Research

CONTRIBUTION OF STEERING COMMITTEE AND KNOWLEDGE HOLDERS

The present study is based on the knowledge shared by Cree land users who participated in interviews and the guidance of a steering committee, comprising representatives from community, regional, and provincial organizations, governments, and co-management agencies.

INTRODUCTION

Participatory modeling of social-ecological systems has gained increasing attention in recent years as a method for communicating local and Indigenous knowledge of complex and dynamic system change. Ecological research often excludes Indigenous Peoples and their knowledge from studies of nature, generating findings that are, at best, irrelevant and, at worst, harmful to Indigenous self-determination and livelihoods (Smith 1999). Participatory modeling involves the co-production of knowledge by engaging stakeholders, rightsholders, and communities in the creation of models that represent their knowledge, experiences, and values within a system (Voinov et al. 2018). Knowledge co-production is widely recognized as improving the quality, legitimacy, and trust of research and as enhancing decision-making processes by the inclusion of diverse and underrepresented perspectives (Norström et al. 2020, Zurba et al. 2022). As Indigenous voices and allies call for reconciliation through more inclusive research approaches within and among nations (Kovach 2009, McGregor 2018, Wong et al. 2020) and around the globe (Smith 1999, Bussey et al. 2016, Mataira 2019), collaboration with local Indigenous experts has become a priority for natural resource research and sustainability science (ICE 2018, ITK 2018, FSC-IF 2020, GC 2022). Indigenous Knowledge reflects long-term and place-based experiential knowledge of the natural world and human relationships with nature, meaning that Indigenous Knowledge is fundamentally connected with Indigenous Peoples’ lived experience and ways of living (McGregor 2004, Kimmerer 2013). The diversity of Indigenous knowledge systems can be paired with Western scientific approaches to provide a rich understanding of the functioning of social-ecological systems (Berkes and Berkes 2009, Reid et al. 2021, Stern and Humphries 2022).

Fuzzy cognitive maps (FCMs) are graphical models of knowledge systems that aim to capture individual or group understanding of possible causes of a particular issue (Gray et al. 2015). FCMs reflect how human cognition functions by representing complex causality with multivariate interactions, loops, and feedbacks, and by expressing causal strength with linguistic statements (Kosko 1986). FCM is a participatory mapping approach that can be used to identify social-ecological knowledge gaps and uncertainties and to gain a more comprehensive understanding of a complex system by incorporating multiple perspectives and diverse knowledge types (Gray et al. 2015). This method bridges the gap between qualitative story lines and quantitative analysis (Jetter and Kok 2014) and hence provides an opportunity for local knowledge systems to inform environmental decision making. Cognitive maps are also an excellent communication tool that can promote mutual learning, facilitate constructive debate, and build trust among researchers, partners, and policy makers (Andersson and Silver 2019).

Circumboreal forests account for about one-third of all remaining global forests, acting as important reservoirs for global biodiversity and providing essential ecosystem services and resources for many human populations (Bradshaw et al. 2009, Burton et al. 2010). Large-scale forestry has become one of the most significant drivers of boreal disturbance (Kuuluvainen and Gauthier 2018) with impacts on the ecosystems and wildlife species on which local populations rely (Schipper et al. 2008, Lindenmayer 2009, Kellner et al. 2019). In Canada, many Indigenous communities attribute recent declines in moose populations across boreal regions to forestry practices and are insisting that their observations, knowledge, values, and life-ways contribute more meaningfully to forestry and moose co-management (Popp et al. 2019).

Our study focused on Cree understanding of the complex interrelationships between forestry practices and moose habitat quality in the boreal regions of Eeyou Istchee in Northern Québec (Canada). For the Cree of Eeyou Istchee, the moose (ᒨᔅ muus [Cree]; Alces alces, also referred to as Alces americanus [Rosenblatt et al. 2023]) is a culturally significant species and, since large-scale clearcutting began in the 1970s, a focal point of conflict with government and industry over forestry and wildlife management decisions (Feit 1973, Potvin et al. 1999, Jacqmain et al. 2008,). A co-management regime, intended to support forestry in the region while also maintaining or improving the habitat of moose and other important wildlife species, was implemented almost 20 years ago. However, the Cree of Eeyou Istchee have ongoing concerns about the moose population, and, in 2021, an aerial survey confirmed a moose population decline in the portion of the territory covered by this co-management regime (Hunting Zone 17, see MFFP 2022).

