Introduction
Wetlands are among the most ecologically significant yet increasingly threatened ecosystems worldwide (Hambäck et al. Reference Hambäck, Dawson, Geranmayeh, Jarsjö, Peacock and Collentine2023). They play a vital role in maintaining biodiversity, regulating hydrological and climatic cycles and supporting the livelihoods of millions of people. As transitional zones between terrestrial and aquatic environments, wetlands serve as critical habitats for numerous species, including rare and endangered migratory birds (e.g., Siberian crane (Leucogeranus leucogeranus); Parchizadeh & Williams Reference Parchizadeh and Williams2018). However, these ecosystems are under growing pressure from anthropogenic activities such as land conversion, pollution and illegal hunting, which undermine their ecological integrity (Brotherton et al. Reference Brotherton, Joyce and Scharlemann2020, Qiu et al. Reference Qiu, Zhang and Ma2024). Wetland conservation increasingly requires approaches that integrate ecological knowledge with social science perspectives, ensuring that ecological threats are addressed alongside governance and behavioural dimensions.
Environmental education fosters ecological awareness, promotes environmental sensitivity and encourages active community participation in conservation efforts (Monroe et al. Reference Monroe, Plate, Oxarart, Bowers and Chaves2019, UNESCO 2020). It draws upon experiential, transformative and participatory learning paradigms (Mezirow Reference Mezirow1997, Kolb Reference Kolb2014, Ardoin et al. Reference Ardoin, Bowers and Gaillard2020). It can produce measurable conservation benefits, such as reduced poaching and stronger community-based policing (Ardoin et al. Reference Ardoin, Bowers and Gaillard2020). It can thus create positive interactions between conservationists and poachers when confronted by law enforcement, strengthen intergroup relations and reduce the escalation of conflicts (Soofi et al. Reference Soofi, Ghasemi, Ahmadpour, Soufi, Islami and Eckert2024).
Freire’s (Reference Freire, Beck, Jenks, Keddie and Young2020) pedagogical principles emphasize the potential of environmental education to advance environmental justice and challenge systemic inequalities (Hart Reference Hart2024). In this study, Freire’s dialogical approach informed the design of participatory workshops, fostering critical reflection and shared decision-making among stakeholders. However, deep-rooted barriers to conservation – stemming from fragmented governance, limited institutional capacity and differing stakeholder perspectives – mean that environmental education alone is not enough and must be complemented by broader social and institutional mechanisms (Rodríguez-Izquierdo et al. Reference Rodriguez-Izquierdo, Gavin and Macedo-Bravo2010, Ardoin et al. Reference Ardoin, Bowers and Gaillard2020).
To address these complexities, scholars have increasingly turned to integrated frameworks that combine environmental education with conservation behaviour change theories tools such as social network analysis (SNA). SNA provides a structural lens for understanding stakeholder relationships and patterns of influence within conservation networks (Borgatti et al. Reference Borgatti, Agneessens, Johnson and Everett2024). It examines how social behaviour emerges from network configurations, using concepts such as centrality, density and clustering to identify key actors and pathways for collaboration (Wasserman & Faust Reference Wasserman and Faust1994, Kadushin Reference Kadushin2012). SNA also reveals how learning and behavioural norms diffuse through social ties and which actors catalyse or inhibit change. SNA has proven valuable in environmental governance for diagnosing power dynamics, facilitating resource mobilization and strengthening collective action (Prell et al. Reference Prell, Hubacek and Reed2009, Ingold & Leifeld Reference Ingold and Leifeld2016).
Complementing this, conservation behaviour change theories offer insights into the psychological and social drivers of pro-environmental action. Models such as the theory of planned behaviour (TPB), protection motivation theory (PMT) and the value–belief–norm (VBN) framework explain how threat perception, self-efficacy, moral norms and social support shape individual and collective responses to environmental challenges (Ajzen Reference Ajzen1991, Stern Reference Stern2000, Bubeck et al. Reference Bubeck, Botzen and Aerts2012, Yuriev et al. Reference Yuriev, Boiral, Francoeur and Paillé2018). In this study, the capability–opportunity–motivation–behaviour (COM-B) model, supported by key constructs from TPB and VBN (e.g., attitudes, perceived norms and moral obligation), serves as the primary behavioural framework for designing and interpreting the necessary intervention (Michie et al., Reference Michie, van Stralen and West2011). We integrate environmental education, SNA and behavioural change theory into a three-layered framework: education builds cognition and motivation; network analysis maps collaboration; and behavioural theory explains collective outcomes. These behavioural frameworks have been used to explain compliance with anti-poaching rules and to design interventions that strengthen community participation in wildlife protection (e.g., Jones et al. Reference Jones, Papworth, Keane, Vickery and St John2021, Sullivan-Wiley et al. Reference Sullivan-Wiley, Shyamsundar and Musengezi2023), yet few studies have empirically shown how education reshapes local governance networks in conservation.
