1. Introduction

The concept of social-ecological systems (SES) reflects that nature and society are interdependent and constantly coevolve, often in a non-linear fashion (Berkes and Folke 1998). Of particular importance are SES transitions, which entail fundamental transformations of the system’s structure and processes that ultimately change how governance challenges are defined and understood (Olsson et al. 2006). A SES transition can be triggered by endogenous feedbacks between the society and the natural system (Lambin and Meyfroidt 2010). For example, unsustainable resource use might invoke a SES transition with either a continued exhaustion of the resource and a collapse of the resource stock, or a changed behavior towards a more sustainable resource use. The particular outcome has been the focus of a large number of recent studies (e.g. Gutiérrez et al. 2011; Cinner et al. 2012; Hausner et al. 2012; Fleischman et al. 2014), and depends on a number of interacting factors pertaining to the governance system in place, local norms-and rules, properties of the resource system and its users as well as the contextual situation (Dietz et al. 2003; Berkes 2007). However, a SES transition can also be triggered by exogenous drivers that emerge independently of the internal SES dynamics, but which can change the conditions for the interactions between nature and society (Lambin and Meyfroidt 2010). Governmental policies may, for example, impact local traditional governance systems resulting in a less (or more) sustainable utilization of a common-pool resource (Borrini-Feyerabend et al. 2013). Similarly, climate change, globalization, changed demography, ­economic growth, market integration and technological innovation are all factors that might have pervasive impacts on all parts of the SES, and in particular the demand and pressure on local natural resources (Young et al. 2006a).

This paper investigates if and how exogenous drivers could trigger broad scale transitions in subsistence-oriented SESs in the Western Arctic (i.e. Arctic Alaska, Canada and Greenland). By subsistence we mean “harvesting of natural, renewable resources to provide food for one’s own household, for gifts for others, and for exchange outside the market economy” (Poppel and Kruse 2009, 39). We focus on the tundra areas (i.e. the areas north of the treeline) in North America and Greenland. This area is dominated by Inuit, Iñupiat and Yup’ik cultures, and the people who live there are highly dependent on fish and wildlife resources for sustaining local livelihoods and cultural identity (Nuttall 2007; Huntington 2013; Chapin III et al. 2015). The wildlife and fish resources in the Arctic typically consist of migratory species with a fluctuating distribution and abundance; e.g. beluga whale, ice-dependent seals, caribou and migratory fish and birds. The indigenous people have traditionally coped with fluctuating resources through a number of adaptations, including flexible hunting and fishing practices, a diverse diet, customary sharing networks, a mixed household economy, and local sets of norms and rules with respect to harvesting (Wolfe 1984; Berkes and Jolly 2001; Usher et al. 2003; Robards and Alessa 2004; Parlee and Berkes 2008; Magdanz et al. 2011; West and Ross 2012). The accelerating changes associated with global warming combined with large natural variability in Arctic ecosystems, the development of extractive industries, increased dependence on formal employment and cash income; and reliance on imported foods and fuels are all factors that can change the premises for the management of fish and wildlife (West and Ross 2012).We postulate that exogenous drivers under certain conditions could change the feedback between resource users and wildlife and fish resources, fundamentally altering the SES and the associated governance challenges.

Based on common-pool resource theory, we first present a dynamic conceptual model for possible transitions of subsistence-oriented SESs and the major exogenous drivers that might invoke such transitions. Guided by the model, we review the general properties of the traditional resource-use systems in the Western Arctic. We assess the status and the recent development of drivers expected to be important, and finally assess whether transitions have taken place. The synthesis seek to answer the following research questions: Is there evidence of broad scale transitions in Western Arctic subsistence-oriented SESs? How does these SES transitions relate to exogenous drivers? Finally, we reflect on how future changes in exogenous drivers could change these systems.

2. A dynamic conceptual model for SES transitions in subsistence-oriented communities

Many of the conditions causing SES transitions on local scales have been documented through decades of research on common-pool resources, and were recently systematized into a SES framework by Ostrom (2007, 2009) and McGinnis and Ostrom (2014). The common-pool resource literature originally characterized resources by two fundamental features (Ostrom et al. 1994): (i) subtractability, which occurs when the exploitation of a resource by one user reduces the amount of the resource available to others, and (ii) excludability, which refers to how costly it is to control the resource-access of potential users. Variations along these two dimensions have been used to delineate four types of goods: Private goods (high excludability, high subtractability), toll goods (high excludability, low subtractability), common-pool resources (low excludability, high subtractability), and public goods (low excludability, low subtractability). In the following we assume that the resource users harvest renewable resources (e.g. fish and wildlife resources) in which the exclusion of other users is costly (i.e. low excludability), and we will accordingly concentrate on the variation along the “subtractability-axis” which delineates public goods from common-pool resources. The characterizing of renewable resources in relation to subtractability is however not trivial. This is because subtractability might vary along a continuum, often depending on the stock of the resource itself (Young 2007). This is known as “congestability” (Weimer and Vining 2005), implying that if a renewable resource is plentiful, then a relatively small withdrawal allowing the resource to regenerate, will have little impact on the resource situation for other users. Subtractability will according to this mechanism be positively, although not necessarily linearly, related to the proportion of harvest rate to stock size (i.e. the harvest pressure). Thus, any factor causing an increase in harvest pressure could in principle change a public good into a common-pool resource and vice versa.

