5. Survival Security Ranks Last

The minimal mean value for the survival protection effect (mean = 2.88) indicates a weak perception among recreationists regarding survival security offered by the Hankou waterfront, particularly in terms of essential life resources, suggesting that the functional design of the Hankou riverbank may be inadequate.

Table 5. Descriptive statistics of well-being effects (N = 318).

Item	Minimum	Maximum	Mean	SD	Cronbach’α	KMO
WB	0.69	4.92	3.33	0.785	0.922	0.882
SG	1.00	5.00	2.88	0.997	0.794	0.703
SG1	1.00	5.00	3.54	0.974
SG2	0.00	5.00	2.63	1.212
SG3	0.00	5.00	2.46	1.337
MH	0.00	5.00	3.89	0.794	0.880	0.701
MH1	0.00	5.00	3.98	0.808
MH2	0.00	5.00	3.80	0.871
PH	0.00	5.00	3.42	1.048	0.855	0.712
PH1	0.00	5.00	3.48	1.134
PH2	0.00	5.00	3.36	1.109
SI	0.33	5.00	3.50	0.962	0.815	0.679
SI1	0.00	5.00	3.68	1.050
SI2	0.00	5.00	3.20	1.247
SI3	0.00	5.00	3.62	1.070
SA	0.67	5.00	3.19	0.959	0.845	0.710
SA1	1.00	5.00	3.57	0.966
SA2	0.00	5.00	2.98	1.165
SA3	0.00	5.00	3.00	1.150

Note: Abbreviations of the above items are consistent with Table 2.

Table 5. Descriptive statistics of well-being effects (N = 318).

Item	Minimum	Maximum	Mean	SD	Cronbach’α	KMO
WB	0.69	4.92	3.33	0.785	0.922	0.882
SG	1.00	5.00	2.88	0.997	0.794	0.703
SG1	1.00	5.00	3.54	0.974
SG2	0.00	5.00	2.63	1.212
SG3	0.00	5.00	2.46	1.337
MH	0.00	5.00	3.89	0.794	0.880	0.701
MH1	0.00	5.00	3.98	0.808
MH2	0.00	5.00	3.80	0.871
PH	0.00	5.00	3.42	1.048	0.855	0.712
PH1	0.00	5.00	3.48	1.134
PH2	0.00	5.00	3.36	1.109
SI	0.33	5.00	3.50	0.962	0.815	0.679
SI1	0.00	5.00	3.68	1.050
SI2	0.00	5.00	3.20	1.247
SI3	0.00	5.00	3.62	1.070
SA	0.67	5.00	3.19	0.959	0.845	0.710
SA1	1.00	5.00	3.57	0.966
SA2	0.00	5.00	2.98	1.165
SA3	0.00	5.00	3.00	1.150

Note: Abbreviations of the above items are consistent with Table 2.

4.2. Influencing Factors of Wellbeing of Hankou Waterfront

This section analyzes the factors of well-being at Hankou waterfront through a two-stage analytical approach. First, one-way ANOVA was employed to examine disparities in well-being dimensions across different human–nature interaction typologies. Second, stepwise regression analysis was conducted to elucidate the specific impacts of natural capital, built capital, human capital, social capital, and human–nature interaction types on both overall well-being and their sub-dimensions, but we did not conduct statistical tests for specific bilateral capital interactions, as we focused on evaluating their combined contribution to well-being through simultaneous inclusion in the regression model. Before the regression analysis, multicollinearity detection was conducted on the explanatory variables of the included models in this study. Specifically, we calculated the Variance Inflation Factor (VIF) of each variable. The results show that the VIF values for most variables are lower than three, indicating that there is no serious collinearity problem in the overall model.

4.2.1. Differences in Well-Being Among Human–Nature Interaction Groups

The results of the one-way ANOVA (

Table 6. Well-being scores and ANOVA results across different human–environment interaction types in Hankou waterfront.

