Introduction

Arctic Indigenous food systems are typified by diverse social interactions, facilitating flows of food and other resources among community members and along extended family lines (Baggio et al. 2016). In modern Arctic Indigenous communities, these interactions manifest in mixed economies. Mixed economies describe the persistence of traditional subsistence activities and food-sharing networks in communities otherwise integrated into wider market structures (Wolfe and Walker 1987). This duality enables diverse food provisioning strategies, pairing culturally relevant and nutrient-rich subsistence foods with store-bought foods to maintain food security (Egeland et al. 2011; Ford and Beaumier 2011; Collings et al. 2016). Definitions of Arctic Indigenous food security include the entire Arctic ecosystem and the relationships between all components within, whilst emphasizing the important role of diversity in these systems, not only for instrumental food security but also for maintaining traditional practices to obtain, process, store, and consume traditional foods (ICC-AK 2015; ACSPI 2023) (Appendix A).

Today, the diversity of these systems is under threat. Challenges such as climate change (Berkes and Jolly 2001; Ford et al. 2009), declining subsistence species (Kenny et al. 2018a), escalating food prices (Kenny et al. 2018b), and the erosion of indigenous, local, and traditional knowledge (ILTK) (Soloway 2015; Parlato et al. under review) (Appendix A) undermine mixed food systems and lead to high levels of food insecurity for the communities that rely on them (Council of Canadian Academies 2014). There is a delicate balance between maintaining traditional Indigenous practices, such as subsistence hunting and fishing, or food-sharing, and market-based food provision in Arctic Indigenous food systems. The latter has an important role in food security but can also threaten Indigenous traditions (i.e., subsistence access to traditional foods) (Kuhnlein et al. 2004; Little et al. 2021).

Given the multifaceted instrumental (i.e., values to the benefit of people), relational (i.e., values related to individual and collective relations), and intrinsic (i.e., values in, of, and for themselves) values inherent in Indigenous food system interactions (Díaz et al. 2015; Arias-Arévalo et al. 2017) and the challenges these systems face, it is important to understand how this diversity is impacted by change. Previous studies have examined the instrumental importance of diversity by examining how it contributes to system resilience. Systems with diverse components are often more resilient, as diversity provides redundancy within a system by allowing some components to compensate for the loss or failure of others (Biggs et al. 2015; Grêt-Regamey et al. 2019). However, the diversity of Indigenous food system interactions says more than its instrumental role in building resilience. It also captures the relational and intrinsic values central to Indigenous food security (ICC-AK 2015; ACSPI 2023). Thus, understanding this diversity is crucial for Indigenous communities across the Arctic, which urgently need effective solutions to navigate the multifaceted challenges of the twenty-first century and safeguard Indigenous food systems (Zimmermann et al. 2023).

While subsistence foods undeniably remain of high value in Arctic Indigenous food systems, it is important to recognize that these are mixed food systems, where engagement in employment limits the time available for subsistence activities (Aslaksen et al. 2008), leading to increased reliance on store-bought foods. Additionally, as subsistence species decline and climate change disrupts traditional means of travel and access (Hauser et al. 2021), the ability to shift to alternative food sources becomes increasingly vital (Wesche and Chan 2010). At the same time, access to healthy and nutritious market foods is often inadequate due to economic factors (Lambden et al. 2006) or logistical challenges (Kenny et al. 2018b). Recognizing the importance of diverse interactions in mixed Arctic Indigenous food systems, we argue that identifying threats and amplifiers to food system diversity is crucial.

Network analysis provides a powerful tool for analyzing diverse interactions in social-ecological systems (SES) (Sayles et al. 2019). In Arctic Indigenous food systems research, most network studies have focused on traditional food-sharing (Collings 2011; Collings et al. 2016; Scaggs et al. 2021) and applied network analysis to investigate food system robustness (Baggio et al. 2016). Until now, research has neither focused on the diversity of these networks nor on understanding how Indigenous food security is achieved through diverse social-ecological interactions. Here, we use the Shannon Diversity Index (SDI) to measure the diversity of food-related interactions. The SDI is a common measure of diversity within systems and has been applied widely in ecology to understand species diversity (Morris et al. 2014). It has also been applied in a network context to understand diversity within networked systems and how the loss of nodes or links can impact system diversity (Klein et al. 2021).