The variation along this axis is particularly important because the two types of goods are associated with different governance challenges (Hinkel et al. 2015). Common-pool resources are associated with appropriation challenges because individual resource users will have incentives to maximize their own consumption of an exhaustible resource at the expense of others (Ostrom 1990). On the other hand, public goods can suffer from provisioning problems due to free-riders in situations where a costly collective investment is needed to utilize the resource. Hinkel et al. (2015) termed the former an “appropriation action situation” defined as “those in which actors face a collective challenge to avoid the overuse of a collective good”. In contrast they defined the “provisioning action situation” as “those in which users face a collective challenge to create, maintain or improve a collective good”. In the following we will elucidate how these two action situations depend on the characteristics of the resource system, and how exogenous drivers might change a subsistence-oriented SES from a provision action situation to an appropriation action situation (Figure 1).

Figure 1 

A dynamic conceptual model of how exogenous drivers could impact the dynamics of a social-ecological system (SES) which, in this case activates a transition from a provisioning action situation to an appropriation action situation. The pressure from a driver can increase or decrease, and thus either inhibit or activate a transition. The colors of the arrows indicate the relationship between the driver and the response; blue for positive effect and red for negative effect. Under the provisioning action situation the resource users are limited by the fluctuations in a large and uncontrollable renewable resource. A number of exogenous drivers working on the resource might have negative consequences for the state of the resource (left side). Similarly, drivers working on the resource users might enhance the harvest efficiency and the resource demand (right side). Note that the list of exogenous drivers is not exhaustive. Increased harvest pressure could be a result of 1) increased harvest efficiency and increased resource demand, and/or 2) increased harvest efficiency and reduced state of the resource. The result of an increased harvest pressure could be a transition of the SES from a donor controlled system where the collective challenge is to secure the subsistence on a local scale, to a common-pool resource system where the collective challenge is to avoid overuse on the scale of the resource system. Once an appropriation action situation has been introduced several pathways for the SES transition is possible, e.g. resource collapse, self-organizing of resource users to avoid over-harvest or governmental top-down regulations.

Ostrom (2009) argued that large resource systems with low predictability and high resource unit mobility make it less likely for resource users to self-organize to avoid overexploitation. Interestingly, the same resource characteristics make it difficult for resource users to sustain a sufficient flow of resources over time, suggesting that overexploitation in general might not be an issue in the first place. Migrating resources with large temporal variation in availability are particularly challenging, and such resource systems might take the form of what Pimm (1982) termed “donor-controlled systems”, in which the resource users do not control the resources by harvest, but on the contrary adapt to their fluctuating availability. By definition, the resource users have little impact on the resource stock in such systems, and subtractability is accordingly low. Moreover, the major collective challenge for households and communities will be to efficiently harvest, distribute and store the resources rather than self-organize to avoid over-use. To reflect the dynamic character of such resource systems, we will use the term donor-controlled systems referring to cases where the large variability of the resources limits the resource users’ ability to maintain a high harvest rate over time. In the terminology of Hinkel et al. (2015), the SES would correspond to a provisioning action situation in which the resource users face a challenge to maintain and develop the local collective partnership to secure local food resources for harvesting, processing and distributing among community members.

Because the resource users have little influence on the resource stock, donor-controlled systems are resilient to changes induced locally by the subsistence users themselves (i.e. endogenous drivers). However, exogenous drivers affecting harvest efficiency, resource demand and the state of the resource stock could potentially increase the harvest pressure and thereby transform the system into an appropriation action situation (Figure 1). For example, Fleischman et al. (2014) showed how the Atlantic bluefin tuna fishery has changed from low to high subtractability and thereby challenged the SES’s governance institutions. The Atlantic bluefin tuna stocks move over large areas and have therefore traditionally been able to escape overexploitation due to their mobility. New technologies such as GPS, sonar and spotter planes providing information about the spatial location of the tuna has, together with high market prices, increased the access to and demand for the resource. As a result, the harvest pressure has increased, leading to increased subtractability, and a change in the governance challenge from one of provisioning, to an appropriation challenge to avoid over-harvest.

Some exogenous drivers work directly on the resource stock with potential effects on the harvest pressure (Figure 1; left). For example commercial exploitation by non-residents and habitat change by e.g. expanding extractive industries, are drivers that might reduce the state of the resource and thereby increase its vulnerability to local utilization. Similarly, climate change is expected to reduce or shift the abundance of traditional resources, making them more vulnerable to local harvest. Finally, large natural fluctuation in the resource stock might increase the frequency and duration of periods when the resource is vulnerable to local harvest. Equally important and probably more complex, are the exogenous drivers that work on the resource users (Figure 1; right). For example, the adoption of new technology and equipment for hunting and fishing, as well as more efficient means of transportation and storage, will enhance the hunters’ ability to increase the harvest pressure. However, new technology and fuel are costly, and whether an increased efficiency is realized depends on the resource demand as well as a sufficient cash income to buy the equipment and fuel needed. Population increase and opportunities to sell country food are among the drivers that may increase the resource demand, and encourage investments in technologies that enable hunting levels above the subsistence need. On the other hand, urbanization might offer alternative livelihoods, wage labor and increased importation of food, thus reducing the demand for the traditional resources. Finally, the traditional subsistence culture and practices might be influenced by globalization through the diffusion of ideas, meaning and values. As a result, the younger generations’ interest in traditional harvests might weaken, reducing the demand for the traditional fish and wildlife resources.