	Sightseeing- Oriented	Outdoor Work- Oriented	Social Activities- Oriented	F	p
WB	3.129	3.527	3.237	7.008	0.001
SS	2.667	3.080	2.780	4.607	0.011
MH	3.696	4.048	3.819	4.805	0.009
PH	3.241	3.671	3.272	5.951	0.003
SI	3.310	3.680	3.412	3.930	0.021
SA	2.959	3.378	3.100	4.758	0.009

Note: Abbreviations of the above items are consistent with Table 2.

) indicate significant differences in well-being among the sightseeing-oriented, outdoor work-oriented, and social activities-oriented group (p < 0.05). Significant differences were found in overall well-being (p = 0.001), survival security (p = 0.011), mental health (p = 0.009), physical health (PH, p = 0.003), social interaction (p = 0.021), and self-actualization (p = 0.009). All tests satisfied the homogeneity of variance assumption.

Post hoc multiple comparisons demonstrated that the outdoor work-oriented group consistently outperformed both remaining groups. For instance, for WB, outdoor-work (3.527) > sightseeing-oriented (3.129, p = 0.001) and social activities-oriented (3.237, p = 0.003). For MH. outdoor-work (4.048, p = 0.001) > sightseeing-oriented (3.696, p = 0.006) and social activities-oriented (3.819, p = 0.019). Although the social activities-oriented group had numerically superior means compared to the sightseeing-oriented group across all variables (WB = 3.237, SS = 2.780, MH = 3.819, PH = 3.272, SI = 3.412, SA = 3.100), these differences were not statistically significant (p > 0.05).

4.2.2. The Impact of Diverse Factors on Well-Being

This section utilizes stepwise regression analysis to examine the impact of natural capital, built capital, human and social capital, and the types of human–nature interactions on overall well-being and its sub-dimensions.

Table 7. Results of stepwise regression analysis on well-being determinants.

Variable	WB	SS	MH	PH	SI	SA
Intercept	1.311 ***	1.322	1.639 ***	0.485	1.028	1.076
Control variables
Education	−0.213 ***	−0.213 ***	−0.132 **	−0.289 ***	−0.235 ***	−0.240 ***
Income	8.679 **	0.102 **	0.092 **	-	0.142 ***	0.063
Gender	8.755	0.079	-	5.363 ***	-	−0.023
Age	-	-	0.065 *	-	-	-
Natural capital factor
NDVI	8.382 ***	8.197 **	4.764 *	2.062 ***	8.648 ***	12.640 ***
Built capital factor
Recreational facilities	-	−0.003 *	0.001	-	-	-
Fitness facilities	−2.293 ***	−0.004 ***	−0.003 ***	−3.950 ***	-	−0.001
Infrastructure	−8.989	0.003 ***	0.001	-	−0.002 ***	−0.001
Human capital factor
Activity frequency	8.418 **	0.102 **	0.117 ***	1.785 ***	-	0.063
Activity duration	2.956	−0.132	-	7.330	0.111	-
Nature preference	0.157 **	0.131 *	0.144 **	4.653	0.280 ***	0.126
Social capital factor
Length of residence	1.407 **	0.150 ***	0.074	0.196 ***	0.116 *	0.101 *
Companionship	-	-	0.452 ***	0.283 *	-	-
Human–nature internation type factor
Sightseeing-oriented group	5.019	-	1.060 *	1.959 **	0.653	0.317
Outdoor work-oriented group	0.792 **	-	0.943 **	1.671 **	0.674	2.021 **
R2	0.681	0.684	0.598	0.72	0.568	0.553
Adjusted R2	0.641	0.643	0.551	0.681	0.534	0.502
F-statistic	9.603 ***	9.231 ***	6.274 ***	10.7 ***	7.885 ***	4.995 ***

***: p < 0.001; **: p < 0.01; *: p < 0.05; “-” indicates excluded variables; abbreviations for the above items correspond to Table 2.

demonstrates that the R² values of the six regression models range from 0.502 to 0.643, signifying that the chosen variables account for a considerable percentage of the variance in various dimensions of well-being. All models produced significant F-values (p < 0.001), indicating substantial explanatory power and robust model fit. The residual plots validated a random distribution of errors, affirming the dependability of the data.

(1)

Overall Well-Being: NDVI and Activity Frequency as Primary Drivers.

