Pursuing Ecological and Social Co-Benefits: Public Hierarchical Willingness for Biodiversity Conservation in Urban Parks

Abstract

Urban green spaces play a critical role in sustaining the urban park biodiversity. The relationship between biodiversity and city residents is complex. Understanding the cognitive preferences of residents toward biodiversity is vital for effective conservation. This study investigated the public willingness to protect the biodiversity in urban parks using questionnaire-based assessments and explored the underlying drivers. The study focused on the residents of Hangzhou, China, and analyzed the effects of respondent and visit characteristics as well as their interactions using ANOVA, PERMANOVA, GLM, and NMDS. The visitor age, education level, satisfaction with plant landscapes, and visit frequency significantly influenced their willingness to conserve. Based on a “cognitive-experience-investment” framework, we uncovered (1) positive synergistic effects between urban park biodiversity and the abundance of urban green space fauna; (2) threshold constraints linking volunteer time for biodiversity conservation and economic expenditures on biodiversity-friendly products; and (3) the complex interactions among these factors. The findings not only elucidate the driving mechanisms and model optimization pathways associated with public willingness for conserving urban biodiversity but also provide actionable strategies to promote both ecological conservation and societal wellbeing.

1. Introduction

Park biodiversity has an important impact on the function of urban ecosystem services and the health of human habitats. With the acceleration of urbanization, urban parks, an important component of urban green spaces (UGSs), are vital for maintaining the ecological balance of cities and improving the quality of life for residents. Indeed, high levels of biodiversity are essential for maintaining the ecosystem services that support human wellbeing by promoting physical and mental health. The public perceptions of biodiversity in urban parks have a direct impact on the awareness and participation in conservation, in turn providing ecological and social co-benefits; however, mainstream research has seldom considered the independent effects of public awareness or perceptions of urban park biodiversity. Therefore, in-depth research is needed to formulate rational UGS planning and sustainable development policies.

With the rapid urbanization, urban development transforms vast natural areas, directly reducing the biodiversity. Numerous studies examining the species diversity in urban areas and along urban–rural gradients have confirmed the negative impacts of urbanization on biodiversity [1], though urban environments may also accommodate a unique biodiversity [2]. Urban parks play important roles in the functioning of urban ecosystems, facilitating citizen activities, and conserving biodiversity. These environments provide diverse services to citizens [3]. Many ecosystem services rely on species interactions [4]. Therefore, there is an urgent need to improve the species-richness within human-dominated landscapes to enhance the biodiversity and ecosystem functioning [5,6]. The multiple ecological and social benefits of urban biodiversity are increasingly recognized; promoting the connections between citizens and the natural environment can not only improve the public’s quality of life but also protect crucial habitats [7].

Urban parks provide habitats for urban biodiversity and spaces for residents to gain a sense of connection/unity with, and find pleasure in, nature. This connection to the natural world represents a fundamental human need, which can be promoted through nature-centered design strategies for urban environments [8]. In the pursuit of socio-ecological benefits, it is essential to recognize the uniqueness of human perceptions and preferences, especially in the context of “ecological aesthetics”. Traditionally, urban parks have been solely designed to meet the needs of residents; however, as the demand for natural spaces by urban residents has increased, so has the demand for biodiversity [9]. Previous studies on urban park biodiversity found that, compared to the ecological value, residents preferred the esthetic value of biodiverse parks [10], whereas the public cognitive preference showed the opposite trend [11,12]. Often, areas with highly valued ecology may not be perceived to have a high landscape value. Therefore, understanding public preferences and attitudes toward biodiversity is critical for designing effective policies and strategies that support biodiversity, inform the design of urban parks, and increase the ecological awareness of residents.

Studies on conservation willingness mainly use structural equation modeling [13,14] and questionnaire survey methods [15] to assess the influence of sociocultural factors on the public perception of biodiversity. Previous studies have shown that the level of ecological awareness of urban residents was positively correlated with the willingness to pay (WTP) [14], and was closely associated with education and community involvement [16]. The results of the study showed that social support was closely related to attitudes toward and the WTP for biodiversity conservation [17]. The contingent valuation method (CVM), used to calculate the monetary value of services provided [18], showed that the WTP was affected by the design of the payment instrument. The CVM is a survey-based technique that determines the awareness of rare and endangered species conservation and the WTP based on a hypothetical market of resources or services for non-marketed species [19,20,21]. The main questionnaire formats include open-ended, dichotomous choice, and payment card (PC) formats, with the latter being the most common [22,23]. The PCs are subdivided into unanchored (UPCs) and anchored PCs (APCs). UPCs ask respondents to select the highest amount that they are willing to pay from a set of data on a given value, or to indicate the highest amount directly. APCs provide respondents with background information and a query about their WTP to provide binding contextual data [22,24]. These methods are useful for quantifying the public’s willingness to conserve.

