Sociocentric relationships drive nature-based enterprises in the human dominated Liuwa Plain National Park

Community network structures and social innovations that support the Sustainable Development Goals (SDGs) play an important role in the performance of nature-based enterprises such as fishing, weaving, and timber harvesting. However, limited empirical evidence exists on how relational dynamics influence these enterprises within contexts of climate change and persistent rural poverty. This study used household surveys and social network analysis to examine community network structures in the human-dominated protected landscape of Liuwa Plain National Park. Chord diagrams were applied to identify the drivers, barriers, and bridging factors shaping enterprise performance and collaboration among entrepreneurs. Gender and resource type were identified as factors related to sociocentric relationships, with inter-household collaborations exhibiting distinct gendered patterns. Households with fewer alliances exhibited greater diversification from fishing into alternative livelihoods, including weaving, caterpillar harvesting, beekeeping, and mushroom production. Enterprise performance was supported by income generation opportunities, traditional practices, and information exchange, while major constraints included climate change impacts, poor market access, limited external support, and weak cooperation among entrepreneurs. Strengthening adaptive capacity, collaborative governance, and social connectivity within shared landscapes is therefore critical for improving enterprise resilience. These findings provide network-based insights for conservation planning and SDG implementation, while underscoring the importance of continuous monitoring to support adaptive management under changing socio-ecological conditions.

Social interactions may reinforce or undermine livelihood and conservation goals within local communities, depending on stakeholders' interests and the quality of communication and engagement (Nunes and Abreu, 2020; Bodin et al., 2020; Ison et al., 2021). These sociocentric relationships, reflected through interactions among community members, influence the performance of enterprises by shaping both horizontal linkages within communities and vertical connections with external actors, such as government institutions and conservation non-governmental organizations (NGOs) (Jaafar et al., 2020). Understanding the structure and function of these social networks is therefore important for identifying opportunities, minimizing constraints, and strengthening enterprise performance under conditions of climate change and environmental degradation.

Within these networks, stakeholder perceptions further shape how communities engage in conservation, cooperate, diversify livelihoods, and respond to environmental and economic challenges (Baird and Gray, 2014; Rockenbauch and Sakdapolrak, 2017; Dapilah et al., 2020; Pereira et al., 2022), thereby influencing the resilience of nature-based enterprises (Botha, 2023). For instance, positive perceptions of trust, governance, and shared benefits may strengthen adaptive capacity and collective action (Dressel et al., 2020; Hulke and Diez, 2020), whereas negative perceptions can weaken collaboration and sustainability (Rodríguez-Rodríguez et al., 2021; Lim et al., 2022). Consequently, understanding stakeholder perspectives is essential for advancing Sustainable Development Goals (SDGs), particularly SDG 1 (No Poverty), SDG 2 (Zero Hunger), SDG 5 (Gender Equality), SDG 12 (Responsible Consumption and Production), SDG 13 (Climate Action), SDG 14 (Life Below Water), and SDG 15 (Life on Land). In this regard, enduring social innovation solutions, including nature-based enterprises, should be informed by stakeholder perceptions (Phillips et al., 2024). As such perceptions are translated into action through the decisions, interactions, and resource-use practices of individual actors, understanding the behavior of entrepreneurs within their social networks becomes critical for explaining how nature-based enterprises emerge, function, and adapt within social-ecological systems (SESs).

Accordingly, this study conceptualizes entrepreneurs as actors with diverse self-interests and influences that shape their interdependencies and actions regarding natural resource use (Mekuria et al., 2021). Entrepreneurs pursuing utilitarian benefits from ecosystem services operate through socially embedded norms and relationships expressed within social networks (Bodin et al., 2006; Oniki et al., 2020). In doing so, they conduct their enterprises within complex and dynamic SESs (McGinnis and Ostrom, 2014) characterized by functional resource heterogeneity (Hopcraft et al., 2010; Fynn et al., 2016).

We evaluate the social interactions in these shared landscapes based on the planned behavior theory, focusing on the understanding of actors' intentions about social interactions (Ajzen, 1991, 2012). Furthermore, we integrate social exchange theory that expresses actions in interpersonal relationships through decision-making powers over associated costs and benefits (Blau, 2017). We apply social networks analysis (SNA) to determine the nature of relationships that exist in the naturebased enterprises in Liuwa ecosystem. SNA can be applied to identify the central or peripheral stakeholders in conservation enterprises (Prell et al., 2009; Yang et al., 2024), and determine the social network structure and collaborative links influencing profitability of uses of natural resources (Kyriakopoulou and Xepapadeas, 2021). Based on the theory of change (ToC) that suggests and stresses achieving community enterprises over time and space in line with the set intervention goals (Boshoven et al., 2022), monitoring at landscape scale will generate new knowledge and insights.

