Determinants of Public Participation in Watershed Management in Southeast China: An Application of the Institutional Analysis and Development Framework

Abstract

Increasingly, adaptive processes and decentralization are vital aspects of watershed governance. Equitable and sustainable water governance requires an understanding that different societal members have unique relationships with the environment and varying levels of interaction with policymakers. However, the factors facilitating public involvement under centralized governance remain less understood. This study combined the Institutional Analysis and Development framework with ordered probit regression to empirically investigate the determinants of willingness to participate (WTP) and actual participation of the public in integrated watershed management (IWM). Data from 933 valid questionnaires collected across 36 counties in Fujian, China, were used to define stakeholders’ perceptions of IWM. Results show that stakeholders are predominantly willing to participate in watershed conservation, management, or planning (85.9%), while only 32.8% frequently attend related events. Pro-environmental intentions were mainly shaped by interactional capacity—information exposure, interpersonal exchanges, and cross-reach support recognition—while actual participation was influenced by perceived biophysical conditions, rules-in-use, socioeconomic factors, and interactional capacity. Frequent observations of poor forest management practices were correlated with higher behavioral intentions, and socioeconomic dynamics significantly affected self-reported actual participation. Information sharing had the most substantial positive impact on both WTP and actual participation. These findings reinforce the necessity for an integrated and holistic approach to regional watershed resource management that fosters inclusivity and sustainability. This study provides workable insights into the social and institutional factors that shape public participation in watershed governance as it evolves toward decentralization.

1. Introduction

Humans throughout history have used various methods to control the level and flow of water resources. A watershed is a geographically defined area where all precipitation collects and drains to a common water body or low point due to its topography. A watershed functions as an integrated ecosystem characterized by the flow of materials, energy, and information flow. It is a key geographic unit for sustainable resource management worldwide [1,2], even if management boundaries do not always coincide with watershed boundaries due to differing objectives (ecological-centered or anthropogenic-centered boundaries) and administrative restrictions. Planners increasingly use watersheds as the basic unit to organize natural resources and manage anthropogenic activities. Since the 1992 UN Earth Summit, there has been a shift toward promoting a decentralized and holistic Integrated Watershed Management (IWM) approach that effectively organizes watershed resources in a locally responsive context. This trend promotes sustainability by incorporating biological conservation into socioeconomic development [3,4]. Wang et al. [5] described IWM as a multidisciplinary, combined, and adaptive approach to managing watershed activities for social, economic, and environmental sustainability. However, the success of IWM requires the active participation of the relevant stakeholders, including policymakers, urban planners, industries, commercial sectors, the scientific community, and civilians, both within and outside the watershed [6,7,8].

The value of IWM is increasingly recognized in densely populated regions, making it a feasible approach for populous, rapidly urbanizing countries like China [9,10]. Widely applied in water pollution control, IWM optimizes the spatial and temporal allocation of pollutants and the rational distribution of discharge permits [11,12]. From 1988 to 2010, China invested over 30 billion yuan in scientific and technological projects for water pollution control, advancing technical capabilities in pollutant detection, treatment, and monitoring. However, social factors (e.g., participation) and management mechanisms (e.g., the relationship between river basin and jurisdictional management, coordination, and cooperation mechanisms) were ignored in the previous IWM design. The revised 2002 Water Law states that the forms of public participation in water management includes providing access to environmental information and empowering the public to report violations for civil compensation [13]. However, the relevant clauses are more principled and framework-oriented, lacking clear guidelines on participation methods and legal consequences [14].

When considering China’s relatively centralized power structure, IWM planning could be prone to bureaucratization and the over-centralization of decision-making institutions [3]. Public participatory processes largely remain undeveloped in China’s urban planning and regional conservation projects [15]. Citizen perceptions do not significantly affect decision-making; rather, the consultation process relies on instrumental rationales for the smooth implementation of specific projects [16]. From the perspective of participating entities, the government-led centralized governance model cannot effectively respond to public demands [17]. Due to limitations in past procedural regulations, the traditional “command-and-control” management has offered few entrances for public participation, resulting in insufficient interactions and the inability to absorb the demands of other stakeholders [18,19]. This lack of meaningful public participation may undermine the public’s trust in the government, which obscures policy implementation and watershed management [20]. Encountering resistance and doubts in law enforcement and judicial decisions poses challenges to sustainable watershed governance [21]. In this regard, there is a need for enhanced public participation and a bottom-up approach to IWM to avoid tokenism or deregulation and secure its effective implementation.

Establishing a dynamic planning process and interaction between citizens and policymakers is crucial to the effectiveness of IWM [22]. Several water resources management studies also recommend enhancing stakeholder cooperation and public participation to improve the monitoring network [23,24,25,26,27]. This study aims to revisit and deepen our understanding of public awareness and perceptions concerning IWM within the Min River Basin. Its primary goal is to gather and assess socioeconomic data and public opinions pertaining to various factors, including the biophysical state, policy implementation, forest management, and watershed administration.

The geographic, temporal, social, and political complexity of watershed management requires a systematic framework for understanding. Ostrom’s institutional analysis and development framework (IAD) was employed to observe how institutional operations influence the participation of respondents, and vice versa, in the socioecological context. The IAD framework is robust for examining social and political dynamics relevant to environmental issues, including collaborative watershed management [28,29]. Local perceptions toward and connections within watershed resources are key factors influencing outcomes. Kadirbeyoglu and Özertan [30] utilized perception data to characterize the performance of common-pool resource governance. Socio-environmental perception surveys are common information sources for policymakers to obtain an integrated view of watershed conditions and public priorities [31]. Additionally, perception-based data are useful for investigating the effectiveness of policy implementation and the influence of institutions on actors [32]. The enhancement of information sharing and monitoring mechanisms can improve government credibility and, thus, increases citizens’ trust in management efforts [33].

IWM is a political process critically dependent upon the effective governance of watershed biophysical elements. Current watershed management studies mainly center on the effects of human activities on ecological conditions or examining the impacts of watershed degradation on a single factor at the local community level [34,35,36], often overlooking the broader, interconnected dynamics within watershed ecosystems. Moreover, these studies do not provide a systematic framework to holistically examine the influence of institutional, socioeconomic, and biophysical factors on individual pro-environmental actions. In this paper, we aim to reveal the interactions and set priorities for IWM from a socioecological perspective. We explore how contextual factors, including individuals’ perceptions of biophysical conditions, rules-in-use, socioeconomic characteristics, and interactional capacity, influence participation in IWM in the Min River Basin, Fujian, China.