If the exogenous drivers increase the harvest efficiency, the resource demand and/or the resource scarcity, then the resulting increase in harvest pressure might trigger a transition from a provisioning action situation to an appropriation action situation (Figure 1). The new situation would shift the collective action challenge from one of securing the subsistence in the local group of resource users to one of avoiding over-harvest at the scale of the resource stock. It is important to note that this model (Figure 1) is dynamic, i.e.; depending on whether the pressure from an exogenous driver increase or decrease, the change can either inhibit or activate a transition. Furthermore, if the system is in an appropriation action situation a “reversed” transition to a provisioning action situation could also be envisaged through a reduced resource demand, an increased resource stock and/or a reduced harvest efficiency.

The long-term outcome of a broad scale transition to an appropriation action situation is dependent on the institutions, i.e. the formal rules and informal norms that structure social interactions, which constrain or enable adaptations to the new governance challenges. Albeit not the main focus in this paper, our conceptual model also illustrates that the new governance challenges could result in multiple pathways of SES transitions. At least three possible outcomes could be envisioned: First, if the resource users do not respond to the challenge, unsustainable harvests may collapse the resource system. Second, resource users might organize at the scale of the resource system and adapt to the new collective challenges, for example through cross-scale arrangements. Third, governments might implement top-down regulations to protect the resource from local harvests. All of these outcomes could have profound impacts on the resource system, the resource users and/or their interactions. Thus, the emergence of this new appropriation action situation would trigger a transition of the SES along different pathways ending in different alternate configurations.

3. Data synthesis

In the following data synthesis we seek to answer our research questions by comparative analyses of subsistence systems in the Western Arctic. The cross-national comparison allows us to examine the influence of large scale exogenous drivers of SES transitions to a greater extent than would be feasible through singular case studies (Popay et al. 2006; Young et al. 2006b). Our approach is a narrative synthesis (Popay et al. 2006) using multiple lines of evidence to examine how the observed system dynamics conform to the conceptual model (Figure 1) by a step-by-step analysis. First, in Section 3.1 we investigate whether the properties of the most important fish and wildlife resources in our study system are congruent to the expectations from a donor-controlled system (i.e. large resource systems with migrating and fluctuating resources; Figure 1, top). In Section 3.2, we synthesize the literature on adaptations and governance challenges associated with donor-controlled systems, and ask whether the traditional SESs conform to a provisioning action situation (Figure 1, upper box). In Sections 3.3 and 3.4 we investigate the major exogenous drivers affecting harvest efficiency and resource demand (Figure 1, right). In Section 3.5, we analyze recent trends in subsistence harvest using available data. Finally, in Section 3.6, we investigate cases were appropriation challenges have arisen due to resource scarcity. We ask what drivers have been involved and what management actions have been taken. We conclude by pointing to the major drivers behind recent transitions to appropriation action situations, and discuss future challenges given the current trends in exogenous drivers.

3.1. Fish and wildlife harvest in the Western Arctic

This section introduces subsistence communities in the Western Arctic, and asks whether the properties of the most important fish and wildlife resources are congruent to the expectations from a donor-controlled system (Figure 1, top).

On a pan-Arctic scale, about 550 communities with a total population of less than one million inhabitants live within or close to the Arctic tundra biome; i.e. communities that harvest from the Arctic fish and wildlife resource systems (Figure 2). This study focuses on the Western Arctic, including 244 tundra communities in Alaska, Canada and Greenland. Fishing and hunting has been the principal livelihood for these predominantly small and remote communities that harvest sea mammals from the Arctic Ocean, fish from the rivers, lakes and the sea, and mammals, birds and plants from the Arctic tundra (Huntington 2013). The area is mainly populated by Inuit, Iñupiat and Yup’ik, however, some of the resources are shared with interior communities further south including Dene, Gwich’in and Métis peoples.

Figure 2 

Human settlements within or close to the circumpolar Arctic tundra biome (Walker et al. 2005) including a 100 km buffer zone to the south. Population data are from the most recent official statistics in Alaska, Canada, Greenland, Iceland, Norway and Russia. See Appendix 1 for data sources.

Based on community-based subsistence surveys in Alaska and Canada in the period 1964–2007 and catch statistics from Greenland in the period 1996–2013 (see Appendix 1), we calculated the relative importance of the different resources in terms of biomass (Table 1). According to these data, the harvest has been dominated by relatively few types of resources. In Alaska, the harvest was dominated by pacific salmon, moose and migratory caribou. In Canada, the harvest was dominated by migratory caribou and ringed seal, while harp seal and ringed seal dominated the harvest in Greenland (Table 1).

