The concept and theory of social networks was first introduced by Alfred Radcliffe-Brown in 1940 and was used by Barnes in the 1950s (Korom, 2015). The main approaches of network analysis are divided into three main groups: the metaphorical approach, descriptive approach, and structural explicit approach (Folke et al., 2005; Hahn et al., 2006). In the metaphorical approach, social networks are often considered as a metaphor. The general idea of this approach is that each of the actors in the social area that form the network are either related to each other and benefit from this relationship to achieve different goals, or they are not related to some people and are more self-reliant. Overall, this approach includes a series of applied studies in which social networks have been considered as a research tool in the governance of natural resources. But it provides little information about the real structure and patterns of social networks. Thus, this approach lacks the analytical and explanatory capabilities of the social network, and it is difficult for the researcher to investigate the effect of a specific factor in social networks through analytical processes such as participation and coordination (Fabricius et al., 2007).

The descriptive approach goes beyond the metaphorical approach; specific features of social networks such as horizontal and vertical networks, links, communication bridges, and network density are studied in this approach (Pretty and Ward, 2001). But this approach lacks a methodological strategy for identifying and differentiating aspects of social network structures and provides little ability to increase our understanding of various issues of social networks in relation to natural resources. The structural approach has transformed the researcher’s perception towards the concept of network structures and their function by introducing computational and quantitative aspects in the process of network analysis (Hahn et al., 2006). This approach includes a set of studies that examines social networks through applying systematic methods of data collection related to network relationships, methodology, and calculation-based modeling. The final goal of this approach is to explain the relationship between official definitions through quantitative and calculable features of the structure of social networks, and to present the results obtained from the analysis of this relationship in the governance of natural resources (Folke et al., 2005). The structural approach was used in this current study.

Social networks consist of a set of relationships and nodes, that is, the various types of stakeholders and the relationships between these stakeholders (nodes). In general, a social network can be considered as a graph of nodes and lines. Nodes in the network represent the actors and lines between nodes indicate the relationships between actors. Network analysis provides an empirical basis or an understanding of the flow of information and influence in governance networks to identify key organizations (Vignola et al., 2013). The main research on social networks seeks to understand how individual patterns of social relationships affect behavior (Ford et al., 2011). Social network analysis allows research into relationships beyond purely physical relationships and very significant relationships, and can provide very useful results which help improve social processes (Molano and Polo, 2015). Network analysis focuses on the relationship between actors, not on the actors themselves and their characteristics. Network analysis criteria can indicate the relationships between organizations, including which ones have the most active members of a network or those whose connections are interrupted by network members (Pedroza et al., 2016).

From a theoretical point of view, social network analysis is based on network theory and graph theory. Network theory is a very general term for evaluating relationships (Celik and Corbacioglu, 2016). Using social network analysis techniques, researchers evaluate social networks with a focus on relationships between individuals and organizations, and social network analysis has been widely used to help improve the efficiency and effectiveness of decision-making processes. In addition, it may identify the groups of people that have a central role, or separate groups or individuals. Further, it can identify opportunities to improve the flow of information and knowledge, improve the efficiency of formal communication channels, and increase the importance of informal networks (Bae et al., 2015). In social network analysis, more emphasis is placed on the type, quantity, and quality of communications and interactions instead of on the underlying variables. Indeed, matrices include the actors involved in interaction as well as the direction and frequency of interaction between them (Celik and Corbacioglu, 2016). An important point about the analysis of organizational networks is that the components of different organizations are not in direct contact with each other and their connections are via complex, extensive, and multiple relationships. Here, the concept of a network becomes necessary for the analysis of organizational relationships.

In many cases, Social Network Analysis (SNA) is used as a tool to determine influential actors in a collaborative process (Narayan et al., 2020). SNA makes it possible to study the interactions between activists, organizations, and other individuals who are involved in the process of natural resource management and governance. It is seen as a valuable method in management and governance studies. Focusing on the system’s relationships and not its elements is the main emphasis of this method. In other words, it examines all the system’s connections and provides a comprehensive understanding of its social structures and actions through discovering the relationships. This strategy is the main basis of the social communication hypotheses (Emirbayer and Goodwin, 1994; Freeman, 2004). This method can identify different social factors in the process of common-pool resource governance and can be used to determine these factors in the success or failure of resource management as well (Bodin et al., 2006; Bodin and Crona, 2009). For instance, SNA will determine what is the pattern of relations between activists and organizations in order to avoid the tragedy of common resources and to establish their self-imposed systems to use these resources in a regulated and sustainable manner (Prell et al., 2009). Obtaining such information is very valuable for external institutions that have the mission of facilitating the process of developing communication patterns between actors (Schneider et al., 2003; Ernstson et al., 2010). Therefore, SNA can be used as an efficient tool to analyze the process of information flow and knowledge exchange in the network of stakeholders and organizations; it can be used to create a successful system for the development and distribution of natural resources knowledge (Dempwolf and Lyles, 2012). In fact, mutual relationships between organizations in social networks play a much more important role than the development of official organizations to define and observe the rules of natural resource exploitation (Scholz and Wang, 2006).