2. Theoretical Framework: Institutional Analysis and Development (IAD) Framework

The institutional analysis and development (IAD) framework proposed by Elinor Ostrom is instrumental in analyzing the management of common-pool resources like woodlands and water sources, where institutional rules, policies, and governance affect actor behaviors ([33], p. 376). It summarizes how different factors shape individual and community actions as well as the process of institution formation. The core is the action arena, which includes the action situation and actors. Exogenous factors, including biophysical conditions, community attributes (socioeconomic characteristics), and rules-in-use (institutional, policy, and governance), together comprise the current action arena.

The IAD framework has been effectively applied to multiple management dilemmas central to common-pool resource sustainability. Researchers have adopted it to understand the impact of actor interactions on decentralized irrigation management in Turkey [30], estimate voluntary waste disposal monitoring in China [37], and analyze local farmer responses to water deficits [38]. At a broader scale, the framework supports watershed natural resource conservation, large restoration projects, and wastewater governance by identifying gaps in institutions and governance [39,40]. Comparative studies of the Yellow and Ganges Rivers further demonstrated its utility in understanding centralized governance narratives of large rivers [41].

The IAD framework guides the research from organizing our literature review and survey design (including socioeconomic characteristics, perceptions of biophysical conditions, the performance of formal institutions, and indicators for informal institution formation—willingness to pay, participate, and actual participation) to model variable selection. We used the IAD framework as a conceptual framework to organize independent variables and then analyzed individual participation decisions in watershed governance for the Min River (Figure 1). By determining the factors influencing the probability of respondents’ involvement, management priorities can be set to maximize public participation. Participatory processes and decentralization of IWM can be prioritized to ensure the long-term sustainability of this approach.

High levels of expressed willingness, commitment to meetings, and engagement in monitoring networks suggest strong potential for institutional development. The IAD framework also considers informal institutions and local-level interactions, making it well-suited to capture how local actors engage with top-down policies within China’s hierarchical governance. It highlights how local actors, even within a centralized system, can modify or impact formal rules through their actions despite restrictions on public participation.

2.1. Factors Affecting Public Participation

The two dependent variables in this study were the actual participation and willingness to participate in the participatory processes of IWM. These processes included any fora, meetings, and activities related to watershed conservation, management, project decision-making, and planning led by local authorities.

Independent Variables

Biophysical conditions. Two variables, perceived environmental quality and future concerns, were adopted to reflect the participants’ perception of the biophysical conditions of the Min River Basin. Perceived environmental quality is affected by the public perception of recent changes in environmental conditions [37]. Based on the most serious and concerning watershed issues identified by participants, a composite variable that summarizes future concerns about water pollution, flooding, and water deficit was developed.

Socioeconomic characteristics. Participation willingness and behavior are influenced by individuals’ satisfaction with management efforts, occupation and income, knowledge, attitudes, perception, and vision of the watershed [42]. The impacts of demographic attributes and socioeconomic factors on direct participation behavior in environmental governance have been recognized [43]. Variables for the respondents’ gender, age, education level, annual income, and length of residence were selected in this study.

Additionally, we assessed perceptions of the watershed and forest value, asking respondents to evaluate the MRB’s significance to Fujian Province’s development and the ecosystem services forests provide, including recreation, economic value/employment, wildlife habitats, global ecosystem balance, water, air/soil regulation, and natural environment protection (

Table 1. The variable descriptions and composite details.

Category/Variable	Variable Definition and Measurement	Composite or Index
Dependent Variable
Willingness to participate	Willingness to participate in watershed-related events (1 = not willing, 2 = unsure, 3 = willing)	Composite
Actual participation frequency	Frequency of watershed-related event participation (1 = seldom, 2 = sometimes, 3 = often)	Composite
Independent Variable
Biophysical Conditions
Perceived environmental quality	Rate the improvement of the MRB environment quality (1 = significantly improved, 2 = improved, 3 = no change, 4 = worsened, 5 = significantly worsened)	Composite
Future concerns	Respondents’ concern about the future issues of water pollution, flooding, and water deficit in the MRB	Indexed
Rules-in-use
Integrated watershed management	Public overall perception of government management (1 = very good, 2 = good, 3 = neutral, 4 = poor, 5 = very poor)
Forest management	Indexed scale of public satisfaction with MRB forest management (lower index indicates less satisfaction)	Indexed
Effectiveness of government sector	Indexed scale of respondents’ satisfaction with effectiveness of various government departments (lower index indicates less satisfaction)	Indexed
Watershed management and operation	Indexed scale of respondents’ satisfaction with various prospects of MRB management (lower index indicates less satisfaction)	Indexed
Presence of poor management practices	Index of witness frequency of poor MRB management behaviors (lower index indicates less frequent)	Indexed
Socioeconomic Characteristics
Gender	1 = female, 2 = male
Age group	1 = below 18, 2 = 20 s, 3 = 30 s, 4 = 40 s, 5 = 50 s, 6 = 60 s
Education level	1 = elementary school, 2 = middle school, 3 = high school, 4 = college, 5 = Bachelor’s degree, 6 = Master’s degree, 7 = PhD
Annual income range	Annual income ranges from 36,000 to 300,000 divided into ten categories (in CNY)
Length of residence	Length of residence in MRB (years) (1 = ≥ 30, 2 = 21–30, 3 = 11–20, 4 = 6–10, 5 = 1–5, 6 = ≤ 1)
Watershed importance cognition	Public perception of level of importance of MRB to Fujian Province (1 = extremely important, 2 = important, 3 = neutral, 4 = unimportant, 5 = extremely unimportant)
Forest value	Indexed scale of public perception of importance of MRB (lower index indicates less important)	Indexed
Interactional Capacity
Interpersonal communication	Willingness to encourage others to participate in MRB-related events (1 = willing, 2 = unsure, 3 = not willing)
Cross-reach support	Public opinion of cross-reach support in terms of economics and technology (lower index indicates less willing to support other reaches)	Indexed
Information sharing	Public perception of the government advertisement of MRB-related information (1 = never, 2 = seldom, 3 = sometimes, 4 = frequently, 5 = always)	Indexed

).