In donor-controlled systems resource users do not control the resources by harvest, but adapt to their fluctuating availability. There are at least four resource system characteristics that are likely to keep the SES donor-controlled and in a situation where the main governance challenge is to secure resource needs rather than to prevent overharvest: 1) extensive seasonal migration, 2) large inseparable resource stocks, 3) seasonal superabundant concentrations of the resources, and 4) low predictability in resource availability. By reviewing management reports and the scientific literature, we classified the most important Arctic wildlife resources according to these four characteristics (Appendix 2, Table 1). For example, several of the most important marine wildlife resources have strong affinity to sea ice i.e.; the “ice seals” (harp seal Pagophilus groenlandicus, ringed seal Pusa hispida, spotted seal Poca largha and bearded seal Erignathus barbatus), narwhal Monodon monoceros, bowhead whale Balaena mysticetus and polar bear Ursus maritimus. These species follow the seasonal advance and retreat of the sea ice, making extensive seasonal migrations during their annual cycle. The stocks are large both in terms of number and habitat area, and one stock is normally shared by several distant communities (see e.g. Allen and Angliss 2013). In general, we found that the main food resources in the Western Arctic largely exhibit donor-controlled characteristics (Appendix 2, Table 1). The resources that deviated from this pattern were all terrestrial herbivores including stationary populations of caribou, muskox (Ovibos moschatus) and moose (Alces alces).

3.2. Provisioning actions

This section investigates whether the traditional subsistence SES conforms to a provisioning action situation (Figure 1, upper box).

Our analysis on resource characteristics (Section 3.1), suggested that large, migrating and fluctuating resources dominate the subsistence harvests in the Arctic SES, making donor-controlled systems more likely. First, these characteristics make it difficult and costly to control resource-access of potential users, but they also make it difficult or even irrelevant to engineer the resource system to enhance productivity by means of traditional management measures such as selective ­harvesting to optimize the sex and age-ratio of the stock, habitat alternation to maximize productivity, or culling of competitors such as wolves. Instead, provisioning actions in the Arctic typically depends on a number of strategies that mitigate the seasonal, annual, decadal and possibly centennial fluctuations in the availability of the resources (Robards and Alessa 2004; Tejsner 2013). Flexibility with respect to hunting location, hunting technique, seasonal migration and the targeted species would be important adaptation to a variable resource situation. Subsistence surveys in Alaska show that the resource users in a community might utilize as many as 69 different animal species (Magdanz et al. 2002) and the substitution of a declining resource with alternative hunting targets has been suggested to be an important strategy to cope with changed resource availability (Robards and Alessa 2004; Brinkman et al. 2007; Wenzel 2009). However, Hansen et al. (2013) found little evidence for switching between major subsistence resources in a study of 19 communities in Alaska. They suggested that resource switching might be precluded by the recent transition of the communities into a more market-based economy in which the resource users supplement declining subsistence resources with store-bought foods instead of switching their hunting target (see also Loring and Gerlach 2009).

Many aspects of the hunter-wildlife relationships embedded in Inuit culture mirror the characteristics of a donor-controlled resource system. For example, traditionally Inuit do not view wildlife as passive subtractable resources, but perceive the animals as active, sentient partners in the harvest (Schmidt and Dowsley 2010). The traditions and knowledge emphasize proper relationships with animals through respectful hunting behavior and distribution of hunt products within the community (Thorpe 2004). According to the beliefs, these habits will persuade animals to return to be harvested in the future, ensuring the flow of food and other hunting products to the community (Fienup-Riordan 1990; Stairs and Wenzel 1992; Kendrick 2000). Thus, the provisioning actions go beyond the hunters decisions of where, how and what to hunt and fish, to include the accumulation and transmission of traditional ecological knowledge on a wide array of targeted wildlife and fish species (Berkes and Jolly 2001; Thorpe 2004; Ford et al. 2006; Berkes 2012), and one governance challenge is, for example, to maintain and transfer the knowledge of subsistence hunting and fishing to the younger generations in the community (Pearce et al. 2015).

Provisioning actions are also enabled by the organization of households into networks engaged in cooperative subsistence activities and sharing of hunting equipment and country food (Magdanz et al. 2002; Collings 2011; Harder and Wenzel 2012). This characteristic and well-documented organization of collective provisioning actions among Inuit in the Arctic is the basis for the mixed subsistence-cash economy (Wolfe 1984; Langdon 1991; Marquardt and Caulfield 1996; Usher et al. 2003), and is also important in terms of buffering subsistence disparities between lower- and higher-income households (Dombrowski et al. 2013; BurnSilver et al. 2016). Fuel, supplies and hunting equipment are costly, and access to such resources either by cash income or through sharing networks is a collective challenge of the resource users (Pearce et al. 2015). Subsistence harvest is accordingly combined with wage labor and transfer payment so that cash income is invested in e.g. store-bought food in times of hardship or hunting equipment in times of prosperity (Tyrrell 2009). Recent studies show that sharing networks remain persistent despite higher engagement in cash economy at the household and community levels (Dombrowski et al. 2013; BurnSilver et al. 2016). Similar analyses of sharing networks have not been undertaken in Greenland, but case studies suggest that networks and food markets depend on whether you are a professional commercial hunter or a non-professional hunter that mainly distribute fish-and wildlife resources locally (e.g. Sejersen 2001; Tejsner 2014; Ford et al. 2016).

3.3. Harvest efficiency and local resource demand

This section investigates exogenous drivers affecting harvest efficiency and resource demand (Figure 1, right).