In this study, we use a knowledge co-production approach (Norström et al. 2020, Zurba et al. 2022) to develop fuzzy cognitive maps (FCM) of a social-ecological system as understood by local Indigenous expertise. Specifically, our aim was to develop FCMs exploring the various ecological and social interrelationships that influence moose habitat quality in Eeyou Istchee almost 20 years after the implementation of a forestry co-management regime. Our main objectives were to use a knowledge co-production and participatory mapping approach to document Cree social-ecological knowledge on moose habitat to inform the development of a co-produced habitat quality index and better communicate Cree knowledge to decision makers at the local, regional, and provincial levels and identify priorities for resource management. Our challenge was to understand how the diverse knowledge of multiple Cree experts, each familiar with a portion of the territory, could be understood and synthesized in a manner that allowed us to identify key habitat factors and their overall influence on moose habitat quality. Fuzzy cognitive mapping allowed us to compare knowledge from different knowledge holders and uncover similarities and differences, while combining maps into a collective knowledge account informed by a multiplicity of understandings and experiences.

RESEARCH APPROACH AND METHODS

About Eeyou Istchee

The Cree Nation of Eeyou Istchee (ᐄᔨᔫ ᐊᔅᒌ, “The People’s Land”) covers a large territory of over 400,000 km² east of James Bay and south-east of Hudson Bay in Northern Québec, Canada. Despite constitutional rights over the land, enshrined in 1975 in the James Bay and Northern Québec Agreement (JBNQA) and Section 35 of the Constitution Act, the Cree have long struggled to assert their influence over resource management policies (Tanner 2018, Cyr et al. 2022). In 2002, the Cree and the Government of Québec established a co-management agreement that sought a more collaborative approach to forest and other resource management in Eeyou Istchee (Governments of Québec and the Cree Nation 2002). Under this agreement, the Adapted Forestry Regime (AFR) was implemented to allow for adaptations that better take into account the traditional Cree Way of Life (Eeyou Pimatseewin), including requirements for Cree participation in forestry consultations and the development of measures to better protect wildlife habitats. Other adaptations included switching to a trapline-based management system that recognized traditional family hunting grounds, or traplines, typically occupied by multiple related families and managed by a land steward, or tallyman (Tanner 2018).

Under the AFR, larger clearcuts (i.e., 250–500 ha blocks separated by 60 m residual forest strips) were phased out across the territory in favor of variable retention harvesting or mosaic cuts (i.e., 50–150 ha blocks separated by equivalent residual forest patches). Mosaic cuts also included riparian buffers, or a protection of the first 20 m of forest adjacent to rivers and lakes (Governments of Québec and the Cree Nation 2002, Jacqmain et al. 2012). Another significant adaptation was the creation of the Sites of Special Wildlife Interest to the Cree (often referred to as 25% areas), where tallymen were asked to identify important wildlife habitat areas on their traplines to a total of 25% of productive area (from a timber harvest perspective) of a trapline. Many Cree land users selected their 25% areas based on valuable hunting areas, most often winter moose habitat or moose yards, while also considering other culturally important sites (Teitelbaum and Lussier 2016). Special management measures were applied within these 25% areas, including the conservation of a higher proportion of mature forest (> 7 m and > 90 years), as well as slowed down harvesting where forests are left to regenerate longer (to 7 m instead of > 3 m) between harvests.

Our study involved four inland Cree communities with traplines within the Adapted Forestry Regime (AFR) territory, which covers about 65,000 km² of southern Eeyou Istchee (Fig. 1). The communities were the Cree First Nations of Mistissini (ᒥᔅᑎᓯᓃ; 50°25′ N, 73°52′W), Nemaska (ᓀᒥᔅᑳᐤ; 51°41′ N, 76°15′W), Oujé-Bougoumou (ᐆᒉᐳᑯᒨ; 49°55′ N, 74°49′W), and Waswanipi (ᐙᔂᓂᐲ; 49°41′ N, 75°57′W). Eeyou Istchee is found within the closed-crown boreal forests of Canada and the AFR is mostly within the western black spruce-feather moss domain under Québec’s forest ecosystem classification system. This domain has a continental subpolar climate, with average annual temperatures of 0 to -2.5 °C, about 1000 mm of precipitation annually, and a growing season of 4 to 5 months (Bergeron et al. 1998). The productive forest of the AFR is dominated by coniferous stands of black spruce (Picea mariana), jack pine (Pinus banksiana), and balsam fir (Abies balsamea; 67%) with scattered areas of mixed (11%) or deciduous stands (2.5%). Mixed stands are typically composed of fir, paper birch (Betula papyrifera), balsam poplar (Populus balsamifera), and trembling aspen (Populus tremuloides). Unforested areas (19%) and waterbodies (11%) cover a relatively large percentage of the territory, and include dry barrens, wetlands, and alder stands (GQ 2019). Lichen-spruce woodlands are found north of the communities of Mistissini and Nemaska, with open-crown forests and scrubland increasing toward the northern limit of Québec’s boreal forest.