Rules-in-use. This study interpreted “rules-in-use” based on stakeholders’ satisfaction with IWM. These indices acted as practical indicators of “rules-in-use”, reflecting the formal rules and the respondents’ direct experiences with them. People’s actual experiences and satisfaction are critical measures of governance effectiveness and are considered when evaluating governance rules [44]. In scenarios involving air pollution, water, and waste management, public surveillance, and frequency of monitoring informally establish community standards [37,45]. Satisfaction and effectiveness indices, therefore, offer a measurable representation of the IAD’s conceptualization of “rules-in-use” in the specific action situation of MRB governance.

Whether or not a regulation’s intended purposes can measure the effectiveness of management in environmental regulation [45] aligns with “perceived effectiveness” as it encompasses stakeholders’ assessment of whether regulatory objectives are being met. Stakeholder satisfaction is a critical insight into how well the governance process functions [46]. Ostroff [47] also noted a strong correlation between individual satisfaction and organizational performance. Thus, we incorporated the assumption that stakeholders’ satisfaction and perceived high management effectiveness were reliable indicators of rule functionality. We evaluated satisfaction with various aspects of IWM, covering overall IWM performance, watershed management and operations, perceived effectiveness of government sectors, forest management, and the observed frequency of poor forest practices.

The composite index of watershed management and operations aggregated individual responses on the perceived performance of governance, including planning and decision-making, investment, public involvement, education and publicity, environmental protection and supervision, law enforcement and legislation, cooperation among government sectors, and cross-reach cooperation. Another composite index measured perceived effectiveness across government sectors, incorporating a range of specific management actions and outcomes. We also investigated the effectiveness and appropriateness of forest management, assessing respondents’ satisfaction with various aspects, including silvicultural practices, forest resource management, stand and species diversity, fertilization, wildlife protection, fire protection, logging operations, riparian protection, and pest control. The last composite index examined the presence of poor forest practices by aggregating the observed frequencies of various activities, including burning slash and shrubs, soil preparation, large-scale clear-cutting, monocultures, logging within riparian zones, herbicide and pesticide use, logging on steep slopes, and natural forest harvesting.

Interactional capacity. The first variable, interpersonal communication, focused on information exchange within the community. Respondents evaluated their willingness to spread IWM-related information and encourage engagement in IWM-related activities among friends, families, colleagues, or classmates. The second variable, cross-reach support, examined the respondents’ perceived necessity of coordinating technical and financial resources between the upper and lower reaches. The last indicator of interactional capacity was information sharing, which measured respondents’ exposure to discussions, meetings, and activities associated with watershed conservation, management, and planning; it indicated the interaction level between institutions and the public (individuals within communities).

3. Materials and Methods

3.1. Study Area

The Min River is the largest river in Fujian Province, China; it has a length of 580 km and an average annual flow of 62.9 billion m³. It is located between 25°23′ and 28°19′ N and 116°23′ and 119°43′ E [48]. The river originates in the Wuyi Mountains and flows southeast to the East Sea. Its catchment area is 60,992 km², occupying more than half of the total area of Fujian Province. The MRB covers three administrative cities (Nanping, Sanming, and Fuzhou), including 36 counties in the province. Around 12.95 million people live in this area, accounting for 30.9% of the total provincial population [49].

In the last 18 years, the total gross domestic product (GDP) of the MRB has increased dramatically, from 32.49 billion to 1.68 trillion USD [49,50]. The central part of the GDP has shifted from primary industries (e.g., agriculture) to secondary industries (e.g., manufacturing), then to tertiary industries (e.g., the service industry) in the past decade. The basin’s natural resources support agriculture, industry, and cultural heritage and are crucial for the livelihoods of millions. It provides valuable ecosystem services, such as the provision of materials (timber, water), water purification, and carbon sequestration services.

MRB’s rapid economic development has led to the emergence of environmental externalities like water pollution, invasive aquatic species (e.g., water hyacinth booming), flooding, and soil erosion. These challenges have raised concerns about long-term ecological resilience and socioeconomic security among local authorities and residents.

3.2. Questionnaire Design

The questionnaire distributed in 2021 was an adapted version of the Integrated Watershed Management: Questionnaire Survey on Sustainable Management in the Min River Basin developed in 2006 [51]. The 2021 questionnaire investigated the public’s awareness, concerns, and understanding of the local environmental conditions and sustainable watershed management in the MRB and collected comments and suggestions for further development in the area. Given the high forest coverage (76.8%) and soil erosion concerns, proper forest management is crucial for watershed health. Accordingly, the survey includes a section on forest management, investigating perceptions of forest value and forest management practices, such as wildfire prevention, logging, and riparian zone protection. This echoes China’s national emphasis on natural conservation, viewing forests as reservoirs for water (purification, conservation, drought replenishment), sources of grains (timber, bamboo, non-timber products), and carbon sinks. This community-based approach is vital to developing governance models incorporating public participation, aligning with China’s transition towards more inclusive, stakeholder-responsive resource management.

Environmental conditions and socioeconomic development vary significantly across the lower, middle, and upper reaches of the basin. The upper reach features mountains with steep slopes and fast-flowing waters, while the lower reach consists of floodplains with higher population density. Given these differences, the questionnaire captured unique challenges and opportunities across these regions, including the capacity for cross-reach collaboration. The MRB’s mountainous, forested landscape is intertwined with agriculture and other production activities that directly impact the watershed and are closely tied to local livelihoods. This survey was initially created with these socio-ecological linkages in mind. The collected data help inform governance strategies for sustainable management, ensuring the provision of tangible goods and services to support the basin’s socioeconomic development.

Questions about annual income and education level were adjusted to reflect economic, social, and technological development over the past 15 years. A draft of the questionnaire was created and pre-tested by distributing around 30 copies to local stakeholders at Fujian Agriculture and Forestry University. This pre-test helped identify and resolve issues such as ambiguous wording, errata, or insufficient question information. The estimated average completion time for the survey was approximately 15–20 min. The final version included 38 questions organized into the following six sections: personal information, present state and issues, forest management, government role, public awareness/willingness/participation, and comments and suggestions. The detailed questionnaire information is provided in

Table 2. Structure and components of questionnaire.