Overall well-being was significantly positively associated with NDVI (β = 8.382, p < 0.001), activity frequency (8.418, p < 0.01), and length of residence (1.407, p < 0.01). Conversely, education level (−0.213, p < 0.001) and fitness facilities (−2.293, p < 0.01) had significant negative effects. Furthermore, individuals with outdoor work orientation (0.792, p < 0.01) reported higher well-being compared to those oriented towards social activities.

(2)

Survival Security: Predominant Positive Effects of NDVI and Significant Negative Effects of Education Level

Regression analysis revealed that survival security was significantly enhanced by NDVI (8.197, p = 0.01), length of residence (0.150, p < 0.001), nature preference (0.131, p < 0.05), activity frequency (0.102, p < 0.01), and infrastructure (0.003, p < 0.001). Conversely, education level (−0.213, p < 0.001), fitness facilities (−0.004, p < 0.001), and recreational amenities (−0.003, p < 0.05) exhibited significant negative associations. Although fitness and recreational facilities showed negative effects, their coefficients were minimal and may be considered negligible in practical terms.

(3)

Mental Health: Combined Influence of NDVI, Social Capital, and Interaction Type

Mental health exhibited a positive correlation with NDVI (4.764, p < 0.05), companionship (0.452, p < 0.001), natural preference (0.144, p < 0.01), and activity frequency (0.117, p < 0.001). Detrimental predictors were educational attainment (−0.132, p < 0.01) and fitness facilities (−0.003, p < 0.001). Negative predictors included education level (−0.132, p < 0.01) and fitness facilities (−0.003, p < 0.001). Both the sightseeing-oriented (1.060, p < 0.05) and outdoor work-oriented (0.943, p < 0.01) groups reported significantly higher mental health outcomes than the social activity-oriented group.

(4)

Physical Health: Strong Inhibitory Effect of Fitness Facilities

Positive contributors to physical health included gender (5.363, p < 0.001), NDVI (2.062, p < 0.001), companionship (0.283, p < 0.05), length of residence (0.196, p < 0.001), and activity frequency (1.785, p < 0.001). Negative effects were observed for fitness facilities (−3.950, p < 0.001) and education level (−0.289, p < 0.001). Both the sightseeing-oriented (1.959, p < 0.01) and outdoor work-oriented (1.671, p < 0.01) groups showed higher physical health scores compared to the social activity-oriented group.

(5)

Social Interaction: NDVI, Nature Preference, and Income Arise as Principal Predictors

Regression analysis reveals that NDVI (8.648, p < 0.001), nature preference (0.280, p < 0.001), income level (0.142, p < 0.001), and the length of residence exhibited a positive correlation (0.116, p < 0.05). In contrast, education level (β = −0.235, p < 0.001) and infrastructure (−0.002, p < 0.01) demonstrated statistically significant albeit practically insignificant negative effects. Significantly, none of the typologies of human–nature interaction exhibited statistically meaningful effects on social interaction.

(6)

Self-Actualization: Central Role of NDVI

Self-actualization was positively predicted by NDVI (12.640, p < 0.001) and length of residence (0.101, p < 0.05), while education level had a significant negative impact (−0.240, p < 0.001). Compared to the social activity-oriented group, individuals with an outdoor work-oriented (0.792, p < 0.01) exhibited significantly higher levels of self-actualization.

4.2.3. Robustness Test

To validate the stability of the predictors identified via stepwise regression, LASSO regression was conducted as a robustness check. The results demonstrated complete consistency between the predictors retained by stepwise and LASSO models across all well-being dimensions (WB, SS, MH, PH, SI, and SA). Specifically, 100% of the variables identified as significant by stepwise regression were also selected by LASSO (see Figure 4).

Further analysis reveals that NDVI, education level, income, natural preference, activity frequency, length of residence, and group type are all retained in the six models, demonstrating extremely high explanatory power and stability, and are the core elements influencing subjective well-being. In contrast, recreational facilities, gender, and activity duration were excluded in some dimensions, indicating that their influence is dimensionally selective or has relatively weak explanatory power. Overall, LASSO regression effectively enhances the refinement of variable selection and further verifies the robustness of the model structure and key variables. (see

Table 8. LASSO regression analysis.