There are several publications that have used SNA to study different aspects of environmental, natural resources, and watershed management and governance. For instance, Raum (2018) argued that appropriate models for ecosystem services need to identify different actors. The study showed that ecosystem governance and management would be effective and sustainable if various stakeholders and ecosystem services and their abilities are analyzed using the SNA method. In a similar study by Giurca and Metz (2018), SNA was used to identify some of the main organizations involved in Germany’s wood-based bio-economy innovation system and their relationship to each other. The figures indicated high density but this was only due to the information exchange of the actors. Based on the results, although the connection between organizations was weak, the level of trust was high due to their cooperation. Application of SNA was also used in the study by Ghorbani and Azadi (2021) to analyze trust and collaboration networks in rangeland co-management. It found low levels of trust and collaboration between different stakeholders including rangeland users and the experts and managers of governmental and non-governmental institutions. This caused severe challenges in the governance of natural resources.

Narayan et al. (2020) studied the governance of decentralized watershed treatment in four cities of India. According to the complexity of the different groups of stakeholders involved in the governance of Water, Sanitation and Hygiene (WASH) and the diversity of their interests, SNA method was used. The results revealed the key differences between mega- and secondary cities in terms of institutions, community engagement, and overall sanitation, including aspects of decentralized wastewater treatment plants, based on the city types. Another study by Nabiafjadi et al. (2021) analyzed the knowledge and information exchange networks among governmental organizations and NGOs that contributed to water governance in the Zayandeh-rud basin of Iran. The study showed that the collaboration of the institutions was necessary to implement better strategies in water management. It also indicated that high participation of the mentioned organizations as well as identifying power relations and densities in different processes of water governance could be used in other countries of the Middle East. The role of various actors and their social capital in groundwater management was analyzed using SNA in the study by Rahimi-Feyzabad et al. (2022). It revealed that the social capital of these organizations was low. In fact, the heterogeneous nature and contrasting strategies of these organizations led to the poor collaboration.

Reviewing the most recent year of SNA study publications, this method was used in different research areas as well: Joyez and Laffineur (2022) studied the occupation space, network structure, centrality, and the potential of labor mobility in the French labor market; Flemming et al. (2022) used the SNA method to identify networks of physicians responsible for the care of specific patient populations; Blanken et al. (2023) studied intersectoral collaboration at a decentralized level, and information flows in child welfare and healthcare networks; and Arnold et al. (2022) studied information exchange networks regarding chronic diseases in primary care practices in Germany.

The network analysis method is used for identifying the main stakeholders and actors in the community (Bodin and Prell, 2012). The appropriate understanding of the position and capacity of the main actors and stakeholders can lead to their appropriate organization in collaborative management programs (Ghorbani and Azadi, 2021). The current study aims to identify the organizations involved in the watershed management of Chehelchay in Golestan province, Iran, and analyze the communications and interactions between the organizations in terms of intensity and type. The research objectives are as follows:

• Identifying the various organizations active in Gorganrood watershed management

Golestan province (Figure 1A) has six sub-watersheds including the Voshmgir watershed dam with 8,100,000,000 square meters (m²), Western Gorganrood watershed with 3,500,000,000 m², Atrak watershed with 8,000,000,000 m², Qareh Sou watershed with 1,650,000,000 m², Gorgan bay watershed with 400,000,000 m², and Nekarood watershed with 1,000,000,000 m². The highest level of altitude of the province and consequently the maximum water available in the province comes from the Gorganrood watershed which is considered one of the most important rivers in Golestan province, flowing through all plains of Gorgan and dividing it into two parts. The main bed direction is from northeast to southwest, is 350,000 m long, and originates from the heights of Golestan forest. Chehelchay watershed is one of the mountainous regions of Iran with an area of about 61,777 acres located between 55° 23 and 55° 38 east longitudes and 36° 59 and 37° 13 north latitudes (Figure 1B). This watershed is located in Minoudasht in terms of political divisions and is one of the largest sub-watersheds of Gorganrood.