Because the yield in a donor-controlled system is constrained by demanding resource characteristics, an improved harvest efficiency might be a prerequisite to increase the harvest pressure, and eventually transform the system from a provisioning to an appropriation action situation. Harvest technologies have changed considerably since the rifles and steel traps were first introduced in the 19th century, and since then, outboard motors, modern boats, manufactured fishing nets, snowmobiles, freezers, ATVs, two-way radio communications and GPS have all become common technologies used in the subsistence activities (see e.g. Thorpe 2004; Wenzel 2009; Pearce et al. 2015). The increased hunting and fishing efficiency has been assumed by some scholars to disrupt the balance between the subsistence users and the fish and game populations, possibly leading to excessive harvest and declining resources (Hansen and Worrall 2002; Gunn et al. 2011). However, evidence that the use of new technology among subsistence users in the Arctic has resulted in increased or unsustainable harvest pressure is equivocal (Collings 1997). High costs of fuel, hunting equipment, and poor markets might prevent an increased harvest pressure although improved technologies are available (e.g. Fazzino and Loring 2009; Kofinas et al. 2010). For example, Brinkman et al. (2014) found that more than 80% of 178 subsistence users in Alaska reduced the number of trips and the distance travelled, most likely because of higher fuel prices. In other words, while motorized transport potentially could increase harvest efficiency, the resource users depend on cash income to fully access the resources (Wenzel 2009). The governmental programs for settling Inuit in larger sedentary villages in the 20th century often increased the travel distances to the traditional hunting grounds. However, the introduction of snowmobiles allowed people to get further away from the community, and back to their original hunting grounds (Wenzel 2009). Thus, in order to understand how increased harvest efficiency potentially could result in an appropriation action situation, harvest technology and transport need to be coupled to other trends, such as participation in labour markets, the cost of imported goods, changes in demography and settlement pattern.

The Western Arctic is sparsely populated, however in our study area the population has increased several fold during the 20th century (Figure 3, see Appendix 1 for data sources). At the beginning of the century the Arctic part of Alaska, NWT/Nunavut and Greenland each held approximately 10,000 inhabitants. This number had increased to 34,000 for Arctic Alaska, 77,000 for NWT/Nunavut, and 57,000 for Greenland by 2012 (Figure 3). Assuming a proportional increase in the consumption of resources, one could accordingly expect a 3 to 8-fold increase in the harvest of fish and wildlife in the same period. There are however several factors that might have prevented such an increase. First, the shift from the use of dog-teams to the use of snowmobiles and ATVs for transportation reduced the need for dog feed and thereby the harvest of typical “dog feed species” of fish and seals (Collings 1997). Second, according to the analyzes of Rasmussen (2011), most of the population growth in the Arctic since the 1960s has occurred in urban centers, often linked to industrial activities, social services and public administration (see also Magdanz et al. 2002). This development is accompanied by increased wage labor and importation of goods, while the importance of traditional harvesting for subsistence is, according to Rasmussen (2011) declining. The increase in consumption of imported food on the expense of local food has been referred to as a nutritional transition, and is particularly evident among the younger generations and in centers with good access to store-bought food (Council of Canadian Academies 2014). Thus, while an increase in harvest pressure could be expected due to a combination of increased harvest efficiency, increased access to cash income to buy equipment and fuel and increased population size, a stable or decreasing harvest pressure is also possible due to urbanization combined with increased wage labour and consumption of imported food.

Figure 3 

Population development in Greenland, Northwest Territories, Nunavut and Arctic Alaska from the beginning of the 20th century. Arctic Alaska includes Nome Census Area, North Slope Borough, Northwest Arctic Borough and Wade Hampton Census Area. Note that the southern part of Northwest Territories is not within the Arctic definition of Figure 2. Data from Nunavut, which is entirely within the Arctic definition, was available from 1986 to 2012. See Appendix 1 for data sources.

3.4. Commercialization of resources

This section addresses how commercialization affects resource demand (Figure 1, right).

Commercialization is an important factor that could both increase the resource demand and provide cash income to buy fuel and expensive hunting equipment. User-rights are often linked to a particular utilization of the resource, and in our study system the legislative authorities’ prioritization of commercial vs. subsistence use has been the source of a long-standing debate (Wenzel 1991; Marquardt and Caulfield 1996; McGee 2010). Commercial use of wildlife and fish resources might provide important monetary income for rural communities (Armitage 2005; Dowsley 2010), and introducing the resource to a larger market might increase the immediate economic value of the resource and thereby the incentives for hunting and fishing. Interestingly, the commercialization of country food differs considerable among the three study regions. Greenland has a long history of selling fish and wildlife resources on local as well as international markets, while on the other side, the Alaskan legislature put strict restrictions on the commercial utilization of fish and wildlife resources among Alaskan subsistence users. In between we find northern Canada where local food markets have developed at a smaller scale than in Greenland.