We apply this model in the Sorkhrud Wetland of northern Iran’s Mazandaran Province because this seasonal wetland is one of the most important wintering habitats for migratory birds in the region (Ahmadpour et al. Reference Ahmadpour, Karimi, Ghasempouri, Ahmadpour and Yaghobzadeh2011, Ashjar et al. Reference Ashjar, Keshavarzi, Moore, Zarei, Busquets, Zebarjad and Mohammadi2023). It is also ecologically vulnerable: illegal hunting practices, particularly aerial trapping by local fishermen, driven mainly by market demand and long-standing damgah traditions (local traditional hunting structures or traps), have severely threatened bird populations and degraded surrounding habitats such as rice fields (Ahmadpour et al. Reference Ahmadpour, Ahmadpour, Hoseini, Ghasempouri, Jafari, SinkaKarimi and Amouie2012, Ashjar et al. Reference Ashjar, Keshavarzi, Moore, Zarei, Busquets, Zebarjad and Mohammadi2023). Conservation efforts are further complicated by contested land ownership and limited community involvement, with no prior structured education or participatory governance initiatives in the area, underscoring the need for inclusive and context-sensitive interventions.
The educational intervention was designed around five core components: (1) raising awareness; (2) deepening ecological knowledge; (3) reinforcing environmental values; (4) enhancing decision-making skills; and (5) fostering collective participation. The study examined changes in network structure and conservation behaviour before and after the intervention to offer actionable insights for designing effective policies to address critical threats such as illegal hunting.
Given that the Sorkhrud Wetland network is centralized and fragmented, with government actors being dominant and local communities being peripheral, we hypothesized that the pre-intervention network would have low density, high concentration, long geodesic distance and uneven centrality. Secondly, given that education was expected to enhance cohesion and redistribute influence, we hypothesized that with intervention density would rise, concentration would fall, geodesic distance would shorten, transitivity and mutual ties would increase and centrality would become more balanced across actors. Thirdly, because the three-layered model (education → network dynamics → behaviour) might explain how awareness and communication shifts foster collaborative conservation, we hypothesized that post-intervention patterns of awareness, ties and cohesion would align with the model, reflecting greater cognition and motivation in stronger networks and improved behaviour.
These three hypotheses thus tested how education-driven network changes might translate into measurable conservation behaviour in wetland contexts.
Methods
Study area
The Sorkhrud Wetland (Fig. 1), covering c. 1200 ha, forms part of the Fereydunkenar International Wetland (FIW) complex, located in Mazandaran Province (northern Iran) along the southern coast of the Caspian Sea (36°38′–36°39′N, 52°26′–52°36′E). Sorkhrud Wetland comprises three damgah units (traditional bird-trapping zones) embedded within low-slope rice paddies that serve as critical wintering habitats for migratory birds. These areas are often enclosed by straw barriers to minimize disturbance and facilitate bird settlement (Ahmadpour et al. Reference Ahmadpour, Ahmadpour, Hoseini, Ghasempouri, Jafari, SinkaKarimi and Amouie2012, Ashjar et al. Reference Ashjar, Keshavarzi, Moore, Zarei, Busquets, Zebarjad and Mohammadi2023). The Wetland’s semi-closed morphology and shallow freshwater pools, formed after rice harvest, create ideal feeding grounds for waterfowl and waders. Dominant species include common teal (Anas crecca), white-fronted goose (Anser anser) and mallard (Anas platyrhynchos), with average seasonal bird counts exceeding 130 000 individuals (Ahmadpour et al. Reference Ahmadpour, Karimi, Ghasempouri, Ahmadpour and Yaghobzadeh2011, Ashjar et al. Reference Ashjar, Keshavarzi, Moore, Zarei, Busquets, Zebarjad and Mohammadi2023). The Wetland’s seasonal hydrology, rice field mosaics and high bird diversity make it a critical ecological system in northern Iran.