The SNA conceptualizes a network as a system of interconnected nodes linked through ties, relationships, or interactions (Wey et al., 2008; Farine and Whitehead, 2015). These nodes may represent individuals, households, or groups, while the ties reflect patterns of association, cooperation, or information exchange among them. By examining these relationships, SNA provides important insights into how social interactions influence decision-making and natural resource use, particularly within egalitarian resource-dependent communities. In such settings, social relations can shape innovations that advance the goals of actors, while simultaneously influencing broader social and ecological outcomes (Antadze and Westley, 2012). Consequently, social interactions play a critical role in strengthening the resilience of SESs through adaptation, cooperation, and collective action (Folke et al., 2010).

The application of SNA in natural resource management, particularly within the wildlife sector, has evolved considerably over time (Wey et al., 2008; Krause et al., 2009; Sosa et al., 2020). Recent studies show increasing use of SNA in wildlife-related research and conservation practice (Webber and Wal, 2019). However, its application at broader landscape scales remains limited, despite its strong potential to generate insights for integrated and effective landscape-level natural resource management (Sueur et al., 2011; Pinter-Wollman et al., 2014; Farine, 2020). Conceptualization of interest-based collaborative governance – interactions amongst social actors with emphasis on relational social networks – can lead to resilience and sustainability of both actor communities and ecosystems faced by challenges (Folke et al., 2003; Olsson et al., 2004; Berkes, 2007). This study investigates relational factors, nature of social networks, and drivers, barriers and bridges affecting the natural resource enterprises in Liuwa Plain National Park (NP), western Zambia. We hypothesize that the performance of nature-based enterprises is influenced by the configuration of relational capital, the structural position of enterprises within social networks, and entrepreneurs' perceptions of drivers, barriers, and bridging mechanisms across the landscape. The study answers the questions:

What are the relational factors influencing nature-based enterprises?

What is the nature of social networks associated with nature-based enterprises?

How are drivers, barriers and bridges, perceived by the natural resource entrepreneurs, influencing the performance of these enterprises?

By developing models, infographics may assist with possible illustrations pointing to solutions that can be employed to enhance livelihoods and natural resource utilization under accelerated climate change in Liuwa Plain NP landscape.

Study area

The study was conducted in Liuwa Plain NP, located in western Zambia and covering an area of approximately 3,660 km 2 (Figure 1). The park lies within the broader Barotse landscape, an understudied biogeographical region of southern Africa (Linder et al., 2012). The Liuwa Plain NP landscape situates within Agro-Ecological Regional IIb, which is dominated by sandy soils and receives moderate mean annual rainfall of 800 mm−1,000 mm and has mean annual temperatures ranging from 8.1 °C to 34.9 °C, with high climatic vulnerability from reducing rainfall projected at 20%−30% points (Ngoma et al., 2021). The floodplain communities surrounding and within Liuwa Plain NP are largely impoverished and highly dependent on natural resources for their livelihoods (Tweddle et al., 2015; Banda et al., 2022).

Liuwa Plain NP represents a shared “working” landscape characterized by human–wildlife coexistence (Western et al., 2020), making it one of Zambia's unique protected areas. Established in 1972, the park is among the oldest conservation areas in Africa, with origins tracing back to the 19th century when Litunga (King) Lubosi Lewanika of the Lozi people designated the area as a royal hunting reserve (DNPW, 2019). Since 2003, the park has been jointly managed by African Parks in partnership with the Department of National Parks and Wildlife and the Barotse Royal Establishment (DNPW, 2019).

The Liuwa Plain NP lies in a seasonally inundated floodplain ecosystem characterized by rich faunal and floral diversity that supports livelihoods of local communities living within and around the landscape through fisheries, agriculture and other natural resource-based activities (Tweddle et al., 2015; Banda et al., 2022). The General Management Plan (GMP) for Liuwa Plain NP documents a wide array of biophysical aspects of Liuwa Plain NP (DNPW, 2019). The area is a renowned Important Bird and Biodiversity Area (IBA) for its high species diversity of mammals, birds, reptiles, insects and amphibians (Leonard, 2005). Notably, the migratory Blue Wildebeest (Connochaetes taurinus, Burchell 1823) population of Liuwa Plain NP is considered second to that of Serengeti in East Africa (Estes and East, 2009). The Liuwa Plain NP is drained by two large rivers (Luambimba and Luanginga), which are the tributaries to the Zambezi River. The vegetation of economic importance includes the reeds (Phragmites australis, Cavanilles 1799), papyrus sedges (Cyperus papyrus, Linnaeus 1753) and clusters of raffia palms (Raphia farinifera, Gärtner 1791) used in making local mats and baskets. The ecosystem inadvertently faces several environmental threats inter alia fish overharvesting, climate change, poaching, wildlife poisoning, agricultural pollution, invasive species, wildfires and illegal timber harvesting (DNPW, 2019).

The Liuwa Plain NP local communities are engaged in multiple natural resource enterprises that include fishing, weaving, and beekeeping (Banda et al., 2022). There are also 450 designated fish pans in Liuwa ecosystem owned by individuals, families, the Senior Chief Mboanjikana at Libonda palace and BRE Litunga in Mongu. Specific headmen administer some fish pans for Senior Chieftainess Mboanjikana and the Litunga. The community-based tourism initiatives are a notable collective community enterprise supported by the AP, and both government and BRE traditional leadership have hegemony over the natural resource harvesting (Nyirenda and Nkhata, 2013). The local entrepreneurs sell locally within the locations where the resources are harvested within Liuwa Plain NP as well as at Kalabo market (Baidu-Forson et al., 2014). Inside the park the only conservation NGO supporting entrepreneurs is AP and Worldwide Fund for Nature (WWF) funds some activities that AP implements.