Section	Questions	Item Content
Personal information	12 questions (filling-in, single-choice, multiple-choice, open-ended)	Background information on respondents, relationship with sustainable watershed management, general views on MRB
Present state and issues	9 questions (single-choice, multiple-choice, ranking, open-ended)	Perception and concerns about watershed states, changes, and issues
Forest management	3 questions (Likert scale)	Perspectives on forest values and forest management activities
Government role	4 questions (Likert scale, ranking, single-choice)	Extent of satisfaction with local government’s management of MRB
Public awareness, willingness, and participation	6 questions (single-choice, filling-in, open-ended)	Extent of WTP and willingness to financially support watershed conservation and management
Comments and suggestions	4 questions (open-ended, single-choice)	Suggestions for sustainable watershed management of MRB in the future
Total	38 questions

.

3.3. Data Collection and Analysis

The survey was conducted physically, with participants choosing to fill out the survey in either a digital or paper format. In total, 1203 questionnaires were collected, including 700 completed online and 503 in the paper format. Two-stage sampling was adopted to apply both cluster and random stratified sampling. Based on the population distribution and density, geographic location, and socioeconomic condition data from the Fujian Provincial Statistical Yearbook [52] and proximity to the MRB, 14 counties were selected in the cluster sampling stage. Six counties were chosen each from the upstream and downstream areas, and two were chosen from the midstream area. We sought to investigate the perceptions and opinions of the public and those involved in managing the MRB. To ensure broad sampling, three high-population-mobility public facilities (city squares, community centers, and libraries), areas near parks and rivers, and watershed-related government departments (hydrology, forestry, agriculture, ecology, environment, and natural resources) at different jurisdictional levels were randomly selected as survey sites. Simple street interviews were conducted at city squares, community centers, government institutions, and libraries, gathering responses from 10 to 15 pedestrians in 2021.

Of the 1203 questionnaires received, 933 were deemed valid based on their response time and the completion of the answers. The cleaned sample size provides a 95% confidence level with a 3.2% margin of error, sufficient for social science analysis [53,54]. Considering those who had to redo the questionnaire due to internet errors, interruptions, or any other incidents, data from participants who spent less than 3 min answering the questions were excluded from the data analysis. Approximately 44.6% and 35.3% of valid responses came from the lower and upper reaches, respectively, with a smaller portion from the midstream and other regions.

In the initial exploratory data analysis, different methods were used to assess the internal consistency of the responses and the quality of the questionnaire’s structural validity. At a value higher than 0.9 for each individual section, Cronbach’s α results confirmed internal consistency across the questionnaire [55]. Further exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) supported the questionnaire’s strong structural validity.

Data were first imported and tabulated using Excel before SPSS (Version 27) was applied for exploratory data analysis. R (version 4.1.2) and STATA 17 software were also used for visualization and analytical purposes. As previously indicated, the questionnaire data related to the respondents’ perceptions of watershed-related events (WTP and actual participation frequency) as per the respondents’ biophysical conditions, rules-in-use, socioeconomic characteristics, and interactional capacity (Table 1).

At the preliminary stage, multiple questions were selected and combined from the original survey data based on previous research. For combined features, such as forest management and information sharing, the mean imputation method was applied to any missing values to reduce bias, as these features are a combination of several Likert scale questions. While mean imputation was chosen for its simplicity and to retain as many data points as possible, this method assumes that the missing data are random, which may introduce some bias. The “future concerns” feature was created as a composite of three binary questions, with missing responses assumed as “no concern” (zero). After combining and indexing each feature, descriptive analyses provided an overview of the independent variables. The Bonferroni correction was then applied to multiple comparisons, evaluating respondents’ WTP and actual participation in MRB-related events.

Finally, sensitivity analyses were carried out using three different methods to investigate the associations between various factors and the respondents’ participation and perceptions. Independent factors were grouped into four categories based on the adapted IAD framework, including biophysical conditions, rules-in-use, socioeconomic characteristics, and interactional capacity (Figure 1). Spearman’s rank correlation coefficient was calculated for both participation variables against every factor to identify any linear relationships.

Recursive Feature Elimination (RFE) was applied across three models (random forest, naïve Bayes, and bagged tree), with 10-fold cross-validation repeated five times to validate each variable’s significance. RFE was chosen to enhance model accuracy by removing the least important variables, thereby reducing the potential for overfitting. Through analyzing accuracy changes across different numbers of optimal features, the best combination of features was identified to achieve the optimal balance between bias and variance trade-offs. The bagged tree and random forest models minimized data overfitting, while the naïve Bayes model accounted for the unbalanced participation variable. No significant discrepancies were found among the results of the three models.

Following feature selection, ordered probit regression was conducted for each variable to calculate marginal effects, assessing each independent variable’s impact on participation likelihood. Model assumptions were tested for validity. The proportional odds assumption was confirmed with the Brant test, indicating consistent effects across all categories of the dependent variable. Variance Inflation Factor (VIF) scores, all below 3, revealed no significant multicollinearity, consistent with the results from Spearman’s correlation. The regression indicates the extent to which each independent factor increases or decreases the likelihood of public participation in MRB-related events. Each step contributes to a detailed, bias-controlled analysis to identify the determinants of public participation in watershed governance.

4. Results

4.1. Overview

Section 2 of the questionnaire revolved around the current situation and changes in MRB biophysical conditions. More than 88% of the participants stated that the MRB ecosystem had improved over the past 15 years. Considering watershed-related disasters individually, respondents’ concerns and observed improvements from 2006 to 2021, as well as their concerns for the future, were recorded. Among all disasters, flooding, water pollution, and soil erosion were the top three concerns, both in the past and the future.