The Indigenous food system in question is on St. Paul Island, a small island located in the southeastern Bering Sea. Its native Unangax̂ (Aleut) community relies on a mix of store-bought foods purchased online or at the local community store and various marine and terrestrial subsistence species. This mixed St. Paul Island food system thus includes traditional practices typified by interactions between the community and their environment and market-based interactions typified by interactions between the community and local entities, institutions, or companies. We formulate three main objectives for this research: (i) to assess the diversity of food-related interactions between different actor groups in the mixed local food system on St. Paul Island, (ii) to identify key food security challenges and investigate how they impact the diversity of food-related interactions in the food system, and (iii) to simulate how food system diversity can be increased through different interventions. Through this research, we hope to contribute to finding strategies for supporting Arctic Indigenous communities in their food security and the resilience of their food systems.

Methods

St. Paul Island’s mixed food system

St. Paul Island is the largest of the five Pribilof Islands (Fig. 1), located in the southeastern Bering Sea about ~ 400 km (250 mi) north of the Aleutian Chain and ~ 480 km (300 mi) west of the Alaska mainland. St. Paul has one small village with an estimated 399 residents, with > 85% identifying as Unangax̂ or Alaska Native (World Population Review 2023). Like many Indigenous communities, the Aleut Community of St. Paul Island (ACSPI) relies on a mixed economy, combining a cash-based wage economy with subsistence harvesting and sharing resources (Kruse et al. 2008). The St. Paul Island food system combines the Aleut Community Store and the subsistence harvest of laaqudax̂ (northern fur seal, Callorhinus ursinus), qawax̂ (Steller sea lion, Eumatopias jubatus), itayax̂ (domesticated reindeer, Rangifer tarandus tarandus), chagix̂ (Pacific halibut, Hippoglogssus stenolepis), qimgiitan (multiple crab species), and intertidal foods such as sea urchins, berries, and sometimes seabirds and eggs. According to the St. Paul food security assessment performed in August 2022, St. Paul residents mostly obtain their food from the local store (n = 32; 100%), from traditional subsistence sources (n = 23; 72%), online orders (n = 20; 63%), and stores in Anchorage or other communities outside of the region (n = 16; 59%) (APIA 2023). The report also showed that 56% of the surveyed St. Paul residents often or sometimes worry that their food will run out before they have money to buy more and that 44% feel no or only a low level of control over getting the foods they want in St. Paul (APIA 2023). A more detailed description of the study area, including its current (crab) fishery challenges, is given in Appendix B.

Fig. 1
figure 1

The five Pribilof Islands: St. Paul Island, St. George Island, Otter Island, Walrus Island, and Sea Lion Rock. The photos at the top left and mid-right of the map were both taken by SZ in St. Paul on St. Paul Island. The overview map at the bottom left shows the location of the five Pribilof Islands and their wider surroundings in the southeastern Bering Sea

Full size image

Research co-design

Responding to complex challenges in Arctic Indigenous contexts calls for inclusive approaches that acknowledge and integrate the rich knowledge and perspectives of the affected communities (Yua et al. 2021). We adopted a participatory approach to ensure our research is rooted in the lived realities of the ACSPI and the Tribal Government of St. Paul Island (TGSPI). Our approach for identifying current food security challenges and food-related interactions between different actor types in the St. Paul Island food system was co-designed by the authors and TGSPI representatives. It was clear from discussions with stakeholders that a diversity of interactions around food typifies the mixed food system on St. Paul Island. Thus, we developed a network-based approach that could capture how interactions between individuals gave rise to the structure of this mixed food system and simulate how food system challenges may impact the diversity of this system and may be mitigated in the future.

The author team has diverse backgrounds and levels of familiarity with the island (Appendix C). The TGSPI was involved in designing the research objectives and methodological approach, identifying interviewees, and validating food security challenges (Fig. 2). Meetings with TGSPI representatives to co-design the methodological approach occurred between February 2021 and January 2022. The insights of this research will be used in a co-created process to design pathways to a diverse and food-secure mixed food system on St. Paul Island (Appendix C).

Fig. 2
figure 2

The workflow from research co-design and data collection to post-processing interview data and data analysis

Full size image

Data collection

Data for the network were collected from March to May 2022 by SZ and KLZ through in-depth interviews with 36 local food system actors, namely ACSPI and TGSPI members who partake in the local food system and either manage food security challenges through their profession and work, have an interest in solving those challenges, and cause or are affected by them. Three of the interviewees were also part of the co-design process. We focused on actors at the local level to shed light on and empower local perspectives.