Outside the park the entrepreneurs are technically supported by the multiple stakeholders inter alia government departments, timber companies, fish companies, and civil society (DNPW, 2019). The Kalabo forestry department has two (2) groups of timber concessionaires focusing and confining their timber harvesting outside the park. Furthermore, the forest department also have eight (8) groups involved in non-forestry products activities, such as mushroom growing/harvesting and caterpillar harvesting taking place in isolated non-protected areas. The Fisheries department has some individuals involved in aquaculture activities supported by NGOs, such as Scaling-Up Nutrition.

Data sampling

To investigate aspects affecting nature-based enterprises in Liuwa Plain NP, we randomly sampled 206 out of the 356 required households, as a sampling unit, in Liuwa Plain NP, derived from entrepreneurs' registers because the remainder were inaccessible due to remote terrain and poor infrastructure that restricted travels. The 356-household sample size was calculated with 95% confidence level and 5% margin of error (e), determined as the sampling size using the method stipulated by Krejcie and Morgan (1970). With human population in the Liuwa Plain NP (Liuwa constituency of the Kalabo district) estimated at 24, 318 inhabitants (52% females and 48% male) occurring in 4, 840 households, 206 households were actually surveyed. The study targeted 356 households based on a priori sample size calculations using a 95% confidence level and a 5% margin of error. Due to accessibility constraints in several remote settlements, data were successfully collected from 206 households (57.9% of the target sample). While this shortfall reduced the statistical precision of the estimates and increased the margin of error to approximately 6.8%−7.0%, the achieved sample size remained sufficiently large for the planned descriptive and inferential analyses. The results should, however, be interpreted with caution, particularly with respect to generalizing findings to households located in the most remote and inaccessible areas that were underrepresented in the final sample.

In addition to the four Village Action Groups—VAGs (Lumei, Mishulundu, Mulinga, and Mukola), representing grassroot functional units for the community-based natural resource management (CBNRM) within the park, people (n = 36, 17.5%) from Kalabo market involved in the nature-based enterprises were also included in the sample size. These people were included because they belonged to the same analytical system having originated from the communities inside the park, as the study adopted a network-based approach where entrepreneurial interactions, rather than merely administrative boundaries, defined participant inclusion. To avoid confounding data with socioecological variations resulting from changing social-ecological conditions over time, data were collected at once in September, 2024. We adopted the case-study based on three-tier criteria postulated by Leach et al. (2002), where the respondents are: (1) active participants in nature-based enterprises; (2) successfully implemented nature-based enterprises; and (3) perceiving impacts of the nature-based enterprises.

Data collection

This study was conducted under the ethical clearance No. 2023/2/13/1 issued by The Copperbelt University, and research permit No. NPW/8/27/1 issued by Department of National Parks and Wildlife (DNPW). Both male-headed and female-headed household natural resource entrepreneurs were targeted in this survey using Kobo Collect—an open-source application to input responses from the questionnaire (Appendix A1). Only people above the age 18 (nationally recognized adult age) were allowed to participate in the survey. The questions in the questionnaire were workshopped for clarity among carefully selected and trained nine research assistants. Prior to the survey administration, these questions were further pre-tested for validation to 30 respondents, who were not part of the final survey. The research assistants were familiar with the local languages in the study area, and sought free, prior informed consent from the survey respondents. Thereafter, research assistants spent approximately 30 min−40 mins to complete a survey with these voluntary respondents. The survey was administered in English or vernacular, whichever convenient language to the respondents. We applied a five-point Likert-type scale to rank the responses from the respondents (Milfont and Duckitt, 2010; Cruz and Manata, 2020). Prior to conducting the survey, the respondents were guided on how to properly use the Likert scale for ranking purposes.

Data analysis

Data were analyzed in R statistical software version 4.4.0 (R Core Team, Vienna, Austria) using the igraph, tidygraph, ggraph, tidyverse, and faraway packages (R Core Development Team, 2024). To assess potential multicollinearity among predictor variables (i.e., VAG, gender, age, marital status, ethnicity, household size, education level attained, and type of enterprise), variance inflation factors (VIFs) were calculated. Multicollinearity can inflate standard errors and reduce the reliability of regression coefficient estimates. Predictor variables with VIF values below two were retained, indicating very low collinearity among explanatory variables. Although higher thresholds, such as VIF <5 are commonly accepted in regression analyses (Jeng, 2023), a more conservative threshold of two was adopted in this study to minimize even moderate collinearity effects and improve interpretability of predictor contributions. The VIF was calculated according to (Equation 1):

where R i 2 represents the coefficient of determination obtained by regressing the i th predictor variable against all other predictor variables in the model. Higher R i 2 values indicate that a predictor is highly explained by other explanatory variables, suggesting increased multicollinearity.