Geographical location of Sorkhrud Wetland, along with its three damgah units, within the Fereydunkenar International Wetland in northern Iran.
Figure 1. Long description
The map illustrates the geographical location of Sorkhrud Wetland and its three damgah units within the Fereydunkenar International Wetland in northern Iran. The map includes labels for Fereydunkenar Damgah, Azbaran Damgah, and Sorkhrud Damgah, each represented by different colors and patterns. The Fereydunkenar International Wetland is outlined in black, and the map provides a detailed view of the wetland’s boundaries and the surrounding regions, including parts of Iran, Turkmenistan, and Afghanistan. The Caspian Sea is also visible to the north. The map uses a scale to indicate distances in kilometers and miles, and a compass rose shows the north direction.
Threats to the Wetland include illegal hunting, with up to 3000 birds being killed daily during peak migration seasons (Ashjar et al. Reference Ashjar, Keshavarzi, Moore, Zarei, Busquets, Zebarjad and Mohammadi2023). Illegal hunting in Sorkhrud Wetland is driven by market demand and supported by traditional and modern trapping methods. The Wetland also faces pollution from agricultural runoff, wastewater and plastic debris, including microplastics from nets and bullet casings, linked to poor waste management (Mirzaei et al. Reference Mirzaei, Knierim, Nahavand, Shokri and Mahmoudi2019, Ashjar et al. Reference Ashjar, Keshavarzi, Moore, Zarei, Busquets, Zebarjad and Mohammadi2023). Land use is dominated by rice cultivation under farmer cooperatives, along with surrounding aquaculture, settlements and roads, posing major challenges for conservation governance and community engagement (Ashjar et al. Reference Ashjar, Keshavarzi, Moore, Zarei, Busquets, Zebarjad and Mohammadi2023).
Research methodology
To test the applicability of the proposed three-layered socio-educational framework in a real-world wetland context, the study employed a single-group pre–post mixed-methods design in Sorkhrud Wetland, with baseline data collected in early 2022, the educational intervention implemented during 2022–2023 and the follow-up assessment conducted immediately after the intervention.
A mixed-methods approach was employed, emphasizing SNA, educational intervention and principal component analysis (PCA). There were three key stages: (1) identifying actors and their interactions within the conservation network; (2) analysing the network’s structure and dynamics prior to the intervention; and (3) designing and implementing a targeted educational intervention, followed by an assessment of behavioural and structural changes post-intervention.
A qualitative and participatory approach was first employed to accurately map the network of partnerships involved in the conservation of Sorkhrud Wetland. Data collection involved semi-structured exploratory interviews with individuals from a diverse group of stakeholder organisations (n = 44), including government officials, non-governmental organizations (NGOs), environmental activists, local hunters and academics. A snowball sampling technique was then applied, beginning with key institutional actors such as municipal officials and local cooperatives. Referrals continued until saturation was reached, defined as no new actors or ties emerging after three consecutive referrals. In total, 370 questionnaires (Appendix S1) were distributed, and 350 were completed (c. 95% response rate). This process ensured broad coverage and produced a final set of 44 unique network nodes representing distinct organizations or individuals. The network elicitation question was: ‘Please list the organizations, groups, or individuals with whom you have had regular communication or collaboration regarding wetland conservation activities during the past year.’
Before intervention, interaction data were systematically coded using content analysis and then processed with UCINET v6.610 software (Analytic Technologies, Harvard, MA, USA) to quantify relational ties and calculate network metrics, which enabled the construction of a detailed map of the conservation and exploitation network. Ties were analysed as undirected and binary, defined by at least one reported interaction in the past year. Discordant reports were symmetrized, considering a tie as being present if either actor reported it.
Structural analysis was conducted at micro, meso and macro levels. At the micro level, indices such as input, output and betweenness centrality were used to assess each actor’s relative position and influence, their intermediary roles and their ability to access information. The meso level was the level at which structural patterns such as core–periphery were assessed. At the macro level, metrics including network density, centralization, reciprocity, transitivity and average geodesic distance were examined to evaluate the overall structural cohesion of the network.