Variables	WB	SS	MH	PH	SI	SA
NDVI	1	1	1	1	1	1
Fitness facilities	1	1	1	1	1	0
Infrastructure	1	1	1	0	0	0
Recreational facilities	0	1	0	0	0	0
Activity frequency	1	1	1	1	1	1
Activity duration	1	0	0	1	0	0
Companionship	1	0	1	1	0	0
Nature preference	1	1	1	1	1	1
Education	1	1	1	1	1	1
Income	1	1	1	1	1	1
Gender	0	1	0	1	0	1
Age	1	1	1	1	0	1
Length of residence	1	1	1	1	1	1
Sightseeing-oriented group	1	1	1	1	1	1
Outwork-oriented group	1	1	1	1	1	1
Stepwise variables number	13	13	12	13	9	10
LASSO Variables number	13	13	12	13	9	10

Note: “1” indicates selection in the LASSO regression analysis, while “0” indicates not selected. abbreviations for the above items correspond to Table 2.

).

5. Discussion

5.1. Upper Middle Well-Being Among Hankou Waterfront Recreationists: Mental Health Most Prominent, Survival Security Perceived Lowest

We thoroughly assessed recreationists’ well-being in Hankou waterfront, revealing that the overall well-being was moderate to high. The sub-categories of well-being, in descending order, were as follows: mental health, social interaction, physical health, self-actualization, and survival security. The mental health scores were the highest, underscoring the significant value of UERS in facilitating emotional regulation and psychological rehabilitation. Hankou waterfront, regarded as the “green lung” of central Wuhan, is characterized by a rich array of natural landscapes, including varied flora, aquatic environments, wetland ecosystems, and seasonal migratory avifauna. By engaging with these natural components, recreationists can attain attention restoration, emotional relief, and an overall sense of psychological well-being [52].

Conversely, survival security received the lowest score, primarily influenced by two factors: firstly, although Hankou waterfront possessing significant ecological safety functions, such as flood control and disaster mitigation, field surveys indicate that these functions have not been effectively promoted, leading to a lack of awareness among visitors regarding the ecological safety benefits of Hankou waterfront. At the same time, we take the recreationists as survey samples, as the Hankou waterfront exhibits a significant level of host–guest sharing [53]; hence, the research subjects encompass both local people and a proportion of tourists. Tourists engage in brief recreational and leisure activities and exhibit minimal reliance for daily life resources, including work opportunities, economic assistance, and essential resource provision on the Hankou waterfront. Consequently, their assessment of survival security is comparatively feeble, leading to a diminished rating of subjective well-being in this dimension.

5.2. Negative Effects of Education and Fitness Facilities, Varied Influences Across Well-Being Dimensions, and the Significant Role of Human–Nature Interaction

NDVI showed a significant positive effect in all well-being types, especially in mental health and self-actualization, which verified the H1 and effectively echoed previous studies, confirming that the natural environment plays an important role in improving human well-being such as restorative experience and emotional connection [22,26,52]. Activity frequency shows a positive effect on overall well-being, survival security, physical health, and mental health. This is consistent with the empirical research by Nath [42], who pointed out that frequent participation in green space activities is a key path to improving residents’ happiness and health perception. Kondo [3] also emphasized, from an experimental design standpoint, that brief, high-frequency interactions with green spaces can markedly enhance mood, concentration, and psychological recovery. The length of residence significantly influences perceptions of survival security and self-actualization, indicating that these perceptions are contingent not only on the ecological conditions, but also on the emotional bonds and daily familiarity developed with the place [24]. Moreover, nature preference and companionship significantly influenced self-realization and mental health, respectively. This illustrates the influence of an individual’s inherent aesthetic orientation on self-actualization and underscores the significant role of interpersonal interactions in enhancing psychological recovery, further affirming that UERS serves not only as a physical environment but also as a crucial medium for social and emotional support and identity formation [54].