The social networks were constructed from reported interactions among entrepreneurs. Each entrepreneur was represented as a node, and a tie was established when a respondent identified a relationship with another entrepreneur. Responses measured on a five-point Likert scale were converted into binary outcomes for proportion testing. “Agree” and “Strongly Agree” responses were classified as positive responses (coded 1), whereas “Neutral,” “Disagree,” and “Strongly Disagree” responses were classified as non-positive responses (coded 0). The strength of a relationship among the entrepreneurs was considered as a product of exchange, conflict, cooperation and accommodation. In determining the structure of social networks (Bodin et al., 2006; Bodin and Crona, 2009), for nature-based enterprises in Liuwa Plain NP, the following measures of network structures were quantified and analyzed: density of relationships/connections, closeness and betweenness of the relationships. Identifying network structures of closely or sparsely connected relationships for partitioned latent functions (Jin et al., 2023), community detection was employed to reveal clusters of discrete and overlapping relationships among the natural resource entrepreneurs. This distinction is important for providing insights into developing or strengthening sustainable interventions associated with nature-based enterprises.

For each factor, the proportion of respondents expressing agreement was calculated as the number of positive responses divided by the total number of valid responses to that factor. Relationships were classified as either alliance ties (cooperative interactions such as information sharing, joint activities, or resource exchange) or competition ties (rivalry for markets, resources, or customers). Ties were unweighted as interaction intensity could not clearly be defined, but coded as binary (presence/absence, 1 or 0, respectively) and treated as undirected because the analysis focused on mutual relationships rather than the direction of exchange. Alliance and competition ties were stored as categorical edge attributes and visualized using purple and red edges, respectively.

Prop.test function in R-statistical software was used to determine the level of statistical significance among diverse drivers, barriers and bridges in the proportion of responses provided by the natural resource entrepreneurs. A one-sample proportion test with Yates continuity correction was then used to determine whether the observed proportion of agreement significantly exceeded the contribution of other proportions. Denominator was the total number of valid respondents for the specific statement or factor. Each respondent contributed to only one driver, barrier, or bridge.

Demographic characteristics

There was a wide range of demographic characteristics of sample distribution in Zambia's

Liuwa Plain NP (Table 1). Both females and males were equitably sampled (n = 112, females; n = 94, males), having insignificant proportional differences between them for the statistical test (χ 2 = 1.403, df = 1, p -value = 0.118). Furthermore, there was a wide age range of the community natural resource entrepreneurs (19–90 years old), with diverse marital status, ethnicity, household sizes, education levels and enterprise types.

Network measures

The Liuwa Plain NP network of natural resource entrepreneurs was characterized by individuals maintaining between 1 and 8 ties (Figure S1). This network's density of connections was very low at 0.7%, with 2.7% reciprocity, due to collective effects of multiple social factors, such as inadequate cooperation, lack of trust, negative attitudes, and highly seasonal floodplain system on social structure and sociocentricity. The unweighted closeness was expressed by all individuals (alters) and involved at least 1% of the nature-based enterprise actors each. The potential betweenness (i.e., edge betweenness) within this network was expected to vary between 1 and 2, 906 relationships in this network of 206 individuals, and the actual degrees of connections between the individual enterprise actors (nodes) totalled 590 connections.

Social networks in nature-based enterprises

Figure S2 aggregates relationships in various community-based nature-based enterprises in the Liuwa Plain NP in 2024 into multiple overlapping clusters representing different types of ties with similar interests, structures and spatial distributions. Later, we decompose these interests, and structures and spatial distributions in the following analyses. Furthermore, with the directed ties, networks reveal that only few individuals (nodes) nominate several actors or receive nominations by several actors directly supporting or competing with them in their nature-based enterprises (Figure S3).

The largest ties occurred among male entrepreneurs (Figure 2). Both females and males fostered alliances and competition in their enterprises, though most of the competition occurred involving males. Thus, each participant depended on a particular relationship for their enterprise. Nearly all participants had an inter- or intra-alliance in their enterprise. Individuals (both females and males) with high numbers of ties with other entrepreneurs were exclusively in alliance with others, and not in competition. Furthermore, these ties involved specific enterprise types (Figure 3). Entrepreneurs commonly mixed fishing with other enterprises, such as weaving, timber and beekeeping. In particular cases where individuals have less endowment of alliances, there were complete shifts from fishing to alternative enterprises, leaving individuals endowed with alliances still fish dependent for their enterprises (Figure 3). There were remarkable differences between gender and household size, with more diversity of enterprises in females with household sizes between 2 and 9, while diversity of enterprises was more elaborate in household sizes between 4 and 9 in males, with varying levels of relationships (Figure 4). Inter-household size collaborative relationships in enterprises occurred among females of household size of 3 with varying enterprises, and males with household size of 9, specializing in fish enterprises (Figure 4).