The international academic community has focused on the interaction between economic incentives and spatial heterogeneity, mostly using discrete choice modeling [25] and ecosystem service valuation methods [26]. The elasticity coefficient of the resident WTP was found to increase around biodiversity hotspots, reflecting the strong influence of geospatial characteristics on the willingness to protect [27]. A previous related study identified five core incentives including environmental benefits, ecological diversity, nature–culture interaction, landscape recreation function, and intergenerational sustainability, revealing the comprehensive economic value of urban natural resources [28]. In addition, a regression analysis showed that positive resident attitudes toward conservation, a sense of local dependence, and social bonds positively influenced the WTP and bid level, whereas place effects (e.g., overreliance on local resources) were negatively related to the WTP [29], highlighting the complex roles of socio-psychological and spatial attributes in conservation decision making. Using the CVM, questionnaire surveys, and multiple regression analysis, we measured the economic value of plant diversity and its spatial heterogeneity in older communities [30]; through field research and scale analysis, we analyzed the cognitive preferences of residents for biodiversity in different types of green spaces and the mechanisms associated with the recovery benefits, confirming the transmission path of ecosystem service perception in terms of health and wellbeing [31].

Various techniques have been developed in the assessment of biodiversity perception and value. The Biodiversity Loss Perception Scale (BiLoPS) systematically quantifies the changes in public perceptions of biodiversity and highlights the scientific value of the standardized scales in cognitive assessments of ecology [32]. Related experiments on the marginal impact of interventions in biodiversity information on the public’s WTP have provided empirical support for policy communication strategies [14]. The mixed methodologies of these studies promote the application of biodiversity valuation in decision making through theoretical modeling, providing a strong scientific basis for urban green space planning and biodiversity strategy development for sustainable development policies to guide the construction of biodiversity-friendly urban areas.

Various case studies on biodiversity conservation with varying geographic coverage and sample selection strategies have revealed the interactive effects of socio-cultural and spatial attributes: Lu et al. demonstrated the heterogeneity of urban resident preferences for green infrastructure in Wuhan, China [33]. Another study across multicultural groups compared the differences in the perceptions of urban natural spaces between gardeners and park users, highlighting the role of social roles in value judgments [34]. In Japanese agroecosystems, most respondents were willing to pay a premium of 0–30% for biodiversity-friendly rice, and the public’s landscape preferences for biodiversity-friendly agriculture were hierarchically differentiated, with rice paddy landscapes having the highest acceptance, followed by plants and birds, and with least support for insects and frogs, reflecting the social and cognitive biases of species esthetic and functional attributes [35]. A study on the mechanisms of community participation in water resource protection in sacred groves integrated the three-dimensional indicators of the WTP, willingness to accept (WTA), and willingness to work to quantify the behavioral tendencies of local residents toward ecological protection and their economic thresholds [36].

Past domestic and international studies on the resident perceptions of biodiversity, preferences, and value used various methodologies, reported divergent conclusions, and predicted trends in different environments (

Table 1. Research dimensions and common keywords in the literature.

Research Dimension	Keywords
Policy tools	Eco-compensation mechanism [32], government-led [33], market incentives [19,20,21], digital payment [23]
Behavioral drivers	Community participation [16], traditional culture [36], individual utility maximization [34], carbon footprint [37]
Technical approach	Questionnaire survey [15], structural equation modeling [13,14], discrete choice experiments [25], machine learning algorithms [3]
Controversies	Regional fairness [31], payment standards [29,30], moral license [34], payment instruments [35]

); however, few multidimensional studies have assessed the public perceptions of biodiversity in urban parks. Accordingly, this study aimed to analyze and predict the public’s views on biodiversity in urban parks using a questionnaire-based approach and further explored the underlying mechanisms. We assessed the effects of respondent characteristics on the key response variables, and explored the interactions among the response variables through a multilevel conservation perspective. This study not only improves the theoretical analysis of the public’s willingness to conserve biodiversity in urban parks, reveals the influencing factors behind it, and expands the scope of environmental behavioral research; it also provides a strong scientific basis for UGS planning and sustainable development policymaking. This study has far-reaching theoretical and practical significance in the construction of an ecologically minded civilization that promotes coexistence between humans and nature.