Education level has a significant negative effect on all dimensions of well-being, especially in physical health and self-actualization. This conclusion diverges from the findings of prior investigations; Ma [26] discovered that educational level, particularly at the master’s level or higher, significantly enhances well-being. This finding can be interpreted from two aspects: (1) sample type and behavioral patterns, (2) cognitive expectations and perceptual sensitivity. Firstly, this discrepancy may arise from the fact that existing studies (e.g., Ma et al.) utilized local inhabitants as the survey sample, and URES is a Usual Environment for them. Consequently, the assessment of well-being serves as a societal and prevalent indicator. However, we primarily focus on UERS visitors as research subjects, using the experience of CES as the entrance point. The respondents may prioritize immediate pleasures and emotional gratification over well-being as defined by institutional or societal metrics. Secondly, studies have shown that higher-educated visitors to be more motivated to use parks to relieve stress and enjoy the views within the park in contrast to those from lower-education backgrounds [55]. For this reason, people with higher education levels may hold a more critical attitude toward public environments, thus assessing their experiences not only in terms of emotional satisfaction, but also in alignment with personal ideals of autonomy, quality, and meaning. If the design details, order management, or crowd density of the space environment do not align with an individuals’ psychological expectations, a significant “expectation–experience gap” may arise, which can in turn lead to a decline in perceived well-being [24].

Similarly, fitness facilities have a significant negative impact on all types of well-being. Gavin argues that the proximity of facilities and amenities appears to influence physical activity participation [56]. Field survey revealed that, although the Hankou waterfront is equipped with extensive fitness infrastructure, actual utilization remains low. Key barriers have been identified from high entry costs, mismatch between infrastructure supply– demand, and spatial conflicts arising from mixed-use by tourists and residents. First, some large sports venues operate on a paid-entry basis, limiting access for casual recreationists and generating localized noise that disrupts surrounding relaxation spaces. As a visitor noted, “sightseeing carts require a 120 RMB deposit and can only be returned at distant endpoints, creating inconvenience and reducing recreational flexibility”. Second, there is a clear mismatch between infrastructure provision and demand. Many fitness facilities were observed to be used primarily for sitting rather than exercise, particularly by elderly people, thus failing to fulfil their intended fitness functions. Last but not least, spatial conflicts have emerged due to the overlap between fitness and tourism activities. During weekend peak hours, over half of the jogging lanes were occupied by sightseeing vehicles, forcing pedestrians to detour through greenbelts and resulting in frequent disputes. The negative impact of fitness facilities on well-being can also be interpreted through the lens of person–environment fit theory, which suggests that optimal outcomes occur when individual needs and preferences align with environmental offerings. In UERS, when fitness infrastructure fails to meet users’ expectations for accessibility, usability or comfort, a misfit occurs, resulting in dissatisfaction. For instance, visitors seeking light physical activity or tranquil rest may find the design, noise, or spatial configuration of sports zones intrusive or mismatched, leading to a negative impact on their perceived well-being.

Human–nature interaction types significantly impact well-being. The univariate analysis revealed significant differences in well-being levels among different types of human–nature interaction. Specifically, the outdoor work-oriented group exhibited significantly higher levels compared to the sightseeing-oriented and social activities-oriented groups. Nonetheless, this advantage became more differentiated across well-being dimensions in the multivariate regression analysis. Compared to the social activities-oriented group, the outdoor work-oriented type significantly enhanced overall well-being, mental health, physical health, and self-actualization, while the sightseeing-oriented group experienced a significant effect on mental health and physical health. These differences can be interpreted through both behavioral and psychological mechanisms. From a behavioral perspective, drawing on the concept of the co-production of ecosystem services, CES are essentially co-created through human–nature interactions, and the benefits individuals derive from ecosystem largely depend on the depth and intensity of human–nature interaction [57]; the outdoor work-oriented group typically engages with UERS through frequent, physically active, and goal-directed behaviors. These high-frequency, immersive activities tend to promote deeper physical engagement, sustained attention, and a greater sense of environmental connectedness, and have been shown to enhance both physical recovery and psychological resilience. From a psychological perspective, the outdoor work-oriented group, characterized by higher activity frequency and deeper situational immersion, is more likely to activate what Huynh [11] refer to as regenerative, formative, and identity or autonomy mechanisms, which contribute to physical restoration, self-worth affirmation, and enhanced subjective well-being [11,58]. According to self-determination theory, environments that support autonomy and competence enhance intrinsic motivation, leading to more profound well-being. In contrast, the sightseeing-oriented group engages in more passive, observational behaviors, which primarily support short-term cognitive restoration and aesthetic pleasure but lack the embodied or goal-driven depth required to activate broader well-being outcomes. This aligns with attention restoration theory, which suggests that nature environments can restore attention capacity but are less likely to lead to the deeper, goal-directed psychological engagement required to activate broader well-being. The passive nature of this interaction style may limit the full potential of restorative benefits, as it lacks the active involvement needed for sustained well-being improvement. The social activities-oriented group, while socially engaging, may be influenced by factors like crowding, noise, or fragmented attention, which could limit restorative benefits and create more variable well-being.