The Lozi ethnic group who had attained primary or secondary education levels dominated the relationships among the entrepreneurs (Table 1 and Figure 5). Alliances were typically formed among female entrepreneurs (Figure 5). With exception of Lozi, Luvale and Mbunda ethnic groups, age distribution across gender was under-represented in nature-based enterprises. There is a concentration of intra-ethnic ties among Lozi ethnic group, unlike all other groups (Figure 6). Based on a mix of enterprises, the trend was toward diversification from the historic focus on fishing enterprises across the VAGs (Figure 7). In Lumei and Mishulundu VAGs, and other areas, such as Kalabo market, there were competition relationships driving the enterprises alongside intra-VAG entrepreneurial alliances (Figure 7). Stronger ties existed among male- and female-based sociocentric clusters at Mukola and Mulinga VAGs and other areas, such as Kalabo market. However, competition relationships were only in male entrepreneurs in Lumei and Mishulundu VAGs, and other areas, such as Kalabo market (Figure 7).

Community social Interactions' drivers, barriers and bridges

Drivers

The data reflected by Figure 8 identifies several drivers of enterprises in Liuwa Plain NP. By ranking using Likert scale, the most important driver for the social interactions regarding naturebased enterprises were income generation for individuals (χ 2 = 42.241, df = 1, p -value < 0.001). Similarly, the contribution of financial capital, as determined from the Likert-based rankings, was statistically significantly higher than the combined contributions of social, cultural, and natural capitals as drivers of nature-based enterprises (χ 2 = 9.218, df = 1, p -value = 0.001). Within the financial capital space, this driver was supported by local market availability, water transport and the limited existence of social amenities, such as credit facilities. The second most important driver was traditional practices within the strong cultural setting (cultural capital) of the area, where enterprises were governed by strong influence of the BRE. The third most important driver was information sharing within existing social capital of the ties. Other minor contributing aspects of social capital driving the enterprises in the area included social innovation, cooperation among the actors, government support in form of materials and skills transfer, shared vision to better actors' well-being, and self-motivation to uplift oneself from poverty. Furthermore, the availability of resources (natural capital), such as fish, reeds and timber also to a lesser extent drove the performance of enterprise in Liuwa Plain NP. In particular, the timber comes from areas, such as Lueti in Sikongo District, and not from Liuwa Plain NP.

Barriers

Several barriers and factors were associated with the nature-based enterprises in Liuwa Plain NP (Figure 9). Negative effects of climate change were recognized by the nature-based entrepreneurs as the most important environmental-related barrier for the development of enterprises, though not significantly more important than inaccessibility to markets (χ 2 = 1.000, df = 1, p -value = 0.159). However, when combined—the negative effects of climate change and inaccessibility to markets were the most important barriers compared to non-cooperation or inadequate external support or any individual factor (χ 2 = 32.051, df = 1, p -value < 0.001).

Other reported environmental-related barriers included the seasonally inundated terrain and persistent human-wildlife conflicts involving crocodiles, hyenas, and lions. Seasonal flooding often restricts mobility, damages infrastructure, and limits access to fishing grounds, grazing areas, and enterprise locations. The flooded landscape also increases transportation costs and reduces the ability of entrepreneurs to move goods and services efficiently. In addition, the presence of dangerous wildlife creates insecurity among local communities, discourages participation in livelihood activities, and increases the risk of livestock losses, and human injury. These environmental pressures collectively reduce productivity and constrain the expansion of nature-based enterprises.

Secondly, limited access to markets hindered enterprise development among entrepreneurs. Poor transport networks, long travel distances, seasonal inaccessibility, and inadequate market infrastructure restricted the movement of products to profitable markets. As a result, entrepreneurs often faced low product prices, reduced bargaining power, high transaction costs, and limited opportunities to diversify their businesses. Among all the identified constraints, governance-related barriers emerged as a key factor, particularly those linked to inadequate infrastructure, weak market systems, and limited institutional support. Other governance-related barriers included inadequate external support by other actors, such as cooperatives and clubs, lack of transparency among some leaders, inequitable access to natural resources, and failure to manage human-human conflicts over resource exploitation.

These were followed by socio-economic barriers, such as non-cooperation, inadequate communication, laziness, lack of trust, poverty, lack of motivation, negative attitudes, loadshedding and language, all of which further constrained enterprise growth and sustainability. Lack or inadequate financing facilities and equipment are the critical financial factors contributing to the barriers for the enterprises in Liuwa Plain NP. Furthermore, lack of skills, and inappropriate fishing methods are also technical factors that add to the list of barriers in the area. However, some respondents (n = 13, 6.3%) were uncertain of the barriers or factors associated with enterprises they engaged in.