2. Methods

2.1. Study Area

Hangzhou is the capital of Zhejiang Province, China, and an important city in the Yangtze River Delta. Hangzhou’s unique climate and geography have shaped its rich UGSs.

Four central urban areas of Hangzhou, namely the Shangcheng, Xihu, Gongshu, and Binjiang districts, were selected as study sites (Figure 1), covering an area of ~625.63 km² and collectively comprising 48 subdistricts. The study sites have experienced rapid expansions in the population and built-up area over the past 40 years, is inhabited by 4.245 million people, and is the center of Hangzhou’s economy, culture, and commerce [38]. A high-density population and high-intensity urban activities have led to increasing conflicts between biodiversity and development demand. Hangzhou, a typical high-density polycentric city, exhibits large differences in population density, land type, construction levels, and socioeconomic levels among city clusters as well as differences in the levels of biodiversity supply and demand. Hangzhou provides a representative case study of ecosocial co-benefits in multicenter high-density cities. As the host city of the G20 Summit, a number of large-scale urban construction and landscape renovation projects have been conducted in the past two decades around urban blue spaces, including the Xixi Wetland Comprehensive Protection Project, Grand Canal Protection and Development Project, and Five Waters Co-management Project. Urban construction projects have greatly improved the quality of the urban blue spaces and promoted their transformation into UGSs. For example, during the development of the Grand Canal, beautiful linear green spaces were created on both sides of the waterfront. Industrial factories in built-up urban areas have been transformed into parks to address the growing demand for spaces supporting outdoor leisure activities. In the suburbs, many rural and wetland parks have been created through land exchange. The number of open parks grew from 181 in 2011 to 355 in 2021 [28].

2.2. Data Collection

This study used a questionnaire to collect the data. The questionnaire was divided into four parts: (1) the respondent’s basic information, including their age, sex, and education level; (2) the respondent’s perceptions of urban biodiversity friendliness; (3) the respondent’s concern and satisfaction with biodiversity in urban parks and green spaces, including the characteristics of the respondent’s park and green space visits, concerns about urban organisms, and preference for urban organisms; and (4) the respondent’s willingness to protect urban biodiversity in terms of biodiversity perception, biodiversity impacts, willingness to invest time, willingness to invest money, and basic information (

Table 2. Variable assignment from questionnaire.

Category	Variable	Definition and Value
Basic information	Sex	1 = “Male”, 2 = “Female”
	Age	1 = “18–40”, 2 = “40–60”, 3 = “≥60”
	Education level	1 = “Junior high school and below”, 2 = “High School or technical school”, 3 = “Bachelor or junior college”, 4 = “Master or above”
	Perception of urban biodiversity	1 = “Species”, 2 = “Quantity”, 3 = “Genetic information”, 4 = “Culture and folklore”, 5 = “Other”
Visit characteristics	Frequency	1 = “Other”, 2 = “Six months and more”, 3 = “Monthly”, 4 = “Weekly”, 5 = “Daily”
	Distance	1 = “Other”, 2 = “Within 10 min of driving”, 3 = “Within 10 min of biking”, 4 = “Within 15 min of walking”, 5 = “Within 5 min of walking”
Biodiversity concerns	Birds	1 = “1 species”, 2 = “2 species”, 3 = “3 species”, 4 = “4 species”, 5 = “5 species”, 6 = “6 species”, 7 = “7 species”
	Other animals	0 = “None”, 1 = “Worms, Abandoned pets”, 2 = “Other”, 3 = “Amphibians, Reptiles”, 4 = “Fish”, 5 = “Mammals”
Environmental characteristics of parks	Tree coverage	1 = “Very low”, 2 = “Dense”, 3 = “Low”, 4 = “High”, 5 = “Medium”
	Plant landscape satisfaction	1 = “Very dissatisfied”, 2 = “Dissatisfied”, 3 = “Average”, 4 = “Satisfied”, 5 = “Very satisfied”
Willingness to protect biodiversity	Diversity	1 = “Very poor”, 2 = “Poor”, 3 = “Fair”, 4 = “Good”, 5 = “Very good”
	Effect	1 = “Lower”, 2 = “Low”, 3 = “Average”, 4 = “High”, 5 = “Higher”
	Period	1 = “No participation”, 2 = “Half a year or more”, 3 = “Once a season”, 4 = “Once a month”, 5 = “Once a week”
	Pay	1 = “Within 50 yuan”, 2 = “50–500 yuan”, 3 = “500–2000 yuan”, 4 = “more than 2000 yuan”

). The data were also validated via the pre-survey in March 2024 (n = 120, Cronbach’s α = 0.79), and the complete questionnaire in English and Chinese is provided in the Supplementary Materials.