4.2. Public Attitudes Toward Forest Management and the Government’s Role

In the third and fourth sections of the survey, participants valued the provisioning, supporting, and regulating services of forests while placing less value on recreational and economic benefits. Satisfaction with nine forest management practices was measured on a 5-point Likert scale, with responses mostly ranging from neutral to somewhat satisfied. Exceptions were noted for wildlife and fire protection, where satisfaction was higher; somewhat satisfied participants outnumbered neutral responses by about 10%. Dissatisfaction was noted primarily with riparian protection, although those satisfied still exceeded those dissatisfied. Over half seldom or never observed poor forest management, with only 4% frequently noting issues in natural forest harvesting.

The survey then evaluated government roles in watershed management, finding that most participants were either neutral or somewhat satisfied with government efforts, excluding funding and public awareness, where opinion leaned more towards neutrality. Integrated management planning was the most needed government action, whereas enhancing cooperation across reaches was considered the least needed. A majority (67.2%) were content with the current government management, with dissatisfaction below 2%.

4.3. Public Awareness

With 795 votes, water provisioning services were considered the most important value of the watershed. There is a trend of rising public awareness regarding supporting and regulating services. Notably, more than half of the participants rated natural environment and resource protection, maintaining an ecosystem network, and wildlife habitat conservation as extremely important.

The focus of the last section of the questionnaire was on public awareness. We examined the respondents’ awareness of singular events, news, and willingness to protect the Min River environment. Despite high awareness, with 87% informed on related news, only 65% had participated in protective actions. However, 78% were willing to promote participation to others, suggesting a preference for collective action over individual initiatives. On resource allocation, a substantial majority agreed that the more affluent lower reach should aid the upper reach, with 92% and 87.5% backing technological and financial support, respectively. However, regarding personal financial contributions, only 48.5% were willing to donate, with 36% uncertain about donating. Among those uncertain, the primary concerns were the lack of transparency regarding the use of funds, the impact on their living expenses, and the belief that the government or those benefiting from the environment should bear protection costs.

4.4. Statistical Analysis

Four main variable categories were chosen to investigate the respondents’ WTP and actual participation regarding the MRB (

Table 3. Sample descriptive statistics.

Category/Variable	Observations	Mean	Std. Dev.	Min	Max
Dependent Variable
Willingness to participate	931	2.847	0.391	1	3
Actual participation frequency	929	1.989	0.817	1	3
Independent Variable
Biophysical Conditions
Perceived environmental quality	933	2.02	0.823	1	5
Future concerns	933	1.436	0.831	0	3
Rules-in-use
Integrated watershed management	911	2.193	0.719	1	5
Forest management	933	33.135	6.654	9	45
Effectiveness of the government sector	933	42.805	9.552	12	60
Watershed management and operation	933	28.389	6.688	8	40
Presence of poor management practices	933	19.884	6.922	8	39
Socioeconomic Characteristics
Gender	924	1.397	0.49	1	2
Age group	932	3.286	1.15	1	6
Education level	931	4.987	0.858	1	7
Annual income range	928	3.159	1.976	1	10
Length of residence	924	2.522	1.653	1	6
Watershed importance cognition	933	1.205	0.454	1	4
Forest value	933	25.202	4.881	6	30
Interactional Capacity
Interpersonal communication	928	2.77	0.462	1	3
Cross-reach support	933	5.737	0.705	2	6
Information sharing	928	3.541	1.004	1	5

). Many participants preferred not to respond to certain questions. The missing values in the data were treated as missing, and only complete observations were included in the model estimation. Overall, approximately 85.9% of participants expressed a willingness to participate in discussion meetings or activities related to watershed conservation, management, or planning organized by the local government (800 willing, 120 unsure, 11 unwilling). However, only 32.8% of participants actively engage in such activities frequently (315 often, 309 sometimes, 315 seldom). The lower reach shows a slightly higher WTP (61.9% vs. 57.8%) and significantly higher actual participation (47.1% vs. 29.4%) than the upper reach.

Multiple comparison tests with the Bonferroni adjustment were conducted to determine the relationship between the dependent and independent variables. As shown in

Table 4. Mean differences in WTP and actual participation frequency.

	Willingness to Participate			Actual Participation Frequency
	MD (US-UW)	MD (US-W)	MD (UW-W)	MD (O-S)	MD (O-ST)	MD (S-ST)
Dependent Variable
Biophysical Conditions
Perceived environmental quality	0.08	0.38 ***	0.3	−0.58 ***	−0.43 ***	0.15 *
Future concerns	0.11	−0.07	−0.18	−0.01	0.01	0.02
Rules-in-use
Integrated watershed management	0.36	0.54 ***	0.18	−0.52 ***	−0.39 ***	0.13 **
Forest management	−0.69	−2.53 ***	−1.84	3.91 ***	2.92 ***	−0.99
Effectiveness of the government sector	−0.27	−4.73 ***	−4.46	6.96 ***	4.93 ***	−2.03 **
Watershed management and operation	1.01	−2.57 ***	−3.58	4.8 ***	3.46 ***	−1.34 **
Presence of poor management practices	5.1 *	0.5	−4.6 *	1.48 **	1.4 **	−0.08
Socioeconomic Characteristics
Gender	−0.11	−0.06	0.05	−0.13 ***	−0.06	0.07
Age group	0.17	−0.14	−0.31	−0.11	−0.17	−0.06
Education level	0.03	−0.06	−0.09	−0.04	−0.19 **	−0.15 *
Annual income range	0.37	−0.4	−0.77	0.45 **	0.03	−0.42 **
Length of residence	−0.09	0.32 ***	0.41	−0.14	0.07	0.21
Watershed importance cognition	−0.35 **	0.21 ***	0.56 ***	−0.29	−0.19	0.1
Forest value	0.64	−0.66	−1.3	−0.38	0	0.38
Interactional Capacity
Interpersonal communication	0.55 ***	−0.47 ***	−1.02 ***	0.11 ***	0.07	−0.04
Cross-reach support	0.47 *	−0.42 ***	−0.89 ***	−0.01	0	0.01
Information sharing	0.24 **	−0.35 ***	−0.59 ***	1.53 ***	0.76 ***	−0.77 ***

Notes: *, **, and *** indicate statistical significance at 10%, 5%, and 1% levels, respectively. MD denotes the mean difference between two groups. Willingness to Participate (UW: unwilling, US: unsure, W: willing). Actual Participation Frequency (S: seldom, ST: sometimes, O: often). Bonferroni adjustment was made for multiple comparisons.