Upon arrival on St. Paul Island in March 2022, SZ promoted the project through the local radio station, flyers, and posting to relevant Facebook groups. Interested ACSPI and TGSPI members were invited to contact SZ to schedule an interview. Additionally, two co-authors (VP and LD), who have worked with the TGSPI for a combined 18 years, compiled an initial list of local food system actors. SZ established contact with these actors and invited them for individual interviews. The interviews lasted approximately 90 min, and all interviewees were compensated for their time. All interviews were audio-recorded, and photos of the different steps were taken.

A mixed methods approach was used throughout the interviews to generate the network data. Quantitative data on actor interactions was used to develop the network structure, and qualitative data complemented our understanding of actor types and their interaction types. We used an inductive bottom-up approach in which the interviewees put forward what they perceived as food security challenges and important actors in their food system.

We started each interview by informing the interviewees about our aims, namely identifying food security challenges, local food system actors, and food-related interactions between them. Interviewees were then asked to consent to participate in the study (Appendix D). Next, we openly discussed Indigenous food security to ensure SZ, KLZ, and the interviewees shared similar understandings. Interviewees were shown photos of the different food security pillars to make the discussion more accessible (Appendix E). The interviewees were then asked to identify challenges related to local food security, and SZ collected all challenges on post-its. In the next step, the interviewees were asked to name all important actors in the local food system. Again, SZ collected the named actors on post-its. Lastly, interviewees were asked which actors they directly or indirectly interact with in relation to food and what the nature of that interaction was. For actors, interviewees identified unique individuals or organizations by name. For interactions, interviewees described them in any way they liked but were reminded of the Indigenous food security dimensions if they struggled to think of certain interaction types. The questionnaire used during the interviews is given in Appendix F.

Data post-processing

The interview recordings were transcribed and coded by SZ (Fig. 2). We used an inductive bottom-up approach from the interview transcripts to code for food security challenges, actor types, and food-related interaction types (Thomas 2003). This allowed for capturing the mixed St. Paul Island food system as perceived by the interviewees.

The most frequently mentioned food security challenges were presented and validated in a meeting with TGSPI representatives. Based on this, out-migration and knowledge loss were identified as two key food security challenges in the local food system. The St. Paul Island community is currently declining at 4.91% annually, and the population has decreased by 13.93% since the most recent census in 2020 (World Population Review 2023). In 2021, the population further declined by 19 inhabitants (ACSPI 2023). Due to the high costs of food and living and limited education on the island, many residents are moving to mainland Alaska. As a result, active participation in the St. Paul Island food system is declining, and ILTK centered on traditional foods and subsistence activities is being lost. Additionally, many residents lament the decrease in knowledge exchange between remaining community members due to the introduction of new technologies or the repercussions of the COVID-19 pandemic.

Regarding the network data, the first step was to inductively group actors into actor types. SZ allocated actors to actor types according to their primary role as identified through qualitative analysis of the interview data. This led to the identification of five actor types, namely Public institution, Company, Institutional representative, Subsistence provider, and Non-subsistence provider, of which the first two represent organizations and the latter individual community members. Next, interactions were inductively coded based on the interviewees' descriptions, which led to the identification of eight food-related interaction types, namely Food donation, Food purchase, Food sharing, Food shipping, Knowledge exchange, Resource exchange, Subsistence activity, and Working together (Table 1). In the next step, ego networks were developed for each interviewee, in which node categories represented actor types and link categories represented interaction types, respectively. Ego networks are networks that specifically map the connections of and from the perspective of a single person (Gupta et al. 2015), paying respect to the individual perspectives of the interviewees and making them suitable to capture diverse interaction types typical of a mixed economy. The food system network of St. Paul was constructed from these 36 ego networks and comprised 92 unique actors and 584 unique links. The network design allowed up to one link per category between two nodes. No conflicts (e.g., contradicting links) emerged during ego network aggregation. Additionally, links were formalized as directed or undirected to account for the interaction each node contributes to the network. In calculating the SDI, undirected links count for both nodes, whilst outlinks count only for the source node. It is important to note that because this network is based on ego networks, it reflects the reality of the St. Paul Island food system as perceived by the interviewees.