From 14 May to 10 September 2024, the questionnaire survey was conducted on the construction of biodiversity-friendly areas in the main urban parks of Hangzhou. All the questionnaires were distributed during interviews with the participants selected through random sampling to promote the efficiency of the questionnaire collection and the quality of answers. A total of 2000 questionnaires were distributed, and 1885 valid questionnaires were recovered (recovery rate of 94.25%).

2.3. Data Analysis

2.3.1. Analysis of Variance (ANOVA)

ANOVA is a statistical method for testing the difference between the means of two or more groups [39]. ANOVA was used to test the effect of the respondent characteristics on each of four response variables. A one-way ANOVA can reveal the degree of influence of single factors, such as age, sex, and education, on the resident perceptions of biodiversity, the impact of biodiversity, willingness to invest time, and willingness to invest money. Permutation multivariate ANOVA (PERMANOVA) was used to test the effects of the respondent characteristics on the combined differences in four response variables. PERMANOVA enables the assessment of the overall variation in public perceptions in terms of multiple characteristics.

2.3.2. Generalized Linear Model (GLM) with Stepwise Regression Analysis

The GLM is an extension of the classical linear regression model but departs from the traditional assumptions about the normal distribution and linear relationship of response variables by introducing the link function and exponential family distribution [39,40]. This framework enables the modeling of non-normal distributions (such as binomial, Poisson, and gamma distributions) and nonlinear relationships [40,41].

Stepwise regression is used for variable selection based on the Akaike information criterion (AIC) in which predictor variables are added or removed stepwise to construct the optimal simplified linear model. This reduces the risk of multicollinearity and overfitting while promoting the explanatory power of the model.

$$AIC=2k-2\ln \left(L\right)$$

(1)

Here, k is the number of model parameters and L is the value of the likelihood function. In the GLM framework, the stepwise method is used to screen explanatory variables that are significantly correlated with the response variable and is particularly suitable for high-dimensional data [42]. We used the GLM combined with stepwise regression to analyze the specific effects of each factor on each of the four response variables, identify the most significant explanatory variables, and optimize the model fit.

2.3.3. Non-Metric Multidimensional Scale (NMDS) Analysis

NMDS is a rank-ordering method based on rank–order relationships that aims to map high-dimensional data into a low-dimensional space (e.g., two dimensions) while preserving the original order relationships among the samples. Unlike PCA, PCoA, and other methods based on eigendecomposition, NMDS does not directly rely on numerical distances, but rather approximates the ordering structure of the original distance matrix by iteratively optimizing the minimization of the “Stress function” [43] as follows:

$$Stress=\sqrt{{\displaystyle \frac{\sum {\left(d_{ij}-{\widehat{d}}_{ij}\right)}^2}{\sum d_{ij}^2}}}$$

(2)

where $d_{ij}$ is the distance between the samples in the original high-dimensional space; ${\widehat{d}}_{ij}$ is the Euclidean distance between samples in the low-dimensional space; stress values range from 0 (perfect fit) to 1 (irregular) and are generally considered usable at <0.2, well ordered at <0.1, and extremely well represented at <0.05.

NMDS can construct a low-dimensional space based on ordinal relationships, which can be used to analyze the spatial differentiation pattern of multidimensional influencing factors of Hangzhou citizens’ willingness to protect biodiversity. Due to the large amount of data in this survey, the relative order of the original distance between the samples can be retained through iterative optimization, which can compress the high-dimensional data such as satisfaction into the low-dimensional space, and directly reflect the degree of difference in the original high-dimensional space through the point spacing. At the same time, the model robustness and anti-interference ability are strong, and the model is not sensitive to outliers or data noise.

3. Results

3.1. Effects of the Respondent Characteristics

3.1.1. Descriptive Characteristics of the Respondents

The basic survey demographics are shown in

Table 3. Basic demographics of residents.

Variable		Sample Size	Percentage
Sex	Male	811	43.02%
	Female	1074	56.98%
Age	18–40	985	52.25%
	40–60	733	38.89%
	≥60	167	8.86%
Education level	Junior high school and below	380	20.16%
	High school or technical school	352	18.67%
	Bachelor or junior college	1083	57.45%
	Master or above	70	3.71%
Occupation	Student	29	1.54%
	Government agent/officer/civil servant	94	4.99%
	Business manager	66	3.50%
	Company employee	328	17.40%
	Professional	22	1.17%
	Manual laborer	59	3.13%
	Commercial employee	196	10.40%
	Self-employed	33	1.75%
	Freelancer	114	6.05%
	Agriculture, forestry, fishery worker	4	0.21%
	Retired	152	8.06%
	Unemployed	11	0.58%
	Other	777	41.22%

. Slightly more than half of the respondents were male and most were 18–60 years old, with an average age of 38.40. Slightly more than half had college, bachelor’s, or higher degrees. Additionally, the respondent occupational backgrounds were diverse and covered a range of industries, with the main occupations being company employees, commercial service industry employees, and others.