Network analysis is based on the common social science method of using different techniques, such as interviews and questionnaires, to collect the required data (Butts, 2008; Bodin and Prell, 2012). In this study, the social network of related organizations was analyzed. Thus, the related organizations involved in the participatory and comprehensive management of the Chehelchay watershed (30 organizations) were introduced during the trips by the research team to the site and using specialized questionnaires. Based on purposeful sampling, key informants were selected among the experts, managers, and authorities of these organizations in order to collect the required data in the network analysis. In these panels, a total of 84 specialists (three individuals from each organization) were interviewed. Then, the data were entered into Excel and UCINET6.631 software in the form of a matrix, and data processing and analysis were performed. This data processing was conducted using the concepts which are the basis of all of the mathematical algorithms (Klenk et al., 2009; Herzog et al., 2014). The most important of these concepts is centrality, power, and grouping (Narayan et al., 2020; Fatemi et al., 2021). Graph theory and NetDraw software were used to plot the data (Scott, 2012). In this research, a number of social network analysis measures were analyzed to assess the different network criteria between individuals, between organizations, and between individuals and organizations (Table 1).

In order to conduct the present study, eight steps were performed to examine the existing interaction and communication patterns among the actors using the social network analysis technique (Brandes and Erlebach, 2005; Vignola et al., 2013; Bae et al., 2015; Fatemi et al., 2021).

Step 1. Determining the type of analysis

In the present study, various social networks were studied and analyzed at organizational and social levels:

All of the respondents had an academic education, except for two individuals finishing their high school. Nearly half (54%) of the samples had a master’s degree, and bachelor’s degree and Ph.D. were the subsequent most frequent education levels achieved, respectively. Based on the findings, 50% of respondents were in their second decade of working. After that, the individuals with more working experience with more than 20 years of work experience constituted 28% of the sample, and the remaining 21% included young workers with less than 10 years of experience (Table 3). One third of respondents (28 individuals) were female experts and two thirds were male experts.

The age range of the respondents was 28 years, with a minimum and maximum age of between 26 and 54 years old. The average age of respondents was about 41 years and the highest recurrence age among the sample was 40 years. In terms of the degree of education, the minimum and maximum number of years of study was between 12 and 23 years to give a range of 11 years. The average study years of the studied sample was 18 years (to master’s degree). The results included a diverse range of occupational backgrounds, meaning that both people with 30 years of work experience and on the edge of retiring and freshly recruited staff with 1-years’ experience could be observed. The average years of employment was 15 years and a large number of people in the sample had an occupational history of 20 years.

The QAP correlation index was used to investigate the correlation between the studied networks (Table 8). The correlation between the networks of information exchange, coordination, and participation was positive and significant at 1% level. In addition, coordination and participation had the highest correlation, while information exchange and participation had the lowest correlation. Thus, if there is an exchange of information between the two organizational actors, the links of coordination (0.193) and participation (0.107) can established in the network.

The results of the study indicated that the density index in information exchange, coordination, and participation networks is less than average. In the participation network, the values of the density index are very low and coherence is the same among the networks (information exchange, coordination, and participation). Based on the results of this index, no dense network can be expected in the studied links, especially in the network of participation. Therefore, with regard to the direct relationship of social cohesion with density, the degree of social cohesion based on the matrix of participation is very weak. As a result, integrated water management activities face some challenges and effort should be made to increase density in the above-mentioned networks to form the system. Based on the results of link reciprocity among the organizational actors, the total amount of information exchange links and coordination was below the moderate level, and the level of interaction among organizational actors was very low. Based on these results, as well as that of the link transitivity index, all networks of information exchange, coordination, and participation has showed minimum sustainability. Some mutual links in the decision-making processes for water management activities should be conducted to strengthen organizational integrity.

The total network size indicated that nearly one-tenth of the expected links were found in the participation network. The value of this indicator in the networks of information exchange and coordination was more favorable, although less than half of the expected links were established in the above-mentioned networks. Based on the results of the research into the information exchange, coordination, and participation networks, the ratio of centrality increased on the basis of the input links to the output links. In other words, based on the input links related to coordination, the participation in the investigated organizational network was almost weak and was dependent on limited activists. However, the distribution of coordination and participation in terms of output links had less dependence on the presence of the actors with high centrality in the above-mentioned networks. On the other hand, the centrality of the entire network of information exchange, based on output and input links, had a more dispersed structure and were similar to each other, indicating the moderate influence of the actors with a central position in terms of reputation and influence. In other words, nearly half of the input and output links of the information exchange are available for central organizations and the rest of the links are shared with other organizations in the network of information exchange.