Theoretically, our research empirically substantiates the foundational premise of the “four types of capital interactions” posited by Costanza for the generation of well-being, affirming that well-being is not influenced by a singular factor, but rather arises from the synergistic effects of multiple capital interactions [21]. On this basis, we incorporated the variable of “human–nature interaction” type and conducted an empirical analysis of its action pathway, addressing the deficiency in behavioral mechanisms in Costanza’s research. That is, under the same or similar capital allocation conditions, the human well-being stimulated by different human–nature interaction types is significantly different, and the human well-being types and action intensities stimulated by different interaction strategies are heterogeneous. Therefore, the initial static premise of “capital interaction automatically produces welfare” requires additional elaboration into a dynamic mechanistic model of “capital interaction→behavioral choices→well-being generation.” This study enhances the theoretical comprehension of the contribution of CES to well-being and offers conceptual support for how UERS can facilitate more nuanced and behaviorally informed well-being provision.

It is noteworthy that the field survey was conducted after the relaxation of post-COVID-19 policies, a point in time which may have influenced well-being perceptions. During the pandemic, many individuals faced immense stress, particularly university students, who faced significant psychological and physical stress due to social isolation, academic pressure, and an uncertain future. Studies have shown that this led to issues such as anxiety, depression, sleep disorders, and weight gain, severely affecting both daily life and academic performance [39,59]. In this study, various forms of natural interaction may help to alleviate these pressures. Sightseeing-oriented interactions improved mood, reduced anxiety, and helped alleviate loneliness, while also enhancing social support and emotional recovery. Additionally, outdoor work-oriented interaction reduced psychological stress, strengthened emotional belonging, and increased self-fulfillment, promoting overall recovery. Due to the limited sample size, further empirical testing is needed. Future research will focus on college students, including longitudinal tracking, psychological and physiological measurements, and comparisons of human–nature interactions. The easing of restrictions and the re-engagement with public spaces may have played a significant role in these findings, and further analysis of this timing could provide valuable insights for future studies.

5.3. Policy Recommendations

Firstly, the well-being evaluation results show that the perceived survival security is the lowest. In light of this result, it is necessary to enhance the public awareness and education regarding the ecological and security guarantee functions of the Hankou waterfront. Through various channels, such as media campaigns, internet platforms and outdoor promotions, the ecological and security guarantee functions of the Hankou waterfront and their significance should be popularized among visitors. In addition, it is possible to consider strengthening the management of Hankou waterfront and enhancing its security guarantee functions. For instance, additional surveillance equipment can be installed, patrol management can be strengthened, and emergency response plans can be improved to further enhance visitors’ sense of security and trust.

Secondly, concerning the negative effects of fitness facilities, management should focus on the structural mismatch between facility provision and demand at Hankou waterfront. Specifically, adjustments to the spatial layout and functional positioning of fitness facilities are needed to avoid overlap with high-traffic zones such as main walking trails or sightseeing vehicle lanes. In addition, the integration of smart management tools can enhance user experience and operational efficiency. For example, installing real-time usage monitoring systems can inform data-driven adjustments to facility placement and maintenance. Furthermore, establishing buffer zones through landscape elements, such as “quiet area” signage or vegetation barriers around fitness areas can help balance the needs of active users with those seeking tranquil recreation, thereby achieving flexible spatial zoning for multifunctional use. This can be accomplished by optimizing facility layout and functional zoning, namely by avoiding the placement of workout facilities in high-traffic locations such as primary pedestrian walkways and sightseeing paths, to reconcile the recreational needs of diverse visitors.