Our research approach

Almost 20 years after the implementation of the AFR in Eeyou Istchee, a project on moose habitat quality was proposed to partners by a government researcher who had started a GPS-collar based telemetry study in the region in 2018. After consultations, a collaborative project on moose habitat was initiated in 2020 by agreement between different stakeholders and rightsholders, including the provincial and regional governments, as well as community and academic partners. Researchers from McGill University were invited to help coordinate the project. This project was called “Moose Habitat Quality in Eeyou Istchee under the Adapted Forestry Regime” and was designed to support the development of a Habitat Quality Index (HQI) to assess the evolution of moose habitat quality in the region. Based on a knowledge co-production framework (Huria et al. 2019, Norström et al. 2020, Zurba et al. 2022), our overall research project included three key directions: the analysis of moose telemetry data from GPS collars, the analysis of Cree Knowledge from interviews, and a collaborative process with continued dialogue throughout all research phases (Fig. 2). Although we will focus in this paper on the analysis of the Cree Knowledge interviews within the larger project, more information on other aspects can be found in Stern 2022 and Badry et al. 2024.

A project steering committee was established before the start of the project and consisted of representatives from four Cree communities, as well as representatives from the Cree regional government and trapper’s association, a forestry co-management board, and the provincial government’s Ministère des Forêts, de la Faune et des Parcs (which, over the course of the project, was reorganized into the Ministère des Ressources naturelles et des Forêts and the Ministère de l’Environnement, de la Lutte contre les changements climatiques, de la Faune et des Parcs). This committee of roughly 10 participants met monthly to review project progress and documentation, including research agreements, proposals, methodologies, and preliminary results. Review and support for the project was sought from organizational and community partners during frequent meetings of the steering committee and was formalized with community research agreements and signed resolutions of support from the First Nations’ Chief and Councils. This paper was authored by a team of Cree (EG) and non-Indigenous scholars (GM, NB, IS, GH, MH) with experience in wildlife and forest ecology, community-based research, and Indigenous-partnered research approaches, and was realized through the collaboration and guidance of expert knowledge holders and a steering committee who are named in the acknowledgments. The knowledge holders, steering committee members, and authors approached the work as an opportunity and a responsibility to ensure meaningful inclusion of Cree knowledge in the understanding of habitat-wildlife relationships in Eeyou Istchee. Ethics certificates for research involving humans was obtained from the Ethics Committee of McGill University (#21-08-034 and #20-06-078).

To help further define the project’s objectives and scope, the lead author (GM) also conducted 14 semi-structured interviews; 10 with steering committee members and 4 with other representatives of partner organizations using a snowball sampling strategy. Because of the ongoing COVID-19 pandemic, these project scoping interviews were conducted remotely via one-on-one videoconference from November 2020 to January 2021 and focused on clarifying research priorities, methodologies, and the ongoing involvement of research partners. Partners stated that the overall goal should be to address knowledge gaps in moose habitat in the region and to work together respectfully to build common understandings informed by Western and Cree knowledge. The spatial and temporal scope of the project were also defined collaboratively during this co-production phase. The study area was defined as the southern region of Eeyou Istchee administered under the AFR since the signing of the “Paix des Braves” agreement in 2002 (Fig. 1). Each Cree Knowledge interview was focused within the boundaries of one trapline, or traditional family hunting ground (Indoh-hoh Istchee; Awashish 2018), that was either within, or adjacent to, the AFR territory. Participants were selected so that their collective local knowledge covered a broadly representative range of the study area. The guiding question of the interviews could therefore be stated as “what are the key variables and forestry management practices that have influenced moose habitat quality on your trapline since 2002?” (Table S1).

Fuzzy cognitive mapping

During research co-development, fuzzy cognitive mapping was identified as a suitable methodology for Cree Knowledge interviews. FCMs can help identify and prioritize social-ecological knowledge gaps, bridge the gap between qualitative story lines and quantitative analysis, and facilitate constructive debate while building trust among project partners (Jetter and Kok 2014, Gray et al. 2015, Andersson and Silver 2019). Given these FCM attributes aligned well with goals expressed by partners during the project development phase, researchers proposed a protocol of participatory mapping of moose habitat using FCMs and the project steering committee subsequently reviewed, revised, and agreed on the final mapping protocol.