3.1.2. Individual Effects of the Respondent Characteristics

(1)

Sex Effects

Sex affected the willingness to invest time, willingness to invest money (p < 0.00), lightly affected the effect of sex on biodiversity (p < 0.01), and did not affect the perception of biodiversity (Figure 2). The willingness of females was higher than that of males in all the cases, which was consistent with the results of a previous study showing that sex affected the WTP and WTA in Tarkeshwar, India [36]. The effect of respondent perceptions and impacts of biodiversity on the willingness to invest time and money decreased sequentially, but the depth of the impact deepened sequentially. The indication was that sex was not a major factor influencing the public perception of biodiversity in urban parks.

(2)

Age Effects

The ANOVA showed that age affected (p < 0.00) all the variables except the willingness to invest time. Specifically, with an increase in age, the willingness to conserve first increased and then decreased. The attention to park biodiversity, preferences, and the WTP showed different trends (Figure 3). For example, the middle-aged and older adults were less concerned about park biodiversity, while the younger age groups were more concerned about the impact of the number and species of animals on the visitor experience and were more willing to invest time and money. The 12–18 year group had the highest values, reflecting their enthusiasm for new outdoor experiences.

(3)

Influence of Education Level

The respondents’ education level affected the four response variables (p < 0.05). Higher education levels were associated with stronger concerns and preferences for park biodiversity and a greater willingness to pay for green space construction and volunteer. This may be related to the greater environmental awareness and sense of social responsibility among the highly educated group. The group with a master’s degree or above did not show the greatest willingness to invest time probably because the sample size of the bachelor’s degree group was two orders of magnitude larger, which may bias the results (Figure 4).

3.1.3. Combined Effects of the Respondent Characteristics

The PERMANOVA showed that demographics strongly affected (p < 0.001) the combined differences of the four response variables, with individual attributes such as age and education contributing particularly to the overall differences (

Table 4. Combined effect of respondent characteristics.

Variable	Df	Sum of Squares	Mean Squares	F Model	r2	p-Value (>F)
Age	3	0.1841	0.046034	4.1998	0.00868	0.001 ***
Sex	1	0.0980	0.098022	8.9428	0.00462	0.001 ***
Education level	3	0.1274	0.042483	3.8758	0.00601	0.004 **
Residuals	1877	20.7930	0.010961		0.98068
Total	1885	21.2026			1.00000

*** = 0.001, ** = 0.01.

). This suggests that social stratification has a multidimensional structural impact on public ecological perceptions. A previous study employing structural equation modeling [17] verified the indirect effects of sex, ethnicity, and family residence and the mediating role of social support, accounting for 3.94% of the total correlation, emphasizing the role of the socio-cultural context in determining individual attitudes through group interactions.

3.2. Effects of Visit Characteristics

3.2.1. Characteristics of Respondent Visits to Parks and Green Spaces

The park visit frequency and distance data were normally distributed, consistent with the general social activity pattern of urban residents. Weekly visitors were the most common (858, 45.52%), followed by daily visitors (23.02%, Figure 5a); 15 min walks were the most common (752, 39.89%), followed by 5 min walks (28.59%, Figure 5b). Females visited parks and greenspaces more frequently than males overall; 18–40-year-old visitors were the most common, followed by 40–60-year-old visitors. The undergraduate population made the most frequent visits, followed by individuals with less than a high school diploma. Among the workforce, company employees were the most frequent visitors.

3.2.2. Respondent Perception of Urban Biodiversity

The public perception of urban biodiversity showed a hierarchical distribution, with species identification being the most important (88.12%), followed by habitat type (60.32%), population size (60.11%), cultural and folklore associations (55.17%), and genetic information (45.78%), reflecting the dominant effect of intuitive taxonomic perceptions on the public’s understanding of biodiversity. These trends were closely related to the spatial accessibility and community structure of urban ecosystems, consistent with the previous studies [35].