Finally, the diverse well-being perception among various human–nature interaction groups offer empirical validation for the management of UERS. UERS design should prioritize enhancing human–nature interactions to successfully promote the CES and optimize multidimensional well-being [40]. For example, for tourist-oriented UERS (targeting the sightseeing-oriented users), the predominantly experience-based and low-intensity activities of visitors (e.g., photography, flower-viewing, strolling) strongly depend on features such as scenic quality, interpretive signage, and cultural storytelling, which requires managers to reinforce both ecological aesthetics and landscape connectivity, while also expanding cultural and environmental knowledge dissemination, to elevate aesthetic and cognitive engagement of visitors. In contrast, community-oriented UERS should address the deeper participatory needs of residents through intergenerational fitness zones, socially inclusive seating arrangements, and co-design mechanisms. Or, if UERS is to serve as a hybrid social–production space (for outdoor work- and social activities-oriented users), the same logic of goal-oriented CES provision can be applied to enhance well-being perception through targeted infrastructure and experience design. In sum, while CES are fundamentally intangible, their delivery pathways can be structured and governed through the design of tangible landscape features and institutional interventions [31].

6. Conclusions

The relationship between UERS and human well-being has garnered much attention. However, the perceived differences and their impact on well-being based on the human–nature interaction types are often overlooked. Taking Wuhan Hankou waterfront as an example, we analyzed the well-being and its influencing factors of the recreationists. The results are as follows:

(1)

Mean value analysis showed the overall well-being of recreationists in Hankou waterfront is in the upper middle level (mean = 3.33), mental health is the most prominent (mean = 3.89), and the survival security is the lowest (mean = 2.88), indicating that Hankou waterfront does not really meet the diversified needs of the recreationists, revealing certain management deficiencies.

(2)

Stepwise regression analysis showed education level and fitness facilities had a consistent negative effect on all types of well-being, and NDVI showed a significant positive effect on all types of well-being, especially in mental health and self-actualization, with varying factors affecting different aspects of well-being.

(3)

Univariate analysis shows the outdoor work-oriented group had higher well-being scores than sightseeing-oriented and social activities-oriented. Compared to social activities-oriented group, the outdoor work-oriented group significantly improved overall well-being, mental health, physical health, and self-realization, while the sightseeing-oriented type mainly enhanced mental and physical health. Finally, the results supported H1 and H10, partially supported H2–H4 and H5–H6, but rejected H7–H9.

This study has the following contributions: First, we have verified that well-being is driven by the interaction and co-production of multiple capital. On this basis, we introduce the variable of “human–nature interaction type” to make up for Costanza’s insufficient consideration of human behavior types. Second, we have fully verified the significant impact of CES experience and human–nature interaction on human well-being, providing valuable insights for urban planning and design. Although CES are essentially intangible, their action paths can be adjusted through tangible landscape element organization and management intervention mechanisms, so as to maximize human well-being.

This study has some limitations. First, this study employed a convenience sampling method, which, while practical for on-site data collection, presents inherent limitations in representativeness. For instance, the majority of participants were well-educated, which may not reflect the demographic composition of the broader urban population. In addition, the sample did not explicitly distinguish between residents and tourists, limiting the ability to assess how user identity may moderate the relationship between UERS experiences and well-being. As such, future studies should consider using stratified sampling methods and seek to include underrepresented demographic groups to enhance external validity and policy relevance. Second, the sample excludes other types of UERS like community gardens. Expanding to diverse urban ecological spaces would strengthen the theoretical framework of well-being mechanisms. Also, future studies are encouraged to extend this framework by comparing ecological functions and well-being pathways across diverse types of urban green and blue spaces. Also, climatic and seasonal variables, which may influence user experiences, comfort levels, and emotional responses, were not included in the regression models. Future research should consider controlling temperature, air quality, and seasonality to improve the accuracy of causal inference.