In Greenland all residents with a hunting permit (non-professional or professional) are eligible to harvest. However, professional hunters (hunters with more than 50% of the income from hunting and fishing) are prioritized by larger quotas, longer hunting seasons and larger bag limits compared to non-professionals (Marquardt and Caulfield 1996; Sejersen 2003). Harvest not used for one’s own consumption or sharing, is sold unprocessed on the local market, privately to institutions or to other households, or for a fixed price to governmental owned companies (i.e.; Royal Greenland A/S) (Caulfield 1993, 1997). Greenland stands in stark contrast to Alaska, where federal and state laws prioritize subsistence use over commercial use and sport hunting (Behnke 1996; McGee 2010). According to the principle of subsistence priority, the subsistence needs must be fulfilled before commercial or other types of uses are allowed. Marine mammals are protected by the 1972 Marine Mammal Protection Act, and commercial use of marine mammals is not allowed. An exemption was made for Alaska Natives living in coastal communities to allow them to hunt for subsistence and make handicrafts provided that the hunts were not conducted in a wasteful manner. Commercial offshore fisheries are not authorized outside Arctic Alaska (i.e. the Chukchi and Beaufort Seas). Inshore commercial fisheries are however allowed (e.g. salmon, king crab and herring), but are often limited by poor infrastructure and low market interest e.g. (Bavilla et al. 2010; Estensen et al. 2012). Subsistence priority is also adopted as a principle by the Land Claim Agreements in the Canadian Arctic, but selling wildlife and fish products to other subsistence users is allowed. The regulations on how Inuit harvesters could sell local food varies according to the specific Land Claim Agreement (Gombay 2005). In Nunavik, commercial tag and quality standards are needed to sell local foods. In Northwest Territories and Nunavut sponsored programs have supported commercial sale of muskoxen and caribou products, but much of the local food is exported to the south (Council of Canadian Academies 2014). The commercial use of marine mammals is not prohibited, and the commoditization of the traditional harvest of polar bear, narwhal and pinnipeds has caused heated conflicts with respect to management and protection (Wenzel 1991; Armitage 2005; Dowsley 2010). The local food market is still small in northern Canada as harvest is mostly sold locally to the Hunter and Trapping Organization, local restaurants or processing plants, but a recent study by Ford et al. (2016) also reports concerns among resource users of the emerging markets for local food on Facebook.

3.5. Trends in local resource use

This section investigates trends in resource use as a response to changes in resource demand (Figure 1, right).

We used the subsistence data from Canada and North America (Appendix 1) to assess the recent temporal trends in harvest and the relationship to community size (number of inhabitants). Descriptive statistics of the dataset is given in Appendix 1: Table A1. The subsistence surveys have been conducted in different communities in different years, and there are no extensive time series from one single community. To address the temporal trends, we therefore fitted the data to a statistical model controlling for community size and community within region. Specifically, the log-transformed harvest values were fitted to a generalized additive mixed model (GAMM) using the nlme library (Pinheiro et al. 2009) and the mgcv library (Wood 2006) in R, ver. 3.2.0 (R Development Core Team 2015). Year and the log-transformed population size of the community were used as fixed factors and community nested within regions were included as random factors. To account for possible non-linear relationships, harvest was modeled with smooth functions using a thin plate regression spline as basis.

The GAMM model (Figure 4) revealed that the total amount of fish and wildlife harvested per year and person (log-transformed) decreased significantly with year (edf=1.0, P<0.0001) and community size (log-transformed average number of inhabitants; edf=2.7, P<0.0001). For a community size of 386 inhabitants (the median community size), the model indicated that harvest rate per person decreased from 362 kg per person (95% C.I. 285–459) in 1970 to 172 kg per person (95% C.I. 138–215) in 2007 (Figure 4A). Keeping year constant and equal to 2007, the predicted harvest per person was 192 kg (95% C.I. 147–251) in a community with 100 inhabitants, while this figure decreased to 42 kg (95% C.I. 26–69) for a community with 4000 inhabitants (Figure 4).

Figure 4 

Harvest of wildlife and fish resources with respect to (A) year and (B) community size in Canada and Alaska combined. The figures show the predicted values (±standard error) from a Generalized Additive Mixed Model with total harvest rate as a dependent variable and year and log community size as independent variables. In (A) community size is held constant and equal to 386 (the median community size). In (B) year is held constant and equal to 2007 (the last year with data). Random components were community (st.dev.=0.20) within region (st.dev.=0.09). Residual st.dev.=0.13. R2(adj) of the model was 0.44, N=326 (community, years).

Figure 5 

Trends in the harvest of wildlife and fish resources for large and small communities separately. Median community size (386 inhabitants) was used as cut-off. Model formulation was the same as for Figure 4. Figures show the predicted values (± standard error) for the median large and median small community (789 and 154 inhabitants). Effect of year was significant in both models (P<0.001).

To investigate whether the trend differed between large and small communities (i.e.; an interaction term between community size and year), we performed the model for communities larger and less than 386 inhabitants (median community size) separately. In both models, the total amount of fish and wildlife harvested per year and person (log-transformed) decreased significantly with year (#inhabitants<386: edf=1.0, P=0.0002 and #inhabitants>386: edf=1.7, P<0.0001). The predicted trends for the median community sizes is shown in Figure 5. There was a tendency for a slightly more rapid decrease in harvest in large communities compared to small communities, however this difference was small, and due to low sample size it was possibly masked by a large uncertainty for the large communities in the early years (Figure 6).

Figure 6 

Harvests of wildlife in Greenland; (A) seabirds, (B) cetaceans, (C) pinnipeds and (D) reindeer and muskox. * indicates national red-listed species.

The catch statistics from Greenland show different trends for different resources (Figure 6). Declining populations of seabirds were followed by stricter regulations in the beginning of the 2000s, and harvests of these species have decreased dramatically since 1996. The same situation is found for beluga whale, walrus (Odobenus rosmarus) and narwhal (Monodon monoceros). The populations of harp seals and ringed seals (Pusa hispida) are considered to be in good conditions (Appendix 2) and harvests of these species have decreased only slightly. Increasing populations of muskox and reindeer were followed by relaxed hunting regulations and increased catches (e.g. Cuyler 2007). The number of professional hunters and commercial use has decreased (Figure 7); however, the number of non-professional hunters increased temporally, following changes in opportunities for reindeer hunting in the late 1990s and early 2000s (Rasmussen 2005).