The respondents were highly satisfied with the biodiversity (82.49%) as well as the animal species and numbers (89.82%, Figure 6), which should promote their willingness to protect the parks at multiple levels and can inform the development of differentiated conservation policies. Among the ecosystem services, the respondents primarily recognized the landscape esthetic value of biodiversity (81.27%), followed by science education, cultural heritage (51.56%), and environmental biomonitoring (44.85%). They were also concerned about the negative spillover effects of fecal contamination (43.24%) and the health risks associated with the spread of allergens and injuries from animals (43.08%). This suggests that the public’s ecological perception is affected by a combination of utilitarian trade-offs and risk aversion and that a multi-stakeholder cost–benefit analysis should be incorporated into conservation planning.

3.2.3. Respondent Preferences for Urban Biodiversity

The selection of green spaces was mainly affected by their proximity. The main leisure activities were related to fitness, exercise, and relaxation, whereas animal viewing was less common. The respondents were generally satisfied with the plant landscape (63.50% were very satisfied and 24.93% were satisfied). Park selection and usage behavior showed strong function-oriented characteristics. At the spatial decision-making level, the distance to the park had the strongest effect (82.02%), indicating that accessibility is the core geographic factor influencing the frequency of visits. Fitness and exercise (67.67%) and psychological recovery (65.62%) constituted the main public demands, followed by animal viewing (17.82%), reflecting the dominant demand for ecological services related to physical and mental health. Overall satisfaction with park vegetation was high at 88.43% (“very satisfied”: 63.50%, “satisfied”: 24.93%); ornamental plants (with flowers, fruits, and colorful foliage) contributed greatly to the esthetic perception of the landscape (p < 0.01). The satisfaction with (55. 27%) and appropriateness of (34.16%) the tree cover were also high, and too much or too little depression caused negative experiences. Visitors preferred low-risk animals, especially birds (60.58%) and fish (57.35%) owing to their esthetic qualities and safety of interactions, and avoided arthropods, amphibians, mollusks, and mammals (p < 0.05), which were negatively correlated with morphological preferences and the perceived threat. Therefore, UGS planning should seek a balance between biodiversity conservation and public acceptability and improve human–nature interactions through appropriate habitat zoning and design.

3.2.4. NMDS Analysis of Visit Characteristics

The NMDS analysis (stress = 0.156, Bray–Curtis distance, p < 0.001; Figure 7) revealed a broad distribution of multidimensional factors affecting the public’s willingness to protect biodiversity. Plant landscape satisfaction and frequency of visits were the main drivers in the two-dimensional ordination space (explaining 32.4% in NMDS1 and 8.5% in NMDS2), with vector lengths of 0.82 (r² = 0.324, p < 0.001) and 0.30 (r² = 0.085, p < 0.001), respectively, suggesting that they are the core factors affecting the public willingness to protect. Specifically, for every 1-unit increase in plant landscape satisfaction (Likert 5-point scale), the composite index of willingness to conserve increased by 0.38 SD (p = 0.001), whereas the willingness of groups with an average frequency of >8 visits per month decreased by 29.7% compared to the low-frequency groups (F = 6.53, p < 0.001).

The animal taxon visibility (bird richness r² = 0.0045, overall animal r² = 0.0045) was positively correlated with NMDS2 (angle < 15°, p < 0.01) despite the short vector length (0.07–0.09), suggesting that animal interactions had a direct but non-dominant reinforcing effect on the public willingness. Among the spatial and ecological attribute factors, the tree coverage (r² = 0.049, p < 0.001) and residence–green space distance (r² = 0.019, p = 0.055) were distributed along a sub-gradient, reflecting the implicit constraints of ecological quality and accessibility. An increase in cover of 10% increased the WTP by 5.1%, while each 1 km extension in distance reduced the WTP by 3.4%.

NMDS ordination further revealed (

Table 5. NMDS analysis of visit characteristics.

Variable	NMDS1	NMDS2	r2	p-Value (>r)
Perception of urban biodiversity	–0.29093	0.95675	0.0139	0.001 ***
Frequency	–0.95420	–0.29916	0.0849	0.001 ***
Distance	–0.97560	–0.21954	0.0194	0.001 ***
Birds	–0.00358	0.99999	0.0045	0.009 **
Other animals	–0.00358	0.99999	0.0045	0.009 **
Tree coverage	–0.29925	0.95417	0.0492	0.001 ***
Plant landscape satisfaction	–0.56994	0.82169	0.3242	0.001 ***

*** = 0.001, ** = 0.01.

) that the perception of urban biodiversity (r² = 0.014, p < 0.001) was negatively correlated with the main gradient (NMDS1 = –0.29), and that the critical assessment of status quo biodiversity by groups with higher levels of perception may have weakened the satisfaction-driven effect.