To test the statistical significance of these changes before and after the intervention, permutation (reshuffling) tests with 1000 iterations were conducted in UCINET. The core–periphery index was analysed at the meso level using the Borgatti and Everett (Reference Borgatti and Everett2000) algorithm, with threshold criteria based on the correlation fit between the ideal and observed structures. The core–periphery structure was analysed to identify key clusters of activity and to distinguish central actors from peripheral ones (Islami et al. Reference Islami, Azadi, Flores Díaz and Sarvi Sadrabad2024a, Reference Islami, Ghanbari and Azadi2024b).
Based on the findings, a structured, problem-based training programme was designed and implemented for key actors. The programme consisted of four modules, each lasting c. 2 hours: (1) ecological functions of wetlands and the role of migratory birds; (2) threats posed by poaching and legal/ethical implications of habitat destruction; (3) participatory conservation principles and connections to local livelihoods; and (4) ecosystem services including carbon storage. Each module had specific learning objectives, including raising ecological awareness, strengthening conservation values, enhancing collaborative decision-making skills and fostering accountability among stakeholders. The programme was facilitated by university faculty specializing in environmental governance and biodiversity, together with local NGO practitioners experienced in participatory training, ensuring both scientific rigour and contextual relevance. In line with Freirean principles, curriculum topics were derived from stakeholder interviews and refined through feedback. Workshops created dialogical spaces for participants to reflect on governance power relations and to co-produce knowledge. The training was delivered through three face-to-face workshops to 68 attendees, locally produced documentaries, group discussions and educational brochures. Following the intervention, two evaluations were conducted in Sorkhrud Wetland during late 2023, immediately after completion of the training workshops, using the same instruments as in the baseline assessment to ensure comparability.
A network reanalysis, focusing on key actors, first utilized the previous indicators across micro, meso and macro levels to examine changes in centrality, cohesion and core–periphery structure, as well as shifts in behaviours and intersectoral partnerships. The knowledge and attitudes of local stakeholders were then assessed both before and after the intervention using the same seven-item questionnaire, which addressed perceptions of environmental services, hunting threats, conservation importance and the economic benefits of wetlands. PCA was employed to construct a composite index, which was evaluated using criteria such as eigenvalues, explained variance and factor loadings. Responses were measured on a four-level Likert scale (none, little, somewhat, a lot) and compared with network and institutional variables. The seven questionnaire items (Appendix S1) were pre-tested with 20 stakeholders. Sampling adequacy was confirmed (Kaiser–Meyer–Olkin (KMO) test = 0.87; Bartlett’s χ2 = 512.43, df = 21, p < 0.001). The PCA was conducted with Varimax rotation, retaining components with eigenvalues > 1 and scree plot support. Reliability was strong (Cronbach’s α = 0.91). Qualitative insights from interviews and document reviews were used to contextualize and interpret quantitative outcomes, ensuring that network metrics and PCA results were explained in relation to stakeholder narratives and observed governance dynamics. This integration of qualitative and quantitative strands ensured that the mixed-methods design was fully operationalized and strengthened the internal validity of the study.
Results
Identifying actors in the network
Forty-four actors related to the conservation of Sorkhrud Wetland were identified, including government entities at national, provincial, county and district levels; NGOs; cooperatives; the private sector; the judiciary; the military; and local communities (Table 1). Specifically, two natural resources offices were present: Amol Natural Resources Office (county level, Actor No. 5) and Sorkhrud Natural Resources Office (district level, Actor No. 21). Both were separate from the Department of Environment (DoE) and were operating under the Ministry of Agriculture Jihad, whereas the DoE (e.g., Actor Nos. 18, 22, 29, 32) is responsible for wetland conservation. All information regarding the activity level and actor type for each of the 44 identified actors is presented in Table 1. Representatives of the 44 actors participated in the intervention, and their roles were triangulated through interviews and workshop observations. Using the Core–Periphery Index, 11 central actors were identified; others occupied more marginal positions within the network (Fig. 2). Government institutions were dominant in the governance network, as reflected in their core positions and high connectivity based on the Core–Periphery Index, consistent with a centralized structure. The weak role of local communities, especially pastoralists, was evident through their peripheral positions and low degree of centrality, despite their geographical proximity to the Wetland (Fig. 2).