Table 1 Node and link categories in the network of food-related interactions among actors in the St. Paul Island food system. Node categories with an asterisk (*) indicate node types that represent individual community members rather than organizations
Full size table

Shannon-Diversity Index (SDI)

We used the SDI as an indicator of food system diversity. Providing a quantifiable measure of diversity considering richness and relative abundance, the SDI is typically used to assess diversity in ecosystems across different communities (Morris et al. 2014; Santini et al. 2017). In networks, the SDI can indicate the diversity of nodes and links (Klein et al. 2021). It is calculated as the negative sum of the proportion of individuals in a specific class (in our case, food system interactions) multiplied by the logarithm of the proportion. In our analysis, the entire network of the food system is equivalent to an ecosystem, each node is equivalent to a community, and each link type is equivalent to a species class. The number of links is equivalent to the species richness in its classical ecological application.

Data analysis

We began by mapping the number of links between actor types using a chord diagram to gain a general overview of which actor types interact in the food system (Flourish 2022). We then used an alluvial diagram (library ggalluvial) to understand the nature and relative frequency of food-related interaction types between different actor groups (R Core Team 2021). To establish a baseline diversity value for the whole food system and subsystems defined by actor types, we calculated SDI values for the whole network and the different node categories. In this way, we could assess how each actor group contributes to food system diversity respectively (objective (i)). We also calculated summary statistics for all actor types, including degree and betweenness values, which are presented in Appendix I.

Using Monte Carlo methods, we then simulated the effects of the identified food security challenges on the SDI values of the whole network and the actor types representing community members (institutional representative, subsistence provider, non-subsistence provider) to understand their impact on food system diversity (objective (ii)). To simulate the effects of out-migration, nodes representing community members were randomly removed from the network in a stepwise process. Links associated with these nodes were also removed from the network as these represent the interactions of the nodes in the overall food system. Consequently, removing nodes with many different links had a bigger impact on network diversity than removing nodes with few homogenous links. To simulate the effects of knowledge loss, knowledge exchange links between any two nodes in the network were randomly removed in a stepwise process (Appendix G).

Finally, to simulate the effects of different food system interventions that could be increased, we simulated adding links of different types to the network and measuring their average effect on the overall network SDI and the SDI of the three actor types representing community members (objective (iii)). To represent different food system interventions, we used Monte Carlo methods to add Food sharing, knowledge exchange, resource exchange, subsistence activity, and working together links to nodes of relevant categories in the network (Appendix G). We focused on interventions at the community level to explore local agency to meet food security challenges. In all simulations, in each step, one single node or link was randomly removed or added to the network, respectively. Mean and standard deviation in SDI values were calculated over 1000 replicates. Taking a mixed methods approach, we contextualize these quantitative results from our network analysis with the lived realities of the St. Paul Island community through qualitative interpretation and quotes arising from interviews.

Results

General network overview

The network’s connectivity is shown in Fig. 3 (also see Appendix H). Subsistence provider and non-subsistence provider nodes have the highest interconnectivity in the network (81 and 57 outlinks, respectively) (Fig. 3a). They are mostly connected through knowledge exchange and food sharing links (Fig. 3b). Subsistence provider nodes are also strongly linked with other subsistence provider nodes through knowledge exchange and subsistence activity links (Fig. 3b). Notably, public institution and company nodes are not interconnected (Fig. 3a) but are the sole providers of food purchase and food shipping links, respectively (Fig. 3b). Public institution and company nodes have links to all actor types representing community members, especially institutional representative nodes (46 and 28 outlinks, respectively) (Fig. 3a). These are mostly working together and knowledge exchange links (Fig. 3b). Institutional representative nodes also have many working together and knowledge exchange links with other institutional representative nodes, and knowledge exchange links with subsistence provider and non-subsistence provider nodes (Fig. 3b). A table showing the number of links behind each alluvial is given in Appendix H.

Fig. 3
figure 3

Food-related interactions between actor types in the St. Paul Island food system. The chord diagram a shows a general overview of which actor types interact in the food system. The arc lengths indicate the number of outlinks in each node category with the number of nodes given in the arc title. The thickness of the chords represents the number of directed links between node categories with the numerical value of outlinks indicated next to the relevant actor class. The alluvial diagram b shows the nature and relative frequency of the interaction types between different actor groups. The alluvials represent directed links between nodes of the different categories. Colors represent different food link categories or food-related interaction types. A table showing the number of links behind each alluvial is given in Appendix H

Full size image

We identified three key results that offer relevant insights into the St. Paul Island food system: the importance of subsistence providers for the food security of other community members, the lack of interaction between public institutions and companies, and the stance of institutional representatives as important bridge actors between public institutions and the local community.