In their graphical form, FCMs are drawn as a collection of boxes or “nodes” connected by arrows or “directed edges” (Fig. S1). Nodes represent variables (also known as concepts or factors) that are part of the study system and can include elements such as “cultural identity” or “respect” that are hard to quantify. Edges represent causal links (also known as arcs or connections) that denote how much influence one variable has on another. Edges have a direction, a sign, and a weight. The sign can be positive to show a promoting effect, or negative to show an inhibitory effect. Because human perception is not always sharp or precise, edge weights are “fuzzy” estimations of relationships used to describe complex systems (Zadeh 1988). Linguistic statements, such as “winter habitat availability is excellent” and “disturbance is bad for moose” are expressed with numerical values (i.e., weights) ranging from -1 to +1. Weights closer to -1 or +1 will indicate stronger influences, and zero denotes no perceived causal effect. The structure of a FCM is determined by expressing the diagram as an adjacency matrix, where “n” variables are described by a “n x n” matrix. The originating variables (Vi) are listed in the rows and the receiving variables (Vj) in the columns. The value of each cell of the matrix is the edge weight (Wij). An adjacency matrix is typically asymmetric and has values other than zero on the main diagonal when the associated concepts have feedbacks, or self-loops (Andersson and Silver 2019).

The process of building FCMs

Our process for building FCMs involved four main steps: (1) project co-development, (2) knowledge collection, (3) data treatment, and (4) network analysis (Fig. 2). As described above, the process of building FCMs began with collaboratively defining the objectives and scope of the study, the timeframe under consideration, and the key knowledge gap to be addressed (Andersson and Silver 2019). Step 2 was knowledge collection where semi-structured interviews were conducted to create individual and group maps. Step 3 was data treatment and consisted of two key stages. The first was concept matching (or qualitative aggregation) of the 35 individual maps to adopt consistent terminologies for similar concepts (Sarmiento et al. 2022), and the second was quantitative aggregation of the individual FCMs into collective (or social) maps representing the knowledge shared among communities. This step included the condensation of maps into category maps to summarize their content and facilitate communication. The final step was network analysis, where graph theory-based indices were used to describe the complex networks of the collective and community-specific maps (Gray et al. 2015).

Knowledge collection

Selection of interviewees for the FCMs was purposive. The research team and steering committee compiled a list of priority traplines from the four communities; preselected traplines were identified based on estimated levels of forestry disturbance, with the priority being those with the highest levels of disturbance within the 25% areas. Our aim was to interview representatives from preselected traplines, as well as other land users nominated by community coordinators based on their expertise. Of the total completed interviews, about half (48%) were from pre-selected traplines and the rest were with nominated experts. We choose to conduct individual or small group interviews, depending on participant’s preferences, so that the overall model would be guided by multiple perspectives and the localized knowledge of each trapline respondent. Locally based research assistants contacted potential participants by mail, phone, and radio to schedule interviews and those interested were given the opportunity to invite others along who were knowledgeable about moose on their specific trapline.

From October to November 2021, one researcher (NB) led the interviews in the two northern communities and another researcher led interviews in the more southerly communities (GM). Free prior and informed consent was obtained in writing from participants and, according to local custom, an honorarium was provided for participation. No personal data were recorded, other than name and trapline number. Mapping sessions lasted from 1–3 hours with translation from Cree provided by local research assistants. Hand-written notes were taken during interviews, and, when consent was given, audio recordings were also taken. When available, printed topographic maps or projected digital maps of the trapline were used as a visual aide to help participants describe topographic features related to moose habitat.

The session followed a series of open-ended questions designed to facilitate FCM building with participants. Interview questions helped guide the interview and validate relationships drawn on the map (Table S1). Mappers were asked to respond to the questions according to their perception, experience, and expertise. The researchers then drew the FCM diagram during the interview following the instructions of participants. During research co-development, “good moose habitat” was identified as the focus motivating research activities. Facilitators therefore began by posting a post-it with the words “good moose habitat” in a central position (i.e., as the central node) on a whiteboard. Participants were then asked to identify key concepts or variables important for good moose habitat, which the facilitator wrote on post-it notes and placed around the central node on a whiteboard (Andersson and Silver 2019). Next, participants were asked to identify the positive and negative causal relationships among the variables and these were represented with different colored arrows. Participants were then asked to estimate the causal strength of each relationship on a scale of 1 to 5 (weak influence = 1, strong influence = 5; Fig. S2). This final step was the most time-consuming and conceptually challenging for participants. At the end of the interview session, participants were able to review, edit, and confirm the FCMs. All maps were completed to the satisfaction of their authors.