(with mean statistics in Appendix A

Table A1. Mean statistics for WTP and actual participation frequency.

	Willingness to Participate			Actual Participation Frequency
	Unwilling	Unsure	Willing	Seldom	Sometimes	Often
Dependent Variable
Biophysical Conditions
Perceived environmental quality	2.27	2.35	1.97	2.26	2.11	1.68
Future concerns	1.27	1.38	1.45	1.45	1.43	1.44
Rules-in-use
Integrated watershed management	2.3	2.66	2.12	2.41	2.28	1.89
Riparian forest management	31.64	30.95	33.48	31.56	32.55	35.47
Effectiveness of the government sector	39	38.73	43.46	39.9	41.93	46.86
Watershed management and operation	25.18	26.19	28.76	26.4	27.74	31.2
Presence of poor management practices	15.27	20.37	19.87	19.37	19.45	20.85
Socioeconomic characteristics
Gender	1.45	1.34	1.4	1.46	1.39	1.33
Age group	3	3.17	3.31	3.30	3.36	3.19
Education level	4.91	4.94	5	4.95	5.10	4.91
Annual income range	2.45	2.82	3.22	2.87	3.29	3.32
Length of residence	2.73	2.85	2.47	2.64	2.43	2.50
Watershed importance cognition	1.73	1.38	1.17	1.24	1.19	1.18
Riparian forest value	24	24.64	25.3	25.48	25.10	25.1
Interactional Capacity
Interpersonal communication	1.82	2.37	2.84	2.72	2.76	2.83
Cross-reach support	4.91	5.38	5.8	5.74	5.73	5.73
Information sharing	2.27	3.05	3.64	2.78	3.55	4.31

), no significant correlation was found between the future concern, age group, forest value and any of the dependent variables. In contrast, perceived environmental quality (as a biophysical condition) had a significant relationship with the actual participation frequency but little association with WTP.

Participants noticing some improvement in the MRB environment were more willing to participate in related events but were less willing to contribute financially. The four rules-in-use factors significantly affected event participation rates but had little to no influence on WTP. Interestingly, rules-in-use only made a difference in distinguishing between unsure participants and those willing to participate in MRB-related events. Only the perception of watershed importance significantly influenced WTP, with other factors not showing statistical significance. Socioeconomic characteristics did not affect participation willingness in MRB events. Among all examined factors, interaction capacity played a key role in affecting the respondents’ readiness to contribute to the Min watershed.

We conducted sensitivity analyses to further analyze the relationships between the respondents’ WTP and actual participation frequency and the independent variables precisely to determine each factor’s impact on the dependent variables. According to the Spearman rank correlation coefficient results (see Appendix A

Table A2. Spearman’s rank correlation.

	Willingness to Participate	Actual Participation Frequency
Dependent Variable
Biophysical Conditions
Perceived environmental quality	−0.179	−0.322
Future concerns	0.023	−0.023
Rules-in-use
Integrated watershed management	−0.248	−0.297
Riparian forest management	0.151	0.28
Effectiveness of the government sector	0.178	0.318
Watershed management and operation	0.163	0.322
Presence of poor management practices	−0.006	0.058
Socioeconomic Characteristics
Gender	0.032	−0.116
Age group	0.05	−0.012
Education level	0.017	−0.04
Annual income range	0.082	0.095
Length of residence	−0.073	−0.027
Watershed importance cognition	−0.167	−0.055
Riparian forest value	0.111	0.048
Interactional Capacity
Interpersonal communication	0.389	0.101
Cross-reach support	0.217	0.011
Information sharing	0.24	0.635

), most factors were weakly correlated (ρ < 0.1) with the variables of interest. Consistent with our previous findings, socioeconomic characteristics had the lowest association, whereas interactional capacity had the highest. The strongest correlation was between actual participation and information sharing (ρ = 0.64). All but one of the rules-in-use factors were somewhat associated with the two participation scales.

Among the three algorithms we used during RFE, the bagged tree and random forest models consistently produced the same important features, which tended to differ significantly from the naïve Bayes model results. This may have been caused by an imbalance in the WTP and actual participation frequency variables. Therefore, a factor was considered significant in influencing the dependent variable only if it was identified as important by both the naïve Bayes model and one of the tree models.

For WTP, interpersonal communication and information sharing emerged as the most and second-most important features. Meanwhile, actual participation frequency was most strongly associated with three features: information sharing, environment quality, and satisfaction with watershed management and operations. Interactional capacity was the key determinant of WTP, followed by rules-in-use for both WTP and actual participation in MRB events. Conversely, socioeconomic and biophysical factors showed weak associations with WTP and actual participation frequency. One particular feature from each section was highly influential: forest value. Among the biophysical conditions, perceived environmental quality was the second-most influential in actual participation frequency.

Next, we used an ordered probit regression model to calculate the marginal effect of each factor on the respondents’ WTP and actual participation frequency in MRB-related events. Given that the null hypothesis was true (Prob > chi2) for both dependent variables, the probability of obtaining the Chi-squared statistic was found to be 0.0000 (

Table 5. Ordered probit regression model of marginal effects.

	Willingness to Participate		Actual Participation Frequency
Variable	dy/dx	Std.err.	dy/dx	Std.err.
Biophysical Conditions
Perceived environmental quality	−0.004	0.014	−0.058 ***	0.019
Future concerns	0.005	0.012	−0.029 *	0.017
Rules-in-use
Integrated watershed management	−0.037 *	0.02	0.046 *	0.026
Forest management	−0.002	0.002	0.002	0.003
Effectiveness of the government sector	0.004 *	0.002	0.003	0.003
Watershed management and operation	−0.002	0.003	0.004	0.004
Presence of poor management practices	0.003 *	0.002	0.006 ***	0.002
Socioeconomic Characteristics
Gender	−0.002	0.021	−0.073 **	0.029
Age group	−0.008	0.012	−0.035 **	0.016
Education level	0.002	0.013	−0.03 *	0.017
Annual income range	0.009	0.006	0.03 ***	0.009
Length of residence	−0.001	0.007	−0.001	0.01
Watershed importance cognition	−0.034 *	0.02	0.022	0.032
Forest value	−0.001	0.003	−0.009 ***	0.003
Interactional Capacity
Interpersonal communication	0.138 ***	0.02	−0.021	0.032
Cross-reach support	0.039 ***	0.013	−0.023	0.02
Information sharing	0.034 ***	0.011	0.31 ***	0.02
Number of observations	881		882
LR χ2	169.82		524.81
Prob > χ2	0.0000		0.0000
Log likelihood	−309.707		−706.525

Notes: *, **, and *** indicate statistical significance at 10%, 5%, and 1% levels, respectively.