First, our network indicates subsistence providers predominantly engage in subsistence activities with other subsistence providers, but commonly share food with other members of the community. This reveals the importance of subsistence providers for the food security of community members who are not engaged in subsistence activities and would not have access to culturally important and nutrient-rich subsistence foods if they did not receive these foods via sharing networks.

[Subsistence provider] is really known for sharing a lot. […] He’s really important for my food security. […] Now that I have my kids and things are a little more hectic, he delivers food to me and my sister, who also has a kid.—Non-subsistence provider (Interviewee #4)

Second, our network shows no links between public institutions and company nodes. It is important to note that this is partly due to how the data collection has been carried out and only partially reflects the reality of the St. Paul Island food system (see “Outlook and limitations”). However, interviewees repeatedly voiced concerns about the need for more collaboration between public institutions and companies to tackle food security challenges.

[Company] is important because they hold power over the community. I just wish they were, you know, better players in the community […]. There's a lot of potential for us […], and I hope going into the future with that. You know, relations get better over there. […] I mean, we have gotten things done. […] So yeah, I think they can be a major player in the community going forward.—Institutional representative (Interviewee #16)

Third, the network showed that institutional representative nodes are well-connected to all actor types and have the highest average betweenness centrality measure in the network (Appendix I). Hence, institutional representatives are important bridge actors in the food system. As professionals, they are close to the decision-making processes of public institutions and companies, and as community members, they engage with other community groups.

When COVID started, and we went into lockdown, […] the [Public Institutions] got together, […] and we were looking at resources and ways we could continue to have services here and still be safe and not exposed. […] We tried to figure out how we could help each other out because money sometimes is limited. How could we keep the store warm so they could afford to keep it open […] and still have their employees? […] How could we help each other? I think if there is a time where […] food really is becoming scarce, that's going to have to happen. There's going to have to be the entities working together to find the resources to make things happen, you know, to provide the people.—Institutional representative (Interviewee #14)

Loss of food system diversity through out-migration and knowledge loss

The network of food-related interactions between different actor types in the St. Paul Island food system has a baseline SDI value of 0.78. The SDI values for the different node categories are public institution (0.90), institutional representative (0.89), non-subsistence provider (0.84), subsistence provider (0.76), and company (0.35). For out-migration, mean SDI values for the whole network and the three actor types representing community members show a linear decrease after the stepwise removal of one random node from either of the three actor types from the network (Fig. 4a, b). For knowledge loss, mean SDI values for the whole network and the three actor types show a less pronounced decrease after the stepwise removal of Knowledge exchange links from the network (Fig. 4c, d).

Fig. 4
figure 4

Simulated loss of food system diversity through out-migration and knowledge loss. Development of the network’s mean SDI value after the stepwise removal of nodes and links. For a and b, in each step, one random node from the three actor types representing community members (institutional representative, subsistence provider, and non-subsistence provider) is removed from the network. For a all groups (grey), the SDI for the whole network is measured. For b, the SDI for each community category is measured individually (blue, green, orange). For c and d, each step removes one random knowledge exchange link between any two nodes from the network. For c all groups (grey), the SDI is measured for the whole network. For d, the SDI for each community category is measured individually (blue, green, orange). Points show mean SDI values and shaded areas show standard deviations over 1,000 replicates

Full size image

Our analysis indicates that out-migration has a more immediate impact on food system diversity than knowledge loss (Fig. 4). Losing community members directly reduces the number of interactions and, hence, food system diversity. At the same time, out-migration reduces the number of active participants in the food system. As a result, established patterns of food sharing, knowledge exchange, and other social support mechanisms may be disrupted, increasing vulnerability to food insecurity.

There’s a rule in Alaska where, like, if your school has less than ten kids enrolled, then they close the school. And so, it’s a big deal that people are moving away […] because the fewer kids we have, the closer we get to closing the school. […] And if there’s no school, entire families have to move.—Institutional representative (Interviewee #36)

Knowledge loss, while critical, has a more gradual impact as the existing strong communication network in the St. Paul Island food system initially mitigates its effects. Nevertheless, knowledge loss has detrimental effects on the food system.