Data treatment

Completed FCMs were digitized using the free software yEd. The next step was concept matching, or qualitative aggregation, to adopt consistent terminologies for similar variables that ended up with different linguistic labels. For example, participants mentioned “mature forests,” “natural mixed forests,” “mixed stands,” and “old forests” and these variables were combined into a single “mature and mixed forests” variable. Sometimes the same variable was characterized as opposites (e.g., “less snow” and “deeper snow”) and these were converted to one standard label and the corresponding weights of the connecting edges were reversed as required. Using inductive thematic analysis, the two lead researchers (GM, NB) then condensed the list of standardized variables into larger categories to facilitate interpretation and communication of results. Categories represented broad concepts that were discussed during interviews and were identified with the help of detailed notes and transcripts that provided context and the detailed explanation of each variable name.

During interviews, we were not able to obtain causal weights for all relationships and some maps were missing 10 to 30% of edge weights. We therefore could not use participant weighting in the analysis step, but used another procedure based on discourse analysis to derive weights based on the frequency of the relationships across all maps. For consistency, we replaced weights from all interviews (for both the completely- and incompletely-weighted participant maps) with values of either -1 or +1 to indicate negative or positive relationships. Relative frequency was then calculated. We chose this technique because it is independent of researcher assumptions and has been shown to yield similar results to participant weighing when there is a large sample size (Sarmiento et al. 2022). Fuzzy transitive closure was also calculated on the adjacency matrix of each map using the software CIETmap 2.2. Fuzzy transitive closure identifies all possible direct and indirect connections between nodes. The model calculates the total influence of each connection based on the weight of the weakest arrows involved. The maximum influence that one node has on another corresponds to the strongest possible connection between those two nodes. The resulting map has a complexified architecture reflecting a more complete measure of the influence of each variable on others (Niesink et al. 2013, Andersson et al. 2017).

After transitive closure, the next step was quantitative aggregation of the FCMs into five aggregated maps, one map for each of the four communities and one collective map with all individual maps included. The weight of an edge in the aggregated maps was calculated as the sum of the weights for the same edge across the maps, divided by the maximum sum calculated across all edges. The category-level weight was calculated as the sum of the weights of the grouped variable-level relationships. To normalize the category-level weights for comparison between aggregated maps, we divided all category-level relationships by the maximum sum found across all category-level edges. Therefore, the values closer to 1 or -1 indicate stronger influences, while those closer to 0 indicate weaker influences.

Network analysis

We used graph theory-based indices on the category maps, calculating weighted indegree and outdegree centrality for each category (Özesmi and Özesmi 2004). Weighted indegree shows the extent to which each category is affected by other categories considering the number and relative strength of arrows that enter a node. Weighted outdegree indicates the extent to which each category affects other categories based on the number and relative strength of arrows exiting a node. A higher sum indicates a higher centrality measure. Given mapping sessions situated “good moose habitat” as the initial and central starting point, a high degree of centrality for this category was expected, whereas for all other identified categories and variables, centrality was expected to vary. Feedback strengths were also calculated for each category. Feedbacks were created when different variables within the same category activate (or de-activate) each other in a cyclic manner, and therefore increase or decrease the overall strength of the category on itself.

Member-checking

We conducted four member checking workshops and community presentations in July 2022 to establish trustworthiness in the results (Birt et al. 2016). There were a total of 79 participants from the four participating communities, including participants of the initial mapping interviews, as well as other interested community members. Workshops allowed for the clarification of concepts used during the individual fuzzy cognitive mapping interviews. Workshop participants reviewed and elaborated on the categories that were created during the qualitative aggregation step described above, and on their relative importance to good moose habitat. These workshops contributed to the final framing and discussion of the research findings.

RESULTS

We conducted 37 FCM sessions with 56 participants from 4 Cree communities and this resulted in 35 completed maps (two interviews did not result in maps; Fig. S2 and S3). Most participants were tallymen, the designated stewards of family hunting territories, or other land users with expertise on moose within specific traplines. Other participants were family members of tallymen. The majority of participants were middle-aged (~35 to 60 years old). About 60% of interviews were conducted with a single male tallyman or land user and the rest were small groups of 2–4 participants. About 30% of participants in small group interviews were female, either co-tallymen, spouses, or family members. Almost half of the interviews took place in Waswanipi (n = 15), which has traplines that cover over 50% of the study area and whose territory covers many of the southern traplines that are more heavily affected by forestry.