Bridges

This study revealed three pillars (enablers) namely, adaptive capacity; good governance; and traditional practices, for building nature-based enterprise in Liuwa Plain NP (Figure 10). Adaptive capacity as major bridging variable involved cooperation, trust, self-motivation, functional cooperatives and clubs, peace, negotiation skills, and good relations with customers. These adaptive capacity factors were more statistically significant than good governance (χ 2 = 8.733, df = 1, p -value = 0.002), as a perceived bridging factor. The governance factors were also more statistically significant than traditional practices (χ 2 = 17.434, df = 1, p -value < 0.001). Governance variables included communication among the actors (for instance, entrepreneurs, government, conservation non-governmental organizations, and traditional leadership), external support to entrepreneurs, accountability among the leadership, skills transfer, transparency, transport infrastructure, and financial prudence. Traditional practices embedded in the cultural values and co-adaptation of the local communities over resource use were also a bridging variable for enterprises. These practices, though widely co-occurring, were viewed by the respondents as of less effect compared to adaptive capacity and governance.

This study demonstrates that social networks in nature-based enterprises within Liuwa NP varied across gender roles and resource types. Male entrepreneurs exhibited high levels of associations and competition, often assuming leadership roles in enterprise relationships (Milupi et al., 2020). Cultural norms and economic expectations may account for this dominance. Conversely, women—despite their essential roles in harvesting, production, and marketing—remain structurally marginalized, particularly in decision-making and access to enterprise resources (Schneider, 2013; Rajaratnam et al., 2015; Sallah and Caesar, 2020).

Inter- or intra-alliance networks played a pivotal role in enterprise sustainability. Entrepreneurs embedded in strong familial and kinship networks were more inclined toward collaboration, facilitating innovation, conflict resolution, and market access (Wei et al., 2019; Li and Liu, 2023). However, such networks also fostered exclusivity, with elite alliances enabling a few enterprises to thrive, while others remained excluded (Sydow et al., 2022). Rent-seeking behavior by traditional leaders exacerbated exclusion and undermined equitable growth (Benjaminsen et al., 2013; Rajaratnam et al., 2015), underscoring the need for inclusive, “green” leadership (Chen and Yan, 2022). In the study area, traditional governance structures strongly influence the formation of alliance networks particularly within fishing communities through kinship systems, customary authority, and sociocentric social relations. While these systems would promote trust, reciprocity, and cooperation among established groups, they also tend to reinforce bonding social capital that privileges insiders, such as dominant clans, influential families, and politically connected fishers. As a result, marginalized groups including migrants, women, and poorer fishers may be excluded from decision making processes and access to fisheries resources. The sociocentric nature of these relationships can therefore reproduce existing power imbalances and limit the development of broader, more inclusive social networks.

Institutional barriers, such as the use of inappropriate or illegal fishing methods further interact with social capital in complex ways. Although strong community ties can support collective resource management and local enforcement, the same networks may protect offenders through kinship loyalty and patron-client relationships, weakening accountability. In many cases, traditional leaders selectively enforce fisheries regulations, particularly where elite capture interests are involved. Consequently, illegal fishing practices are not only a regulatory problem, but also a reflection of unequal social relations, poverty, and overlapping customary and state governance systems. This highlights the need for more inclusive co-management approaches that strengthen bridging social capital, and improve transparency and equity in fisheries governance.

Environmental stressors, such as climate change and environmental degradation, within SES influence social relations by reshaping access to resources, patterns of dependence, and the incentives for collective action. In fisheries-dependent communities, declining fish stocks, seasonal variability, or ecological degradation alter both the material and institutional conditions under which people interact. Resource scarcity may intensify competition as individuals or groups compete over diminishing fishing grounds, markets, and livelihood opportunities. Under such conditions, social tensions often emerge through exclusionary practices, weakened trust, elite capture, illegal fishing, or conflicts between local users and external actors. In western Zambia, for example, the increased use of inappropriate fishing methods can be interpreted not merely as an environmental problem, but also as a symptom of livelihood insecurity and weak governance. Such stressors may reinforce existing power asymmetries, where wealthier or politically connected actors are better able to secure access to productive fishing areas, equipment, or regulatory influence.

At the same time, environmental stress does not always produce fragmentation; it can also stimulate new forms of cooperation and adaptive alliances. Within the SES framework, communities facing ecological uncertainty often mobilize social capital to reduce risk through collective action, information sharing, reciprocal labor arrangements, or co-management institutions. Fishers may form alliances to regulate access, monitor illegal activities, diversify livelihoods, or negotiate with state institutions and NGOs. However, the effectiveness of these cooperative responses depends heavily on governance quality, trust, and inclusiveness. In sociocentric systems, where social ties are strongly embedded in kinship, ethnicity, or traditional authority structures, cooperation may strengthen resilience for some groups, while simultaneously marginalizing others. Consequently, environmental stressors interact dynamically with institutional arrangements and social networks, producing outcomes that may either deepen inequality and competition or foster collective adaptation and resilience.

Diversification emerged as a key adaptive strategy amid ecological pressures, especially declining fish stocks due to overfishing and insufficient enforcement (Peel et al., 2013; Tweddle et al., 2015). This dual trajectory mirrors trends observed in other natural resource-dependent regions (Adekola and Clelland, 2019; Dey et al., 2019; Ruiz-Ballesteros and del Campo Tejedor, 2020). Although enterprise types were influenced by resource availability, the relative uniformity of resources across sites resulted in minimal locational variation, diverging from findings by Abreu et al. (2019).