3.3. Willingness to Conserve Urban Biodiversity at Multiple Levels

3.3.1. Meaning and Associated Factors of Each Level of Conservation Willingness

The respondent willingness to protect was categorized according to the perceived biodiversity in parks and green spaces, experiential impact, willingness to invest time, and willingness to invest money.

PC1 (perceived urban biodiversity) reflects the respondent perceptions of the richness and diversity of plant and animal communities, such as vascular plant species and the number of terrestrial vertebrates. It was mainly correlated with the complexity of the habitat structure, such as vertical stratification and the ratio of water to terrestrial habitats, which directly affect the level of biodiversity. PC2 (impact of the number and variety of organisms on visitor experiences) reflects the impact of the abundance and visibility of plants and animals on the visitor experience and was mainly correlated with the habitat design, hotspots for animal activity, interactions with visitors, and animal taxa, with birds increasing positive experiences and high reptile densities increasing negative experiences. PC3 (willingness to invest in the protection of urban biodiversity through volunteer hours) reflects the public’s environmental awareness and community involvement and was mainly correlated with education and publicity, the accessibility of environmental organizations, eco-friendly activities, and policy incentives. PC4 (willingness to spend money on biodiversity-friendly products) reflects the economic support for eco-friendly products and was mainly correlated with eco-labeling certification, enhanced consumer trust due to the popularization of science, and policy guidance to promote green consumption.

The PCA revealed the heterogeneous characteristics of the public willingness to conserve biodiversity and its multidimensional mechanism. The first level (PC1 and PC2) integrates cognitive and experiential dimensions, where PC1 focuses on species identification ability and ecological knowledge [29], and PC2 (recreational experience) reflects the influence of animal visibility, which together constitute the cognitive–emotional basis of conservation willingness. The second level (PC3) characterizes the depth of behavioral participation, with volunteer hours as the core indicator, highlighting the key role of community mobilization under the constraints of time and cost. The third level (PC4) reflects the dependence of the WTP on consumption preferences, as evidenced in the premiums added on certified ecological products [44]. This tier division suggests that conservation policies need to be differentiated: the cognitive–emotional tier (PC1–PC2) relies on ecological education to enhance the species visibility, the behavioral tier (PC3) strengthens community self-organization incentives, and the economic tier (PC4) promotes ecological value through market mechanisms.

3.3.2. GLM and Stepwise Regression Analysis of the Willingness to Protect

We clarified the differential mechanisms of the four-dimensional structure of the public willingness to conserve biodiversity (Figure 8). Only plant landscape satisfaction had a strong effect on PC1 (z = 10.57, p < 0.001). For PC2 (the impact of the number and species of animals on the visitor experience), plant landscape satisfaction was also the main driver (z = 5.520, p < 0.001), while the perception of urban biodiversity and the frequency of visits had marginal effects through moderating roles, demonstrating the potential role of a composite mechanism. In PC3, the frequency of visits was positively influenced by volunteer hours (z = 7.745, p < 0.001) and plant landscape satisfaction (z = 2.613, p = 0.009). The tree coverage showed a negative marginal effect (z = –1.848, p = 0.065), suggesting that over depression may inhibit the willingness to participate. The model AIC was optimized from 6750.4 to 6746.6 (8.6% increase in explanatory power). In PC4 (consumption preferences for eco-friendly products, showing spatial and ecological interactions), for every 10% increase in plant coverage, the WTP increased by 4.28% (z = 1.865, p = 0.062); for every 1 km extension of the residence–green space distance, the WTP decreased by 2.63% (z = –1.608, p = 0.108), though this was non-significant.

4. Discussion

This study explored the public’s willingness to conserve urban park biodiversity at multiple levels and the underlying mechanisms using questionnaire surveys. We found that the respondent and visit characteristics had significant effects on the willingness to protect the biodiversity at different levels, and complex interactions were observed among the response variables. This study not only enriches the theoretical system in the field of biodiversity conservation, but also provides new perspectives for understanding the intrinsic mechanisms of public environmental behavior. By revealing the differential effects of individual characteristics such as gender, age, and education on the willingness to conserve, as well as the multidimensional interaction effects between these characteristics, it expands the scope of research in environmental behavior, and reveals the law of group heterogeneity of the public’s willingness to conserve the biodiversity in Hangzhou, as well as its spatial driving mechanism. At the level of practical guidance significance, the results of the study show that plant landscape satisfaction and visit frequency are important factors influencing the public’s willingness to conserve. Therefore, when developing urban green space planning, the plant landscape configuration should be optimized to improve the attractiveness and accessibility of parks. The government can increase the public’s willingness to participate in urban biodiversity conservation by anchoring high explanatory variables to promote the healthy and stable development of urban ecosystems. At the same time, these measures can also help to improve the quality of life and well-being of urban residents and realize the coordinated development of the economy, society, and environment.