). Thus, both models were statistically significant. Both rules-in-use and the interactional capacity significantly affected the respondents’ WTP, consistent with previous findings. The effectiveness of the government sector and watershed importance cognition were the only two factors that negatively influenced the respondents’ WTP. The marginal effects for government effectiveness and poor management (two rules-in-use factors), and all interactional capacity factors, were positive and significant. This indicates that a one-unit increase in interactional capacity boosts participants’ WTP in MRB-related events.

The information sharing factor was highly influential on actual participation. A one-unit increase in the spread of MRB-related news raised the probability of the respondent attending an MRB-related event by 31%. Biophysical conditions also significantly impacted actual participation rate. Respondents who observed less improvement in the MRB area and were more concerned about natural disasters attended watershed-related events less frequently. Socioeconomic factors similarly influenced participation negatively but were less impactful than biophysical conditions and information sharing. Regarding rules-in-use, participants were more likely to attend MRB-related events when they perceived high government management effectiveness or observed poor forest management practices.

In summary, interactional capacity had the greatest influence on the respondents’ WTP and actual participation rate. Rules-in-use was also somewhat connected to the two dependent variables but not as strongly as interactional capacity. Conversely, socioeconomic characteristics and biophysical conditions had little to no influence on WTP but were somewhat significant in actual participation frequency.

5. Discussion

Our findings empirically validate and expand the IAD framework by demonstrating how different variables influence stakeholders’ engagement, thereby deepening the application of this framework to water governance. The public’s environmental perceptions and awareness of management strategies and activities are closely related to governance effectiveness, institutional success, and conservation sustainability [56].

We found that increased awareness of the significance of watersheds would boost willingness in IWM. Similarly, in Japan’s Yodo River area, social capital and public awareness have been vital in boosting individual involvement [57]. Public participation in decision-making addresses collective action challenges and expands opportunities for further participation [58]. Increased participation also builds social capital in local communities, enabling them to become more self-driven and resilient when facing collective action problems over time [59].

A notable disparity exists between expressed WTP and self-reported actual participation in watershed conservation activities. The high level of expressed interest (85.9%) suggests a strong community inclination toward environmental stewardship, yet less than one-third regularly participate, suggesting barriers preventing locals from translating their willingness into action. According to the Theory of Planned Behavior, individuals’ actions are shaped by their perceived ability to engage in a behavior [60]. Participants may express interests, but actual involvement can be hindered by perceived barriers like lack of time, age, financial constraints, or inadequate tools (e.g., transportation to meetings) [61,62]. In this study, barriers identified for actual participation included high concern levels and socioeconomic traits (male, older, higher education, higher income). Incentives for participation included improved environmental conditions, better-integrated watershed management (IWM) performance, awareness of poor practices, and information sharing between the government and individuals.

A common motive is the desire for better conservation outcomes. A sense of community stewardship is reinforced when participants observe a direct benefit or improvements in their local environment, increasing their likelihood of participation [63]. Our identification of improved environmental conditions as an incentive aligns with this broader trend [64], where individuals are motivated by observable, tangible improvements. When management and conservation efforts visibly improve watershed biophysical conditions, citizens’ confidence and motivation to participate actively increased.

However, higher concern levels were associated with lower actual participation. When and Almomani [63] noted that high concern levels can serve as both a motivator and a deterrent interchangeably. Excessive concern may lead to feelings of being overwhelmed or pessimism about the effectiveness of participation, lowering active participation despite high willingness. Other researchers [65] have linked public participation with environmental risk perceptions and trust in governance. Our survey showed significant improvements in water pollution and flooding in the MRB, suggesting that, despite ongoing concerns, trust in the local government’s ability to manage these risks likely led to a nuanced decrease in participation as concern levels rose. Community concerns over flooding heightened awareness of the ecosystem services provided by riparian forests, including water purification and erosion control [5,34,66,67]. Similar trends have been observed in China’s Miyun Reservoir Basin, Heihe River Basin, and Baiyangdian watershed, where experiences with environmental issues have enhanced public environmental perceptions [68,69,70].

We found that a more comprehensive understanding of forest ecosystem services decreased the probability of actual participation (Table 5), which may initially seem counter-intuitive. A possible explanation is that awareness of ecosystem complexity might lead individuals to overestimate the effort required to make a difference, discouraging actual participation. Moreover, if individuals do not perceive an immediate threat or urgency, their motivation to participate may be reduced. Respondents valued regulating and supporting services more than provisioning and cultural services, indicating some understanding of ecological processes. This aligns with findings from Lima and Basto [71], who noted a predominant recognition of regulating services. The lesser emphasis on cultural services may indicate a weaker connection between community well-being and the watershed ecosystem, contrasting with locales where tourism and direct livelihood reliance on forests highly value cultural services [72]. Most respondents in the MRB recognized both the instrumental and intrinsic value of natural watershed resources. Environmental education might, therefore, be expanded to areas such as involvement in problem-solving, responsible behavior, youth engagement, and leadership to guide specific actions.

The universality of environmental degradation and social development has reduced the effect of demographic variables on environmental awareness [68], with a gap in actual environmental participation being more influenced by economic capacity and leisure time [73]. Our findings indicate that while willingness to participate is not closely tied to gender, age, education, or income, these factors do play a role in actual participation. Despite the lower tendency of the more educated to engage in environmental projects, individuals with higher incomes show more activity in IWM-related activities [74,75].

In regions dominated by agriculture and horticulture activities [76], higher education and income levels correlate with greater involvement. Individuals are more likely to contribute when they see clear financial advantages, such as jobs or revenue from non-timber forest products. However, this trend does not hold in the MRB, where higher education and income are associated with competing priorities, time constraints, and a preference for indirect forms of engagement (e.g., financial donations). The lack of immediate, tangible benefits from hands-on participation likely contributes to this contrasting trend. These participation trends, based on gender, age, and income, highlight the need for IWM plans to consider these social dynamics [77,78,79,80].