Figure 7 

Hunting permits (professional and non-professional) issued in Greenland.

In sum, the harvest and subsistence survey data suggest that the use of local wildlife and fish resources in Arctic Alaska, Canada and Greenland is decreasing, supporting the “nutritional transition hypothesis” (see Section 3.3). The subsistence data from Alaska and Canada also suggest that the harvest per person decreased for increasing community size, indicating that the ongoing urbanization might be important. However, the trend is also evident in small communities (Figure 5), suggesting that this is a general negative trend that is likely to be exacerbated by the urbanization trend. It is important to note that the data from Alaska and Canada represent the biomass harvested per person, thus given the increase in population size (Figure 4), the trend in harvest pressure will be closer to constant. On the contrary, in Greenland the data represent the total harvest, and the trends suggest that the harvest of vulnerable (i.e. red-listed) species has decreased substantially, while the harvest of abundant species has been constant or variable (Figure 6).

3.6. Appropriation challenges

This section investigates how exogenous drivers impact resource status (Figure 1, left), the importance of local harvest in causing resource scarcity (Figure 1, harvest pressure), and the emergence of appropriation challenges (Figure 1, lower box).

To investigate cases were appropriation challenges have emerged, we reviewed management reports and the scientific literature to assess the current status and trend of important fish and wildlife resources, the most important drivers of change and possible management responses (Appendix 2). Specifically, we searched for: i) the status of the resource stocks; ii) the current trend in the resource stocks; iii) the major drivers behind current stock dynamics; and iv) implementations of management regulations of local harvest. Based on the reports, we divided the drivers into five major groups: i) local harvest; ii) rebounding from historical industrial/commercial exploitation; iii) climate warming; iv) natural population cycles; and v) unknown. A detailed description of each resource is provided in Appendix 2. A summary of the results is given below, and a resource specific overview is given in Table 2.

Climate change, historical over-exploitation and natural fluctuations were common explanations for the observed dynamics of resources in the literature, and we found several examples were these drivers were the ultimate reason for implementing harvest restrictions on local hunters and fishers (Appendix 2). Commercial sealing, whaling and hunting during the 19th and 20th centuries decimated many important Arctic wildlife resources, and several slow-growing species are still at low levels. Although many resources have recovered or are recovering, low population size due to previous overexploitation is, together with climate change, a major rationale for current restrictions on local harvest of polar bear, beluga whale, bowhead whale, walrus and narwhal in the three Arctic regions (Appendix 2, Table 2). However, climate warming has become an increasingly important driver in Arctic ecosystems, and in particular the reduction in sea ice is currently severely threatening endemic Arctic species associated with the ice habitat (Kovacs et al. 2011; Michel 2013). This is the major argument behind the recent ESA listing of polar bear and ringed seal (FR 2008, 2012), and it is possible that the effect of climate warming will be an increasingly important argument for the implementation of regulations of the local harvest of Arctic marine mammals. Caribou is the most important terrestrial wildlife resource in the Western Arctic and many of the caribou populations undergo more or less cyclic changes with a period of 60 to 90 years (Ferguson et al. 1998; Zalatan et al. 2006; Gunn et al. 2011). The impact of local harvest when the populations are cyclic low has been heavily debated, especially in the Canadian management system, and recently the Nunavut Government banned hunting of the Baffin Island caribou due to historically low numbers (Nunavut Government 2014).

In all the above mentioned examples the harvest regulations were implemented by governments, and the rationale has been that local harvest represented an additional impact on an already threatened population, indicating that an appropriation action situation has arisen due to exogenous drivers working on the resource (see Figure 1). However, such arguments are often heavily contested by local harvest organizations who frequently question the management authorities’ assessment of the stock and/or the impact from local harvest (Collings 1997; Armitage 2005; Clark et al. 2008; Dowsley and Wenzel 2008; Kendrick 2013). In other words, while governments claim that low resource levels has resulted in an appropriation action situation necessitating harvest regulations, local people often claim that the situation is still provisioning. Thus the conflict is rooted in two different action situations, addressing different governance challenges, scales and levels of organization and knowledge systems. For example, in the caribou hunting systems in Canada and Alaska, hunters use Traditional Ecological Knowledge (TEK) to guide effective, non-wasteful and respectful local hunting practices (Kendrick 2000; Parlee et al. 2005; Kendrick and Manseau 2008). However, to address collective action problems in the case of appropriation challenges, there might be a need for delineating the caribou herds, identify the respective resource users, and implement a monitoring regime that can give unbiased and trustworthy estimates of the harvests and stock size. Currently, this knowledge is mainly collected by management authorities using western scientific methodology, and it is accordingly a considerable challenge to combine TEK with western science in the herd specific management boards that have been established to specifically tackle collective action problems (Kofinas 1998; Kruse et al. 1998; Kendrick 2000; Urquhart 2012). Despite these challenges, the implementation of TEK or IQ (Inuit Qaujimajatuqangit) in the Canadian co-management system has empowered the local resource users and has been instrumental for turning the management focus from a one-sided focus on appropriation challenges towards local provisioning challenges (see e.g. Thorpe 2004; BQCMB 2014).