The new generation […], they don't eat native foods like they used to. […] But not everybody knows how to hunt either, you know, so it's good to see some younger kids […] trying to learn how to go hunting. So, hopefully, that will start picking up again with the knowledge of hunting and how to prepare. […] Because everybody could shoot a gun, but can you […] process the food?—Subsistence provider (Interviewee #26)

Increasing food system diversity through food system interventions

For the whole network, mean SDI values show the strongest increase when, in each step, one link of each category is added to the network (all types). Regarding individual food system interventions, increasing resource exchange links has the strongest positive effect on mean SDI values. Conversely, adding additional knowledge exchange links slightly negatively affects mean SDI values because there is already a high richness of knowledge exchange links in the network (Fig. 5a). Similar patterns are found at the category level for the three actor types representing community members, except increasing working together links slightly negatively affects the mean SDI values of institutional representative nodes, and increasing subsistence activity links has a less pronounced positive effect on mean SDI values for subsistence provider nodes than for the other two actor types, respectively (Fig. 5b).

Fig. 5
figure 5

Increasing food system diversity through food system interventions. Development of the network’s mean SDI value after the stepwise adding of links. a The change in mean SDI values for the whole network and b for the three actor types representing community members (institutional representative, subsistence provider, non-subsistence provider), respectively. For b, facet headers indicate for which group the SDI is measured. For all types (grey), in each step, one link of each category (food sharing, knowledge exchange, resource exchange, subsistence activity, working together) is added to nodes of relevant categories in the network. For food sharing (purple), knowledge exchange (blue), resource exchange (green), subsistence activity (yellow), and working together (red), one link of the respective link category is added to nodes of relevant categories in the network. Points show mean SDI values and shaded areas show standard deviations over 1000 replicates

Full size image

The many knowledge exchange links in the actor-network of food-related interactions make it more resilient to knowledge loss than out-migration. Hence, the strong communication network on St. Paul Island provides an opportunity for planning and implementing targeted food system interventions. Our analysis suggests that jointly increasing resource and food sharing, working collaborations, and communal participation in subsistence activities is most effective in increasing food system diversity (Fig. 5; Appendix G). As an individual food system intervention, enhancing resource exchange is most effective in increasing food system diversity as this interaction type is currently underrepresented in the network (Morris et al. 2014).

There’s probably a handful of hunters that hunt sea lion and seal and reindeer that have done it for many years. And they have a history of distributing to elders, […] getting it for community events and stuff like that. […] They definitely hold the knowledge of doing that. If anybody wants to teach somebody or learn, I think they’d be good people to go to.—Non-subsistence provider (Interviewee #28)

Discussion

We sought to investigate the diversity of food-related interactions between different actor types in the mixed Arctic Indigenous food system on St. Paul Island, in the context of different food security challenges and interventions. Having identified eight types of food-related interactions related to subsistence- and market-based food flows, we found food-related interactions between different actor types in the St. Paul Island food system to be diverse. The local food system thus provides a typical example of a modern Arctic mixed economic system in which different actors show diverse social interconnections (objective (i)) (Fig. 3). Out-migration and knowledge loss are two key challenges in the St. Paul Island food system. Both challenges decreased food system diversity and, thus, pose a potential threat to the diversity that typifies this mixed Indigenous food system; however, out-migration presents the more urgent threat (objective (ii)) (Fig. 4). Targeted interventions to meet key food security challenges can effectively increase diversity in the St. Paul Island food system. We found the combination of different food system interventions to be most effective in strengthening this diversity (objective (iii)) (Fig. 5). In the following, using a resilience lens, we discuss our results in the context of Arctic Indigenous food security.

Food-related interactions are diverse in the mixed St. Paul Island food system

As mixed economies, Arctic Indigenous food systems are inherently diverse, enabling adaptability and creating space to sustain traditional indigenous practices (Ford et al. 2015). We found that traditional food-sharing networks contribute to the diversity of the St. Paul Island food system and are important not only for food provision but also for maintaining cultural traditions, a sense of community, and a meaningful connection to the environment. Traditional food-sharing networks represent a fundamental aspect of social-ecological resilience in Arctic Indigenous food systems and have been put forward as a critical element in maintaining food security, especially in times of resource scarcity or economic hardship (Collings et al. 2016). Food-sharing networks are adaptive strategies that allow for flexibility and for communities to efficiently redistribute resources in response to changing ecological conditions or the availability of subsistence resources (Scaggs et al. 2021). Hence, engaging in subsistence activities and food sharing is essential for maintaining nutritional health and cultural practices and serves as a cornerstone for food system resilience in the face of change (Ready and Power 2018).