Key actors involved in the conservation of Sorkhrud Wetland: classification by type, centrality and activity level.
Table 1. Long description
The table presents a detailed list of 44 actors involved in the conservation of Sorkhrud Wetland, categorized by their type, centrality, and activity level. It includes government organizations, NGOs, cooperatives, private sector entities, the judiciary, the military, and local communities. The table has 44 rows and 4 columns, with headers labeled ‘No.’, ‘Actor code/name’, ‘Actor type’, ‘Core-Periphery Index’, and ‘Activity level’. Each row provides specific information about the actors, such as their names, types, centrality indices, and activity levels. Notable trends include the dominance of government institutions in core positions and the peripheral positions of local communities, reflecting a centralized governance structure.
NGO = non-governmental organization.
Participation network map of actors in Sorkhrud Wetland conservation. Relationships among actors are shown with three main features: community structure, where blue lines show one-way ties and red lines show two-way ties, reflecting intra-group cohesion; central actors, where node size indicates degree centrality; and intermediary roles, where actors act as intermediaries, bridging otherwise-disconnected groups. Nodes are shown in blue and red; same-coloured nodes tend to form a cohesive community, with most of their connections within the group.
Figure 2. Long description
A scatter plot visualizes the participation network map of actors involved in Sorkhrud Wetland conservation. The plot features dozens of data points, each representing an actor, with node size indicating degree centrality. Blue lines represent one-way ties, while red lines indicate two-way ties, reflecting intra-group cohesion. The actors form cohesive communities, with same-colored nodes tending to connect within their groups. Some actors act as intermediaries, bridging otherwise-disconnected groups. The x-axis and y-axis represent the network positions of the actors, with values being actual. The plot highlights clusters, patterns, and the overall structure of the network, showing how different actors interact and collaborate in the conservation effort.
Respondents’ awareness of public participation in wetland conservation
Respondents showed strong awareness of public participation in conserving Sorkhrud Wetland (Table 2). Because no item was rated as ‘low’ (score 1) or ‘none’ (0), Table 2 presents only the ‘to some extent’ (2) and ‘high’ (3) response categories. The scale was reliable (Cronbach’s α = 0.91), with the principal component explaining 73.9% of variance (eigenvalue = 5.511). Awareness was highest for household livelihoods (mean = 3.0, rank 1), followed by degradation risks/food security (mean = 2.9) and cultural/recreational value (rank 3). Lowest was carbon sequestration (mean = 2.0, rank 7), showing limited climate awareness. Overall, a median of 3 and low dispersion reflect shared attitudes and the potential for participatory governance. It should be noted that attitudinal data (Likert responses from 350 stakeholders) and network analysis results (44 actors) were derived from different but complementary samples, ensuring that quantitative awareness measures and structural network indicators were integrated without overlap of respondents.
Respondents’ self-reported awareness of public participation in wetland conservation (scoring scale: none = 0, low = 1, to some extent = 2, high = 3; no responses were recorded for ‘none’ or ‘low’, so only ‘to some extent’ and ‘high’ data are included).
Table 2. Long description
The table presents data on respondents’ awareness of public participation in conserving Sorkhrud Wetland. It includes eight items rated on a scale from to some extent (2) to high (3). The items cover various aspects such as protecting migratory birds, combating illegal hunting, preventing soil and water degradation, protecting wildlife for future generations, and more. The table shows the percentage of respondents who rated each item as to some extent or high, along with the mean, median, standard deviation, and priority ranking for each item. Notable trends include high awareness for household livelihoods (mean 3.0, rank 1) and degradation risks/food security (mean 2.9), while carbon sequestration has the lowest awareness (mean 2.0, rank 7). The median response is 3, indicating a general consensus among respondents.
Collaboration cohesion and network structural dynamics before and after the educational intervention
Before training, the low network density (0.22) showed weak cohesion and connectivity, while its high concentration (0.51) revealed power as being centralized among few participants (Table 3). After training, network density rose to 0.55 and concentration dropped to 0.34, reflecting stronger ties, new actor involvement and a more balanced role distribution. The observed increase in network density and decrease in concentration after the intervention were statistically significant (permutation tests, p < 0.05). These changes were significant (p < 0.05), but their interpretation should be viewed in light of the network’s size (44 actors), and indicators such as density, transitivity and mutual ties may be sensitive to sample size and reporting bias.