The impact from and regulations of local harvest followed a regional pattern with less impact and regulations in Alaska, more incidences of impact and regulations in Canada and relatively widespread recent impact and detailed regulations in Greenland (Appendix 2, Table 2). In other words, the effects of local harvest followed the pattern in commoditization of the resources (see Section 3.4), suggesting that commercialization might spur an increased demand and eventually an increased harvest pressure. In Arctic Alaska, where commercialization of fish and wildlife is largely suppressed by national and state legislation and poor infrastructure (Section 3.4), we found relatively few local regulations of subsistence hunting (Table 2, see Appendix 2 for details). In the cases where regulations have been issued resource scarcity was, according to the management reports, mainly due to exogenous drivers such as natural fluctuations, historical overharvest from non-locals or climate change. We found no cases where local harvesting was mentioned as a major driver of population dynamics, and subsistence harvest was neither considered to be a threat to any of the three Arctic species listed under the Endangered Species Act (ESA).

In Arctic Canada, restrictions on local hunting have been implemented to protect declining or vulnerable populations of bowhead and beluga whales, narwhal, polar bear, caribou and muskoxen. The dominant explanations for decreasing resources have been natural fluctuations, historical overharvest from non-locals or climate change. However, local harvest of migratory caribou, beluga whale, narwhal and polar bear have according to the management reports contributed to resource decline in several cases (Appendix 2, Table 2). In general, there has been an increased need for cash in the mixed household subsistence economy, and in Canada were selling resources have been enabled to some extent (see Section 3.4), there are several examples in which commoditization of wildlife products has spurred conflicts among management authorities and local resource users (Wenzel 1991; Armitage 2005; Clark et al. 2008; Dowsley and Wenzel 2008).

In contrast to Canada and Alaska, Greenland has long traditions of commercial utilization of natural resources (Section 3.4), and local over-harvest and dwindling resources were the main causes behind the enforcement of new and stricter hunting regulations on walrus, beluga whale, narwhal, polar bear and several bird species in the early 2000s (Appendix 2, Table 2). For example, of the 115 species/populations considered in the Greenland red-listing process, 36 were listed within one of the six IUCN criteria and hunting was considered to be a serious threat to half (18) of the listed species/populations (Boertmann 2007). Subsequent to the implementation of stricter hunting regulations, the pressures on the threatened species have decreased considerably (cf. Figure 6). The situation for some seabird populations has improved following the reduction in harvest pressures (Merkel 2010); however, populations of thick-billed murre (Uria lomvia) have continued to decrease (Merkel and Labansen 2013). Thus, despite being largely top-down oriented with little legitimacy locally (Nielsen and Meilby 2013), the hunting statistics suggest that the management actions have been successful in terms of reducing harvest pressures on the threatened populations. This has, however come at a cultural and democratic cost (Nielsen and Meilby 2013).

4. Conclusion

Our synthesis of the characteristics of the fish and wildlife resources and the corresponding provisioning challenges in the Western Arctic, suggests that the resource users traditionally do not control the resource level but rather adapt to the fluctuating availability of fish and wildlife. The collective challenge facing the resource users is therefore to maintain and secure the subsistence in the community rather than to prevent overexploitation. The traditional subsistence SESs are accordingly mainly donor controlled and conforming to a provisioning action situation. Given the potential increase in harvest efficiency during the last century by the introduction of new technology, a transition from a provisioning to an appropriation action situation might arise due to exogenous drivers working to 1) increase the cash income to buy necessary equipment and fuel, 2) increase the resource demand, and 3) reduce the standing stock of the resource. Our synthesis show complex interactions among multiple exogenous drivers which have either activated or inhibited transitions. Thus, the results do not indicate a broad scale transition throughout the Western Arctic. However, the data indicate that appropriation challenges have emerged to a varying degree, and mainly through two lines of development. First, exogenous drivers such as climate change and previous industrial overharvest have resulted in several cases of resource scarcity. Although the arguments often have been contested by the resource users themselves, management authorities have as a consequence implemented harvest restrictions on local subsistence users. Secondly, commercialization of fish and wildlife resources have, most notably in Greenland, provided cash income and incentives to increase the harvest. In Greenland and in a few cases in Canada, this has resulted in excess harvest and appropriation action challenges. As a response, the government in Greenland has implemented harvest restrictions which presumably have reduced the pressure on vulnerable wildlife species. A third line of development has worked in the opposite direction to decrease the demand for fish and wildlife. A combination of urbanization, increased wage labor and importation of food has reduced the demand for local country food throughout the study area. Thus despite increasing populations, the resource use has not increased, ­possibly preventing appropriation challenges. Presently, Arctic warming is threatening many important Arctic wildlife resources, and this exogenous driver is likely to result in more cases of resource scarcity and associated governance interventions in the name of preventing additional impacts from local hunters and fishers. The transition to an appropriation action situation involves a change in the objectives and scale for collective actions. The governance challenges may not be perceived equally among all participants and stakeholders, and conflicts have been a frequent outcome in the arenas where these re-framings are negotiated (Kofinas et al. 2013). Such changes are particularly demanding in the Arctic where the fish and wildlife resources typically are large, fluctuating and migrating and where the collective actions traditionally have been directed towards solving local provisioning challenges.