Effectively addressing complex challenges often requires diverse collaborations across scales, institutions, and actor types, as these can facilitate knowledge and resource sharing, leading to more effective and adaptive management (Bodin and Crona 2009; Bodin 2017). In the context of the St. Paul Island food system, it was found that institutional collaboration was perceived to be limited and that increased collaboration among such entities could lead to improved management of the local food system, greater resource efficiency, and enhanced resilience against local challenges. Encouraging such partnerships could be a key strategy in strengthening local food security. In Arctic Indigenous communities, where traditional knowledge and practices are a central part of food security, collaborations between indigenous and non-indigenous actors are particularly important, as culturally appropriate outcomes for these communities can only be achieved where ILTK is integrated into decision-making processes through co-production and co-management (Armitage et al. 2011; Donatuto et al. 2019; Yua et al. 2021). In particular, bridge actors are pivotal in facilitating communication, resource flows, and collaboration across different parts of a network (Bodin and Prell 2011). Bridge actors are individuals whose ties fill a structural whole, providing the only link between two individuals or clusters (Granovetter 1973). The position of institutional representatives in the St. Paul Island food system allows them to understand the perspectives and needs of various actors, making them important in mediating conflict and finding mutually beneficial solutions (Crona and Bodin 2006). In Indigenous contexts, bridge actors can be instrumental in integrating traditional, scientific, and institutional knowledge (Rathwell and Armitage 2016; Reo et al. 2017; Bixler 2021).

Key challenges threaten food security

Removing nodes directly affects a network’s structure and functionality (Bellingeri et al. 2020). On St. Paul Island, out-migration reduces the number of active participants in the food system, weakening local capacities to sustain cash- and subsistence-food flows and social ties that underpin the community’s ability to manage their food sources collectively (Tsuji et al. 2020). While out-migration of working-age individuals, subsistence providers, and elders can result in losing the skills, labor, and knowledge necessary to handle food shipments, maintain grocery stores, or engage in subsistence activities (Huskey et al. 2004), out-migration of women and children profoundly impacts village viability (Martin 2009).

Out-migration also contributes to knowledge loss. While Indigenous communities have long struggled with the suppression and destruction of ILTK due to colonization (Spiegelaar et al. 2019), this trend might further be exacerbated today by out-migration and globalization (Sowa 2016). Such developments can undermine the intergenerational transfer of crucial information about local ecosystems, subsistence strategies, and cultural practices related to food, resulting in less capacity to adapt to ecological changes, such as shifts in species availability due to climate change (Pearce et al. 2015; Wyllie de Echeverria and Thornton 2019). Research has shown that in the long term, knowledge loss can lead to the erosion of adaptive capacity as communities become unable to respond to changes (Fernandez-Llamazares et al. 2015). The loss of ILTK diminishes the role of traditional foods in local diets, forcing communities to rely more heavily on imported foods and ultimately impacting nutritional health and cultural identity (Egeland et al. 2011; Little et al. 2021).

Targeted food system interventions can increase food system diversity

By enhancing different interaction types, communities can build a diverse food system, which has been shown to increase food system resilience and strengthen Indigenous practices (Ready 2018). In the case of the St. Paul Island food system, this could be done through increasing resource and food sharing, working collaborations, and communal participation in subsistence activities. Sharing equipment necessary for subsistence activities enhances food security by allowing individuals who lack the financial means to own such equipment access to traditional food sources (Wenzel 1995; Harder and Wenzel 2012). Moreover, exchanging non-food items, like clothing, fuel, or tools, can indirectly support food security by enabling community members to focus more resources directly on food acquisition activities. For example, to tackle the effects of out-migration, the sharing of hunting and fishing equipment among remaining community members could enable more residents to access traditional foods. Simultaneously, the sharing of traditional food can increase food security, cultural continuity, and, ultimately, community health and well-being (Newell and Doubleday 2020). Increased collaboration in working contexts (e.g., through meetings, joint projects, collaborative lobbying, coalitions, or legislative proposals) might further enhance understanding of local challenges, coordinate efforts to address them, and lead to solutions more in line with local needs (Bodin et al. 2020; Lam et al. 2020). Our analysis further suggests that interventions focusing on incentivizing the return or retention of community members could strengthen local food system resilience, as an increase in community members would also mean the reintroduction of additional knowledge and skillsets and, thus, a diversification of food-related interactions (Rozanova-Smith 2021). Similarly, efforts to revitalize and transmit ILTK to youth, for example, through knowledge exchange among different generations and communities through panels or workshops, could help mitigate knowledge loss and strengthen communal participation (Islam et al. 2017). Such interventions could leverage existing communication networks to enhance knowledge sharing and collaboration across different community groups (Kuhnlein et al. 2006).