Structural network metrics and centrality indicators of key actors before and after the educational intervention.
Table 3. Long description
The table presents structural network metrics and centrality indicators for key actors before and after an educational intervention. It includes data on network density, centralization, average geodesic distance, transitivity, and reciprocity. The table has 15 rows and 5 columns, with columns labeled Category, Indicator/actor, Metric, Before intervention, and After intervention. Key actors include Sorkhrud Municipality, Sorkhrud District Office, Sorkhrud District Court, Amol Department of Natural Resources, Sorkhrud Migratory Swans Support Society, Kalleh Sabz Fishing Cooperative, Beautiful Birds Fishing Cooperative, Sorkhrud Police Station, Provincial Department of Environmental Protection, Sari Forest and Rangeland Administration, and Mahmoudabad Department of Environment. Notable trends include increased network density from 0.22 to 0.55 and decreased centralization from 0.51 to 0.34 after the intervention, indicating stronger ties and more balanced role distribution.
Pre-intervention, the network showed weak connectivity: geodesic distance was 3.45, transitivity was 0.08 and mutual ties was 0.33, indicating slow information flow, poor responsiveness and limited two-way interaction (Table 3). Post-intervention, geodesic distance improved to 3.11, transitivity rose to 0.25 and mutual ties increased to 0.48, reflecting faster communication, stronger structural resilience and greater stakeholder solidarity, although crisis responsiveness remained limited.
Centrality analysis before and after the educational intervention
The municipality, district administration and environmental protection organization had the highest centrality, underscoring their key roles (Table 3). Before training, the Sorkhrud Municipality, Kalhesabz Fishing Cooperative, District Court and Provincial Environmental were most connected. After training, centrality shifted: the Municipality strengthened its leadership with higher input (from 23 to 34) and betweenness (from 887.36 to 1124.77), becoming the main intermediary. The Mahmudabad Environmental Unit and Department gained importance in resource mobilization, while the Migratory Swan Conservation Association grew markedly (input from 4 to 13, output from 4 to 8, betweenness from 124.97 to 631.52), reflecting greater cooperation and acceptance. Enhanced intermediary roles for the District Court and fishing cooperatives further improved information flow and responsiveness.
Conceptual model of the impact pathways of the educational intervention
The proposed conceptual model (Fig. 3) integrates environmental education, SNA and conservation behaviour change theory to depict the observed pathways through which educational interventions and social networks influenced conservation behaviour and contributed to the conservation of Sorkhrud Wetland. The model comprises three interconnected layers: (1) cognitive and motivational empowerment through education; (2) strengthening inter-actor linkages via network analysis; and (3) behavioural change and improved ecological conditions as reported outcomes.
Three-layered conceptual model of conservation behaviour change. This model integrates model structure (education (cognitive/motivational empowerment), social network analysis (linkages) and behaviour change (ecological sustainability)), empirical basis (pre- and post-intervention data showing education shifts in attitudes/skills, while network analysis tracks relational changes) and practical use (a tool for designing education programmes, monitoring networks and evaluating outcomes in participatory conservation).
Figure 3. Long description
The flowchart illustrates an integrated theoretical framework for conservation behavior change. It is divided into three main components: Environmental Education, Social Network Analysis, and Behavior Change. Environmental Education fosters analytical, problem-solving, and decision-making skills, as well as scientific knowledge and conceptual understanding. Social Network Analysis includes a diagnostic role focusing on network indicators such as density and centrality, and a strategic role aimed at strengthening links. This analysis leads to the development of trust, social identity, and self-efficacy, which in turn promotes active and collective participation and trust. Behavior Change is driven by value-based attitudes and motivation to act, leading to the development of social capital through bonding and bridging. The framework ultimately aims to promote pro-environmental behavior, responsibility, cooperation, participatory governance, and socio-ecological resilience.