When asked to define the central node, i.e., explain what the term “good moose habitat” meant to them, participants spoke about areas where moose live, where moose like to go, where moose food is, or where moose stay during different seasons. Many participants mentioned the importance of moose movements between different seasonal areas for feeding, sheltering, calving, and mating. Typical definitions were descriptions of key habitat features, such as, “A nice valley between two mountains,” “A mixed forest of mature stands and young stands preferably,” “A mix of conifers and deciduous trees,” and “A mountainous area with lots of water nearby during the summertime.” Some participants defined good moose habitat as areas with low disturbance, i.e., areas without helicopters, or “no forestry roads, no camps nearby,” or mountainous areas that were undisturbed or not “fragmented” by forestry cuts. Participants said they were familiar with many of the local areas considered to be good moose habitat, stating, for example, that “It’s usually the same areas they hit. And we know where they are, at what time, and the season” (Participant from Mistissini).

During data treatment, we used concept matching to identify 144 standard factors based on the labels of the 608 factors mentioned across all the maps (Table S2). Of these 144 variables, 56 were unique or non-duplicated among the individual maps. The community of Nemaska had the greatest number of unique variables (n = 37) compared to the other communities (n = 3–9). The average individual FCM had 30 variables and 45 edges. After the calculation of transitive closure, the number of edges usually doubled or tripled to about 100. Accumulation plots indicate that the study’s sample size was sufficient, with the total number of new variables decreasing with each new interview until they were close to zero (Fig. S2 and S3). Using thematic analysis, we condensed the 144 variables into 18 broader categories (Table S2). In Table 1, we present the final classification of categories (names in italics), with the number of variables in each category and a short description of the category and key variables.

For the collective FCM, which combined data from all four communities, the top six most influential categories that were directly linked to good moose habitat are shown in Figure 3. From highest to lowest relative strength, these categories were: “Habitat Features, Hunting & Predation, Moose Forage, Forestry & Access, Noise & Disturbance, and Other Resource Development” (Fig. 3, Table 2, Table S4). Overall, “Habitat Features” and “Moose Forage” were perceived to have an overall positive influence on good moose habitat, whereas the other categories were said to have negative influences on habitat. “Habitat Features” had the strongest positive influence on moose habitat (Table 2, Fig. 4). The most influential variables within this category were mountains, valleys, and mature and mixed forests, which were considered to be excellent winter moose habitat, or moose yards (Fig. 5b). Calving areas were also included in this category and these swampy or riparian areas had a strong positive influence on moose habitat. Habitat connectivity was also mentioned as important. The category “Moose Forage” also had a strong positive influence on moose habitat and included food sources (willows, birch, alder, balsam), regrowth after disturbance, and aquatic plants (Table 2, Fig. 4).

“Hunting & Predation” was observed to have the strongest negative direct influence on good moose habitat (Table 2, Fig. 4). High hunting pressure was largely caused by non-Cree hunters, poaching, Cree hunters, and predators, in order of decreasing influence (Fig. 5a). Both non-Cree and Cree hunters contributed to poaching, with non-Cree more strongly linked. The main predator across the study area was wolves, with the occasional mention of bears. Both “Forestry & Access” and “Noise & Disturbance” had relatively strong negative influences on moose habitat (Table 2, Fig. 4). For “Forestry & Access,” all types of forestry cuts and the construction of forestry roads were perceived to reduce the availability of moose habitat, especially in critical winter habitat or moose yards, as well as reducing habitat quality via the increase in windthrow and debris on the landscape, which restricted moose mobility. During member-checking workshops, participants were asked to clarify whether different road types affected the strength or direction of relationships. The consensus was that a “road is a road” and that “roads mean more access for everyone” because even temporary winter roads (built for short-term forestry operations) stayed accessible for snowmobiles in the winter. “Noise & Disturbance” was defined by most participants as noise disturbance caused by human activity, typically related to the use of technologies like vehicles, aircraft, and hunting equipment. The proximity of non-Cree camps and communities was strongly linked to “Noise & Disturbance.” “Other Resource Development” had a moderate negative overall influence on moose habitat (Table 2, Fig. 4) and included mining, hydroelectric development, and powerlines, in decreasing order of relative influence.

These most influential categories with direct links to the central node were similar between the collective map and the four community maps. As shown in Table 2, “Habitat Features” and “Hunting & Predation” were the top two categories for all four communities, and “Moose Forage” and “Noise & Disturbance” were also ranked highly across all communities. “Forestry & Access” had the strongest perceived negative influence in Oujé-Bougoumou and Mistissini. The category “Other Resource Development” had the strongest negative influence on Nemaska maps and was relatively less influential in Waswanipi. The category “Education & Knowledge” had the strongest perceived positive influence in Nemaska and Mistissini and the “Moose Movement” category was mentioned more frequently in the two southernmost communities, Oujé-Bougoumou and Waswanipi.