Regarding institutional capacity and governance, overall IWM performance and poor practices were the two most significant predictors of WTP and actual participation. Echoing Liu et al. [65], we found that effective public engagement boosts environmental regulation, with positive views on IWM enhancing WTP but potentially reducing direct participation due to enhanced trust in authorities’ capabilities [37]. Conversely, frequent observations of poor forest management practices led to a higher probability of involvement by motivating respondents to seek change. Naiga and Penker [81] found that effective law enforcement and both downward and upward accountability promoted WTP. Similarly, we found that a higher perceived effectiveness of government sectors increased respondents’ WTP. While public participation allows individuals to express expectations and demands, it has also positively correlated with higher satisfaction levels [65]. Furthermore, satisfaction with public services has been directly associated with the quality of those services [82]. Our study adds to this narrative by correlating satisfaction and participation with higher institutional trust and efficacy, thus contributing to the IAD framework’s applicability in analyzing water governance. The interaction element of the IAD framework could be expanded to incorporate the stages of perception, intention, and behavioral change. Our results show that while recognizing the importance of watershed value increases willingness, translating this into action requires strategic engagement beyond simply raising awareness.

This study highlights the impact of subjective cognition in shaping public engagement in centralized water governance. Awareness of current watershed conditions, past management experiences, and future expectations are critical in moving individuals from intention to action. Integrating these elements into the IAD framework allows a more comprehensive understanding of how subjective interpretations of information, experiences, and future visions influence willingness and participation, especially in top-down governance structures. Additionally, a dynamic, longitudinal analysis of endogenous variables is essential for better predicting participation behavior over time. A longitudinal approach can reveal how shifts in perceptions, experiences, and external conditions drive sustained participation and support the long-term viability of informal institutions.

Our study revealed that individuals’ willingness to encourage participation in IWM activities did not lead to a higher willingness to participate, contradicting the role of social information flow highlighted by Floress et al. [83]. This observation suggests that encouraging others does not directly reflect one’s trust levels, which Collard and Gambetta [84] identified as pivotal for community and policy engagement. Despite this, 79% of participants were willing to disseminate information within their networks, indicating a strong potential for community-wide information sharing. Additionally, we discovered that recognizing the value of cross-reach support significantly elevates WTP, validating the IAD framework’s emphasis on collaborative stakeholder engagement in resource management. Similarly, the longitudinal connectivity of the Min River, as noted by Wang et al. [51], necessitated the cooperation of stakeholders across spatial scales to achieve positive management outcomes. Thus, understanding the need to enhance cross-reach cooperation would encourage the public to fully understand IWM, thereby potentially increasing their WTP in IWM-related activities. A more integrated understanding of the interconnectedness across biophysical conditions, socioeconomic factors, and governance might encourage individuals to participate in decision-making processes [85].

Interactional capacity is vital in participatory processes and facilitates institution formation [38,86]. This interaction is critical in establishing institutions and associations, leading to collective action and participation in mitigating environmental risks [87]. The information sharing variable had the most significant impact on both WTP and actual participation. This observation is supported by the findings of Dolnicar et al. [88] regarding the capability of factual information to increase public support for water policies and projects. Information provision promotes participation in and adaptive responses to environmental conservation practices [38,89,90]. Our study affirms that interactional capacity is crucial for effective water governance, emphasizing that informed communication and collaboration facilitate stakeholder engagement. This insight complements the IAD framework by highlighting the need for strong interactional capacities to build resilient institutions. To enhance institutional credibility, governments should monitor the authenticity of information and prioritize factual information over persuasive arguments. For example, visualizations can readily convey water treatment information [88]. Large basins like the MRB, spanning multiple administrative regions and departments, often struggle with information disclosure that is fragmented and inconsistent. To address this, we recommend that policymakers appoint designated agencies responsible for integrating and releasing river basin policies, plans, and data uniformly across these regions. Standardizing release methods, timing, and standards will enhance communication accuracy and completeness, thereby improving public responses’ effectiveness.

6. Conclusions and Policy Implications

In this study, adopting the IAD framework as a structural lens, we analyzed the complex interdependencies that drive stakeholder engagement in environmental governance. Our findings, drawn from the marginal effects from the probit regression model, highlight the significant influence of interactional capacity on public willingness to participate (WTP) and actual participation in IWM, followed by the perceived effectiveness of rules-in-use in watershed management. Biophysical conditions and demographic factors in the socioeconomic context had limited impact on WTP but significantly affected actual participation. Environmental cognition strongly predicted WTP but did not necessarily lead to action.

Clear communication about environmental status, management strategies, public participation channels and opportunities, and monitoring and surveillance means is essential. Beyond providing information, all levels of government need to institutionalize decentralization to provide access for stakeholder involvement and to empower grassroots participation. Adopting a “river chief” system could bring decisions directly to applicable areas. However, a lack of institutional support could undermine the river chief’s actual power in the coordination and supervision of the river basin [91]. At its most fundamental level, watershed management is the key to increasing direct stakeholder participation, such as downward monitoring accountability (e.g., monitoring poor forest practices).

To ensure the sustainability and viability of IWM, institutional decentralization and informational accessibility need to be considered. From a unique institutional perspective extended from the interactional arena and enhanced focal action (public participation) of the IAD, several policy innovations and practical recommendations targeted at integrated watershed management could be made. First, cross-department and cross-reach water governance should be improved to reduce administrative costs and complexity. Second, a community-focused approach to IWM increases social cohesion. With high levels of willingness and commitment, there is substantial potential to form informal institutions within the watershed. Policies that facilitate community-building activities, stakeholder collaboration, and cooperative decision-making can strengthen residents’ sense of ownership and responsibility. Establishing community-based monitoring groups and involving residents in decision-making will increase public engagement and lead to better management outcomes by leveraging local knowledge and ensuring that policies address community needs. Lastly, improving transparency and accountability in water governance is necessary for ensuring policy effectiveness and credibility. In this case, resource coordination was optimized to maximize conservation efforts for the sustainable development of the basin.