Discussion
All three hypotheses were supported: pre-intervention, post-intervention changes and the three-layered model were confirmed. To clarify, the three-layered model was not derived from the findings. Instead, it was developed a priori in this study, based on existing theory (Pahl-Wostl et al. Reference Pahl-Wostl, Craps, Dewulf, Mostert, Tabara and Taillieu2007, Reed et al. Reference Reed, Evely, Cundill, Fazey, Glass and Laing2010, Valente Reference Valente2010), and it was used to guide the intervention in Sorkhrud Wetland. The study results subsequently confirmed the predictive validity of the model and showed that it is both testable and effective. This study makes three key contributions. First, it demonstrates how environmental education can function as a mechanism for network building among diverse conservation actors. Second, it highlights the integration of individual cognitive–motivational empowerment with institutional and structural levels of governance. Third, it introduces a three-layered intervention model that links education, SNA and behaviour change theory, offering a transferable framework for participatory conservation. Our findings indicate that environmental education interventions, when implemented within a socially networked context, may foster conservation-orientated practices. Alongside quantitative indicators, the intervention embodied participatory ideals: reflexivity was encouraged, and community agency shaped both content and outcomes, ensuring that the design was not purely top‑down. This dual framing, combining participatory ideals with structured behaviour change elements, acknowledges the tension between dialogical empowerment and managerial approaches, and it resolves this tension by positioning education as both a participatory process and a practical tool for measurable conservation outcomes. The observed behavioural changes demonstrate that education can enhance network cohesion, build trust among stakeholders and foster shared environmental values. These findings ground claims of empowerment, trust and cohesion in observed quantitative changes rather than speculation. While previous studies have explored the role of education in enhancing environmental social capital (e.g., Suškevičs et al. Reference Suškevičs, Hahn, Rodela, Macura and Pahl-Wostl2018, Arodin et al. Reference Ardoin, Bowers and Gaillard2020, Meier et al. Reference Meier, Powell, Stern, Frensley and Sène-Harper2024), our study contributes novel insights by combining longitudinal SNA with behavioural outcomes in a wetland conservation setting.
Beyond the local context, these findings have wider importance for global wetland conservation. Global Wetland Outlook (Dudley Reference Dudley2025) emphasizes that wetlands are among the most threatened ecosystems worldwide, requiring innovative governance approaches. Our three-layered model offers a transferable framework for other Ramsar sites facing similar challenges of centralized governance and marginalized local communities (Davidson & Finlayson Reference Davidson and Finlayson2018). The observed shift towards hybrid governance arrangements responds to calls for decentralized wetland management globally (Folke et al. Reference Folke, Hahn, Olsson and Norberg2005, Bodin Reference Bodin2017). The comparison of network structures before and after training revealed significant increases in participation, trust and actor cohesion, reflected in greater stakeholder involvement in decision-making, improved compliance with conservation rules and stronger collective monitoring of wetland activities.
These findings informed the development of a conceptual model that links individual-level empowerment with systemic mechanisms, offering a practical framework for guiding education-based conservation initiatives in similar local and regional contexts.
The network after the intervention acquired a more coherent and participatory structure. This shift not only reflects improved communication and stakeholder cohesion but also indicates enhanced empowerment and perceived agency among local actors, as they became more actively engaged in decision-making and collective monitoring. This restructuring represents a redistribution of influence from exclusively state-centric actors towards more hybrid governance arrangements, an outcome that is widely recognized as critical for adaptive wetland management globally.
Interactive education enhanced trust, shared role understanding and recognition of environmental values, reflected in expanded ties and clusters (Wals et al. Reference Wals, Brody, Dillon and Stevenson2014, Bodin Reference Bodin2017). Facilitators and local representatives gained mediation roles, improving information flow and trust. Post-intervention, interactions shifted from state to local actors, strengthening horizontal social capital and community communication power, consistent with Friedman et al. (Reference Friedman, Guerrero, McAllister, Rhodes, Santika, Budiharta and Wilson2020). Environmental education proved most effective as a network-building intervention engaging marginalized actors, aligning with Valente’s (Reference Valente2010) network-based interventions. Network analysis revealed both increased interactions and qualitative shifts in influence, exchange and institutional engagement, consistent with Newig et al. (Reference Newig, Jager, Kochskämper and Challies2019). Yet, broader contextual factors such as policy shifts, institutional changes and economic instability were not fully separated from educational impacts. Higher density and centrality may partly reflect short-term training effects or reporting bias, so these results should be interpreted cautiously as indicative rather than definitive evidence of empowerment and sustainability. Moreover, the study did not incorporate modern technologies (e.g., digital platforms, interactive media) in its educational approach, which suggests a direction for future research. Future studies could adopt longitudinal designs to better capture the persistence and nature of behavioural transformation, apply the model in diverse settings, isolate contextual variables and integrate digital tools to enhance the effectiveness and scalability of environmental education.