Indegree and outdegree centrality were also calculated to show the relative significance of each category within the larger system illustrated by the map. “Good Moose Habitat” ranked the highest for indegree centrality on the collective map (Fig. 4), which was to be expected given our methodology involved a priori identification of the central node. This indicates that the central node was the most strongly affected by all the other categories on the map. Other categories with high relative indegree (i.e., arrows going inward) on the collective map were “Hunting & Predation,” “Habitat Features,” “Moose Population,” and “Cree Culture.” For outdegree centrality (i.e., arrows going outward), the categories “Forestry & Access,” “Habitat Features,” “Hunting & Predation,” “Noise & Disturbance,” and “Other Resource Development” ranked most highly on the collective map (Fig. 4). Outdegree centrality shows which categories had the most important relative influence on the map as a whole, i.e., not only on the central node, but also on other categories on the map. The strongest feedback loops were all positive and were found for the categories “Forestry & Access,” “Habitat Features,” and “Hunting & Predation” on both the collective and community maps (Fig. 5, Table S5). The individual variables that contributed to these strong positive feedbacks within each category are illustrated in Figure 5.

DISCUSSION

Our study summarizes Cree understanding of contemporary drivers of moose habitat quality after the implementation of an Adapted Forestry Regime (AFR) intended to better reconcile forestry, wildlife conservation, and Cree way of life. We gained understanding about the perspectives, observations, and experiences of Cree Knowledge holders using a knowledge-co-production approach focused on a local knowledge priority and involving fuzzy cognitive mapping interviews and member-checking workshops. Fuzzy cognitive mapping allowed us to identify key variables influencing moose habitat quality and to examine the relative importance of these variables from the combined perspectives of 56 Cree land users. The analysis identified 18 categories that influence good moose habitat, including physical (“Climate & Weather”), ecological (“Habitat Features, Moose Forage”), and social contributors (“Hunting & Predation, Cree Culture”). Our findings highlight how fuzzy cognitive mapping can bring together individual expertise into a collective knowledge account, which is inclusive of multiple perceptions and experiences and allows for the identification and ranking of variables and relationships. Our results also highlight the diverse interrelationships that land users know to influence moose habitat quality and point to key social variables (hunting activity, noise disturbance) that should be included in wildlife-habitat models. Finally, this study highlights specific aspects of forestry practice and management that continue to negatively impact moose populations despite the implementation of a co-management regime.

Collective knowledge using fuzzy cognitive maps

Adoption of knowledge co-production (Norström et al. 2020, Zurba et al. 2022) and participatory mapping approaches contributed to the successful recruitment of participants and the completion of interactive interviews. Early collaboration via project-scoping interviews and meetings with the steering committee helped us co-construct a context-based project organized around the pre-existing needs, interests, and goals of different partners. Context-based co-production allowed for a shared understanding of the project’s objectives and scope, using methodologies adapted to apply both Western and Cree types of knowledge to address knowledge gaps about moose habitat in the region. Using an interactive, flexible, and goal-orientated approach led to high attendance and engagement from partners at steering committee meetings, which ensured regular feedback and the local engagement and logistical support critical to the success of interviews and workshops. Frequent consultations with the committee meant that documents were written in relevant language, and translated so that key concepts could be properly communicated to interview participants. For example, an infographic on moose collaring was created with the steering committee to help inform concerns about the deployment of collars and their effects on moose. The relative simplicity of the FCM approach, with minimal technical skills required during interviews, facilitated engagement and knowledge sharing with participants, some of whom reported learning from and enjoying the process. Following completion of interviews, our approach enabled multiple, iterative discussions during steering committee meetings and member-checking workshops about how best to categorize and interpret variables from the individual FCMs. These types of conversations helped promote mutual learning, constructive debate, and trust-building between researchers, partners, and decision makers (Andersson and Silver 2019). FCM also provided a framework for participants, researchers, and partners to discuss linkages between system components on the maps and to highlight areas of consensus and divergence (Hobbs et al. 2002, Khan and Quaddus 2004).