4.1. Differential Influence of Visitor Characteristics

Sex had a strong impact on the willingness to invest time and money in conservation, and females were more inclined to participate in volunteer activities, while males were more sensitive to payment behaviors. This is consistent with the group heterogeneity revealed by the PERMANOVA, confirming that the overall variation in public environmental behavior is driven by the dynamic interactions of demographic characteristics, spatial attributes, and socioeconomic attributes and providing an empirical basis for developing differentiated policies.

Age strongly influenced the public perceptions of biodiversity, and the public’s attention to park biodiversity, preferences, and the WTP varied between the age groups. The F-value for age in PC2 (6.837) indicated that age stratification had a strong moderating effect on the ecological experience. This may be related to varying life experiences, values, and social responsibilities in the different age groups [45]. Therefore, policies should carefully consider the impact of age on public perceptions, and differentiated publicity and education strategies should be developed accordingly.

The respondents’ education level had an important influence on their willingness to protect biodiversity. In particular, it was marginally significant for PC1 and PC4, but not for PC3, indicating that more educated groups tend to have a stronger sense of environmental protection and social responsibility, higher level of concern for biodiversity, stronger preference for outdoor activities, and a greater WTP for urban green space construction. Therefore, the government and community leaders/groups should increase the investments in environmental education to improve the public’s awareness and participation.

4.2. Multidimensional Interaction Effects of Visit Characteristics

Through a combined PERMANOVA and NMDS analysis, we revealed the group heterogeneity in terms of the willingness to conserve and its spatial driving mechanism in Hangzhou, China. The PERMANOVA showed that the combined effect of age, sex, and education strongly affected public perceptions, experience, voluntary participation, and payment premiums, suggesting that demographic characteristics had a complex and indirect mediating effect. Using the GLM + stepwise regression analysis, we decoupled the direct effects from the latent variable interactions. The NMDS analysis revealed that plant landscape satisfaction and visit frequency were the main drivers in the downscaled two-dimensional space, consistent with the results of a previous study [46]. The tree coverage and residence–green space distance were closely linked to the sub-gradient distribution. This multidimensional structure suggested that the anchoring effect of these explanatory variables increases citizens’ willingness to participate in urban biodiversity conservation through policy design that promotes accessibility and social equity; differentiates conservation strategies according to sex, age, and occupation; and creates eco-friendly consumerism through third-party certification, scenario-embedded marketing, production chain traceability, and carbon footprint reporting [37].

By employing such a multidimensional strategy, the Hangzhou government can gradually adopt “active co-creation” policies over “passive protection” and develop a biodiversity-based governance model that considers the relevant scientific and social perspectives.

4.3. Driven by the Willingness to Protect and Co-Creation

Our study revealed a multidimensional mechanism underlying the willingness to conserve and the mediating interactions between drivers, and proposes a systematic governance framework focused on “eco-social co-benefits”. Satisfaction with the botanical landscape was the core driver associated with “perception-experience-commitment”, and high-frequency visits indirectly promoted the length of volunteer hours. PC1, reflecting the satisfaction and visit frequency, revealed a mechanism of “aesthetic perception driving behavioral transformation”, PC2 emphasized the basic supporting role of ecological quality (tree coverage and perception of urban biodiversity), and PC3 and PC4 highlight the constraints of payment thresholds, which are closely linked to the synergy between the market and policy instruments [28].

The synergistic influence of PC1 and PC2 explained 52% of the public’s willingness to protect the biodiversity. This provides empirical evidence for the potential role of megacities in resolving the nonlinear relationship between ecological inputs and social benefits. To achieve the “total co-benefits” of ecological protection and social wellbeing, the Hangzhou government can upgrade ecologically minded parks to improve their social value based on the relevant socioeconomic and spatial drivers and their interactions.

4.4. Limitations and Prospects

Although this study provides various theoretical and practical advances, it has some limitations. For example, the sample size was relatively limited and mainly concentrated on Hangzhou. The questionnaire method may introduce some subjectivity bias and distort our statistical inferences. Future studies should expand the sample size and geographic as well as cultural scopes, and adopt more objective data collection methods. Simultaneously, a dynamic “perception-behavior-ecological response” model that combines multi-source data, such as remote sensing imagery, social media analysis, and cultural backgrounds, among others, should be developed to more comprehensively assess the effects of public perceptions on urban greenspace planning and sustainable development policies.