Importance of mixed methods to understand social-ecological system complexity

Our network-based analysis of food-related interactions quantitatively demonstrates the importance of diversity for modern mixed Arctic Indigenous food systems. Hence, it complements previous, often value-driven argumentations for preserving these systems (Sheremata 2018; Green et al. 2019) and underscores the critical role of diversity in Indigenous food systems (Kuhnlein et al. 2009). Food system diversity has been shown to strengthen food security in other contexts, such as recently shown in East European food systems (Jehlička et al. 2018, 2020) or smallholder farming systems (Varyvoda and Taren 2022). System diversity has further been shown to be an important determinant of system resilience (Biggs et al. 2015) both in food networks (Tendall et al. 2015) and in other critical SES networks (e.g., see Kharazzi et al. (2020) for trade networks). We emphasize that a focus on diversity opens the door to capturing often neglected intrinsic and relational interactions in food systems that are crucial to moving beyond an instrumental way of understanding and gaining insight into the interactions that form the deeper levels of social-ecological systems (Abson et al. 2017; Meadows 1999).

We employed a mixed methods approach that combined qualitative and participatory approaches with network analysis to capture and analyze the diversity of food-related actor. There is growing acknowledgment of the value of integrating complexity science methods with ILTK systems in transdisciplinary settings (Zimmermann et al. 2023). Such approaches allow for a more holistic understanding of SES while aligning with Indigenous worldviews emphasizing the interconnectedness of all life (Ali et al. 2021). However, while participatory approaches are increasingly applied in the field of sustainability science (Chambers et al. 2021), participatory complexity methods are still sparingly used to achieve a holistic system understanding and create action for sustainability (van Bruggen et al. 2019). Network analysis, in particular, has proven to be a suitable tool to improve the understanding of the structure of various SES (Domenico and Sayama 2019; Bergsten et al. 2019; Bodin et al. 2019). Here, it allowed us to capture and integrate a diversity of perspectives on the Indigenous food system and integrate them in a systematic way to develop a holistic picture of the local food system. By complementing this with qualitative data, we could effectively describe how specific factors impacted the diversity of this system and what that means for the lived experience of the people of St. Paul Island. We, hence, hope this study inspires scholars from different fields to explore the potential of combining complexity methods in transdisciplinary settings to reveal the complex ways in which systems change and how that change is experienced within these systems.

Outlook and limitations

We see large potential for future research that employs co-produced complexity approaches to provide a more comprehensive understanding of (Arctic) Indigenous food systems (Abgar et al. 2009; Berkes and Berkes 2009) and other SES. However, we are aware that the analysis presented in this paper has limitations. Our analysis lacks validation of how the SDI as a diversity measure relates to real-world indicators linked with Indigenous food security pillars (e.g., local policies as indicators of access; health and wellness indicators based on ILTK, and indicators of natural and physical phenomena) (ICC-AK 2015). Refining how Indigenous food security can be measured using diversity as an indicator could be an important direction for future research. Moreover, our network shows no links between public institution and company nodes, as interviewees were asked to focus on their food-related interactions with food system actors rather than on connections between other actor types. Hence, our network reflects local actors’ perspectives and may not entirely reflect the full structure of the St. Paul Island food system. Although such analyses represent simplifications of reality, future studies of complex SES could specifically address this shortcoming by validating individual perspectives in, for example, a workshop setting.

Conclusions

We set out to investigate the diversity of food-related interactions between different actor types in the mixed food system on St. Paul Island in the context of different food security challenges and interventions. We found diverse interactions related to subsistence- and market-based food flows in the St. Paul Island food system. We identified out-migration and knowledge loss as two key challenges, decreasing food system diversity and threatening local food security. Our analysis revealed a strong communications network between different actor types and highlighted the opportunity for future food system interventions to build on this existing sub-network. We found a combination of different food system interventions to be most effective in strengthening diversity in the local food system. The network analysis of food-related interactions presented in this paper complements previous, often value-driven arguments for preserving Arctic Indigenous food systems by quantitatively demonstrating diversity as a strength of these systems to navigate the complex challenges of the twenty-first century. These findings provide valuable insights for future research endeavors and targeted food system interventions, as they underline the importance of diversity in mixed food systems to maintain and enhance Indigenous food security.