Targeted, place-based education within social networks reinforces norms, trust and legitimacy in conservation (Manfredo et al. Reference Manfredo, Berl, Teel and Bruskotter2021, Zhao et al. Reference Zhao, Liu and Han2024). Choe and Sheffield (Reference Choe and Sheffield2025) highlight the need to address individual, interpersonal and institutional dimensions, which were core to our model. We propose a three-layered framework integrating environmental education, SNA and behaviour change theory, extending beyond TPB (Ajzen Reference Ajzen1991) and VBN (Stern Reference Stern2000). Unlike classic approaches (Hungerford & Volk Reference Hungerford and Volk1990), our three-layered framework emphasizes the interplay of education, networks and institutions, filling gaps in frameworks such as those of Olsson et al. (Reference Olsson, Folke and Berkes2004) and Ansell and Gash (Reference Ansell and Gash2008). Validated by pre- and post-intervention analyses (Bodin Reference Bodin2017), our three-layered framework shows synergy between individual factors (e.g., knowledge, self-efficacy) and systemic components (e.g., connectivity, institutional alignment). Previous studies have confirmed that wetland education enhances awareness (Otto & Pensini Reference Otto and Pensini2017, Rosa & Collado Reference Rosa and Collado2019), while social capital builds trust (Wan & Du Reference Wan and Du2022) and networks shape norms.
Conclusion
This study addressed a critical governance challenge in wetland conservation: how to transform fragmented, state-centric decision-making into inclusive, participatory action. By integrating environmental education, SNA and behaviour change theory into a three-layered intervention model, we demonstrated a replicable pathway to reduce illegal hunting and enhance collective stewardship in Sorkhrud Wetland, Iran. Applying the three-layered model in this intervention led to three breakthroughs. First, the three-layered model redistributed influence from state-centric actors to local communities and NGOs, cutting network centralization by a third. Second, the three-layered model converted weak, unilateral communication into dense, reciprocal ties, doubling network density and nearly doubling mutual collaboration. Third, the three-layered model translated individual awareness gains into collective behavioural outcomes, documented through increased compliance with anti-poaching rules, participatory monitoring and shared stewardship. Key innovations include the use of pre- and post-intervention longitudinal network data to trace causal pathways and the explicit integration of Freirean dialogical principles with COM-B behavioural constructs, resolving the tension between empowerment and managerial control. The framework is not site-specific: it offers a transferable blueprint for any wetland or natural resource system where illegal use persists due to governance fragmentation and marginalized local actors.
Globally, this research responds to the Ramsar Convention’s call for innovative governance approaches. With over 35% of the world’s wetlands having been lost since 1970 and 22% more having been degraded, scaling such education network interventions is urgent. Future work should apply the model across diverse socio-political contexts, integrate digital tools and adopt longer-term follow-ups to validate sustainability. This study provides rare empirical evidence that strategic educational interventions, when embedded within SNA, can restructure conservation governance and deliver lasting behavioural change.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S0376892926100472.
Acknowledgements
We acknowledge the support of Tarbiat Modares University and Iran National Science Foundation. We thank all participants who responded to the questionnaire and thus contributed to the study by sharing their knowledge, experience and perceptions.
Author contributions
I Islami: conceptualization, funding acquisition, project administration, sampling, methodology, drafting, analysing the data, writing the manuscript. M Ahmadpour, MH Gorjian Arabi and SM Ghasempouri: sampling, resources, methodology, review and editing, contributing critically to the drafts and giving final approval for publication.
Financial support
This research was funded by the Iran National Science Foundation (project No.99012971).
Competing interests
The authors declare none.
Ethical standards
The study was conducted in accordance with institutional ethical guidelines. Informed consent was obtained from all participants prior to data collection. According to Tarbiat Modares University, formal ethical approval was not required for this type of non-invasive social research.