Variation in participation by the entrepreneurs was observed across gender and household sizes. Men displayed more competitive behavior, whereas women—especially from medium-sized households—pursued more diversified enterprises. These patterns reflect culturally defined roles, though increasing modernization and supportive policies may promote greater gender equity (Nielsen and Reenberg, 2010; Igwe et al., 2020; Kalinda et al., 2022). Ethnic and household characteristics also shaped enterprise development. The Lozi formed dense intra-ethnic networks, unlike other ethnic groups, such as the Luvale and Mbunda. Older entrepreneurs, though less agile, sustained participation via strong social capital, while lower education levels constrained strategy and innovation, giving a regional character (Rocha et al., 2020). In addressing the underrepresentation of certain ethnic groups, and continued promotion of inter-gender roles to strengthen community-based nature-based enterprises, there is a need for targeted and engendered interventions, positively contributing to gender inclusiveness [SDG 5].

Key drivers of enterprise growth included income generation, access to markets, water transport, and financial services. Traditional institutions, particularly the BRE, played a central role in resource governance and enterprise compliance. Information exchange, government support, social innovation, and peer cooperation were additional enablers. Traditional fishing practices, for instance, are supported through management measures to ensure healthy fisheries. Park management aids these practices by providing fish-drying facilities and transportation to local and regional markets, thereby supporting income generation and market access. These findings underscore the multifaceted nature of drivers influencing nature-based enterprises in Liuwa Plain NP, highlighting the interaction between economic, cultural, social, and natural capitals in shaping enterprise performance.

However, multiple barriers hindered enterprise growth. Entrepreneurs collectively identified climate change as the most severe environmental challenge, along seasonal flooding and humanwildlife conflicts. This assertion aligns with Li et al. (2024), who acclaimed that climate change reduces resource endowment, and requires adaptive behavior to mitigate its effects. Market isolation, especially in Kalabo, and weak cooperative structures limited growth. Governance-related obstacles included inadequate support from cooperatives and clubs, lack of transparency among some government and traditional leaders, inequitable resource access, and ineffective management of human conflicts over resource exploitation. Socio-economic factors, such as non-cooperation, poor communication, lack of trust, poverty, low motivation, negative attitudes, and language barriers also contributed to these challenges. Financial constraints, including limited access to financing facilities and equipment, contributed to inability by the natural resource entrepreneurs to add value to their products, and exacerbated the underdevelopment of nature-based enterprises. Technical issues, such as insufficient skills and the use of inappropriate fishing methods by the local fishers, further hindered enterprise development. Addressing these multi-faceted barriers requires concerted and comprehensive approaches that encompass environmental management, infrastructure development, governance reforms, socio-economic improvements, financial support, and technical training to enhance the viability of nature-based enterprises in Liuwa Plain NP.

This study identified three key pillars or enablers, critical for fostering nature-based enterprises in Liuwa Plain National Park. The first pillar, adaptive capacity, emerged as a key bridging variable, encompassing elements, such as cooperation, trust, self-motivation, functional cooperatives and clubs, peace, negotiation skills, and positive relationships with customers. These factors possibly enhanced the community's resilience and ability to navigate challenges in enterprise development (Folke et al., 2010; Koliou et al., 2020; Ruiz-Ballesteros and del Campo Tejedor, 2020). The second pillar, good governance, was highly valued by respondents as a crucial enabler. Key governance variables included effective communication among stakeholders (entrepreneurs, government agencies, conservation NGOs, and traditional leaders), external support for entrepreneurs, leadership accountability, skills transfer, transparency, robust transport infrastructure, and financial prudence. These factors contributed to an enabling environment for sustainable enterprise growth (Ostrom, 2009; Lockwood, 2010). The third pillar highlighted the importance of traditional practices, deeply rooted in the cultural values of local communities. These practices possibly served as a bridging variable by aligning enterprise activities with the community's cultural norms and values, ensuring broader acceptance and sustainability (Charina et al., 2022). These three pillars—adaptive capacity, good governance, and traditional practices—underscore the multifaceted approach required to develop sustainable nature-based enterprises in protected areas like Liuwa Plain NP.

Compared with studies from other African transfrontier conservation areas (TFCAs) and human-dominated protected landscapes, the Liuwa Plain NP study provides a more integrated understanding of how social networks, traditional governance, gender relations, and environmental stressors interact within a floodplain SES. Similar studies in the Kavango–Zambezi Transfrontier Conservation Area (KAZA TFCA), and Great Virunga Landscape in Rwanda, and Zimbabwe's CAMPFIRE programme have highlighted the importance of social capital, livelihood diversification, and community institutions in conservation enterprises (Child et al., 2009; Mbereko et al., 2017; Sabuhoro et al., 2023; Lekgau and Tichaawa, 2025). However, most of these studies focus primarily on wildlife tourism, conservancy governance, or household adaptation. In contrast, the Liuwa study uniquely demonstrates how largely fisheries-based enterprise networks are embedded within kinshiporiented alliance systems that simultaneously promote cooperation, innovation, and resilience while reinforcing exclusion, elite capture, and unequal access to resources. Therefore, the study advances protected area scholarship by showing how bonding social capital within sociocentric fishing communities may strengthen collective action for some groups, while marginalizing women, migrants, and poorer fishers.

Furthermore, the study also contributes uniquely to debates on governance and environmental change in African protected landscapes, where traditional authorities and local elites frequently mediate access to conservation benefits, often reinforcing gender inequalities and power asymmetries (Matose et al., 2024). The Liuwa study extends these findings by illustrating how kinship loyalty and patron-client relationships may also weaken fisheries enforcement and protect illegal fishing practices, demonstrating that social cohesion can function simultaneously as both a governance asset and a source of regulatory failure. Unlike many African conservation studies that treat environmental stress mainly as a livelihood challenge (Adebayo, 2024; Pereira et al., 2024), this study conceptualizes climate variability, seasonal flooding, and declining fish stocks as drivers that dynamically reshape social relations, alliance formation, cooperation, and conflict within entrepreneurial networks. The explicit integration of social network analysis, environmental stress, and traditional governance within a highly seasonal floodplain SES, therefore, represents a key empirical and conceptual contribution to understanding resilience, adaptive capacity, and enterprise sustainability in African conservation landscapes.

Though the study has several merits, the following inherent limitations exist. First, a key limitation of this study is its cross-sectional design, which captures only a single temporal snapshot of a highly dynamic floodplain SES. In study area, seasonal flooding affects livelihoods, mobility, market access, and social interactions, meaning that the identified drivers and barriers may vary across seasons. During the flood season, factors such as water transport, fishing income, traditional practices, information sharing, and actor cooperation may become more important as communities adapt collectively to changing environmental conditions and resource availability. Conversely, seasonal flooding may also intensify barriers including climate-related vulnerability, poor market accessibility, difficult terrain, human-wildlife conflicts, weak communication, inequitable resource access, poverty, and limited external support. Inappropriate fishing methods and governance challenges may become more pronounced under increased livelihood pressure and weak enforcement during peak fishing periods. Therefore, a longitudinal or seasonal approach would provide a more comprehensive understanding of how environmental variability shapes social networks, cooperation, and resilience within the SES framework. Second, although the achieved sample size was smaller than originally planned, the 206 completed household interviews still provided a substantial dataset for the analysis. Nevertheless, the reduced sample size has important methodological implications: First, the lower number of observations decreases statistical power, thereby reducing the ability to detect significant relationships or differences where they exist. Second, the precision of population estimates is reduced, resulting in wider confidence intervals and a larger sampling error than originally anticipated under the 356-household design. Third, because inaccessible areas were underrepresented, there is a possibility of non-response or coverage bias if households in these locations systematically differ from those that were surveyed. Other limitations include:1. The study did not examine further the antecedents and mediating factors among the drivers, barriers and bridging effects from entrepreneurs' perspectives; 2. It did not also leverage time-series changes in social networks across temporal dynamics; 3. Furthermore, study was restricted to study site that had one dominating ethnic group, the Lozi-speaking people, with strong cultural traditions. As such, our results may be different from areas where there is a stronger mix of ethnicity. Future research should adopt longitudinal and multi-ethnic approaches to enhance generalizability.

Using the Liuwa Plain NP as a case study, we have demonstrated the complexity of naturebased enterprises. Here, the sociocentric relationships underscore the underlying elements characterizing community-based nature-based enterprises. Based on the novel knowledge generated, incorporating multiple factors in the social-ecological system as in Liuwa Plain NP could help enhance nature-based enterprises and positively contribute to entrepreneurs' well-being. Aligning interventions with the identified key factors can help promote policies and practices around sustainable nature-based enterprises. This approach is expected to favor local transformative behavioral change toward sustainable livelihoods and conservation strategies, amidst risks of environmental challenges, such as climate change and biodiversity loss. Therefore, the growth of sustainable nature-based enterprises should consider community social relational needs around livelihood and conservation strategies, alongside positive behavior change. Such understanding about diverse landscape-based nature-based enterprises can further open dialogue and collaboration, vehemently tackling environmental challenges, such as climate change and biodiversity losses, and avoiding protracted and emerging conservation conflicts, and livelihood risks among local communities. Together, these findings underscore the need to understand what underlies sociocentric relationships in nature-based enterprises at landscape level in relation to SDG 1 [No poverty] (Gajdosikiene, 2008; Filho et al., 2021), SDG 2 [Zero hunger] (Owens et al., 2022), SDG 5 [gender equity] (Dengah et al., 2022), SDG 12 [Sustainable consumption and production] (Glavič, 2021), SDG 13 [Climate action](Kawamoto and Kanie, 2020; Smith et al., 2021), SDG 14 [Life below water] (Brunet et al., 2020) and SDG 15 [Life on land] (Esteves et al., 2021). Understanding sociocentric relationships, based on long-term research for knowledge production, is needed to inform effective and sustainable development and conservation planning, implementation, and evaluation.