One of our objectives was to bring together the individual knowledge of multiple Cree experts, each familiar with one or two traplines in the study region, into a collective knowledge account. We were successful in achieving a high level of participation from knowledge holders spread across a large and spatially diverse study area, including from traplines with extensive forestry disturbance, as well as from traplines with little or no disturbance, and from other experts nominated by community members. FCM provided a flexible, intuitive, and transparent method to organize and summarize this wealth of Cree Knowledge (Gray et al. 2015). Bringing together the knowledge of individuals who held multiple perspectives on moose habitat into collective models allowed us to identify key variables with higher influence on the overall system, as well as to compare differences at the community (Table 2) and individual level (Table S2; Voinov et al. 2018). These models also helped us identify and rank variables and relationships, while also exploring the underlying beliefs, assumptions, and consensus (or lack thereof) among knowledge holders (Jetter and Kok 2014). For example, we could highlight six key categories that were observed to have higher influence overall, both positive and negative, on the quality of moose habitat across the whole study region (namely “Habitat Features,” “Hunting & Predation,” “Moose Forage,” “Forestry & Access,” “Noise & Disturbance,” and “Other Resource Development”). We also observed that knowledge holders from different communities across a large region shared relatively similar understanding of top positive and negative contributors to good moose habitat in the region.

Modeling with FCMs is designed to be simple, which has many advantages in a participatory study using a knowledge co-production approach, however there are also limitations. Models with FCMs do not integrate different temporal or spatial scales very easily. We were unable to adequately capture the effects of seasonality on the individual maps, although seasonality was identified as a key variable influencing habitat quality by many of the participants. FCMs capture the state of a knowledge system at a specific moment in time (i.e., autumn 2021 for this study); as social-ecological conditions change so too will knowledge holder understanding of key factors. For example, forest fires were evaluated as having a neutral influence on moose habitat quality in this study, causing short-term habitat loss but also being beneficial in the longer term because of the promotion of good quality regrowth for forage. However, following the extensive and record-setting fires that occurred in the study region in the summer of 2023, we could expect that participants might represent the overall role of fire within the system differently. Nonlinear relationships, like thresholds, are also difficult to include and FCMs are usually limited to defining linear relationships within a system (Gray et al. 2015).

Furthermore, the aggregation of cognitive maps and creation of broad categories may lead to the loss of knowledge heterogeneity, especially within a diverse participant group, and raises questions about how to accurately represent “scaled-up” knowledge (Hobbs et al. 2002, Gray et al. 2014, Edwards and Kok 2021). Although using FCMs can create a wide range of opportunities for semiquantitative models to be combined with more quantitatively based management assessments, key challenges related to ontological pluralism may lead to certain important aspects of Indigenous Knowledge being de-emphasized or de-contextualized within an integrated social-ecological model (see Badry et al. 2024). As described in previous studies, the challenge is to find appropriate methods of bridging or weaving knowledge systems that respect the integrity, complexity, and diversity of Indigenous Knowledges (Johnson et al. 2016, Berkes 2017, Reid et al. 2021). Despite these limitations, fuzzy logic-based approaches, like FCMs, are considered to be a good fit for knowledge co-production with Indigenous types of Knowledge (Berkes and Berkes 2009, Tengö et al. 2014) because they can consider a large number of variables qualitatively, instead of a small number of variables quantitatively as is typical of wildlife-habitat models.

Social-ecology of moose habitat quality

In the literature, wildlife-habitat relationships are most often modeled based on biophysical or ecological factors, which for moose typically include forest and vegetation characteristics, topography, and waterways (Jacqmain et al. 2008, Dussault et al. 2006, Dettki et al. 2003). Although these factors are important, there is growing evidence that social factors such as off-road vehicle use (Shanley and Pyare 2011, Ploughe and Fraser 2022), hunting activity (Rempel et al. 1997, Brown et al. 2018, Brown 2011), and other social relations (Behr et al. 2017) can alter wildlife distribution, movements, and habitat selection. Models of wildlife habitat inclusive of a broader array of social and ecological drivers are rare, in part because information about and understanding of potential drivers is often discipline-, jurisdiction-, and sector-specific. The inclusion of social variables in wildlife models can be challenging, often requiring novel or interdisciplinary methods, and leading to greater complexity in model relationships. The use of participatory modeling, such as FCM, is increasing popular as a method for summarizing complex social-ecological systems (Collins et al. 2011) and has the potential to support planning and decision-making processes (Hedelin et al. 2021). In the present study, the understandings of moose habitat shared by Cree experts highlights the relevance of prioritizing both ecological and social variables in moose habitat models for this region and also points to how local knowledge can inform management decisions related to wildlife and forestry and scenario-planning.

Based on the FCMs, Cree Knowledge holders emphasize many of the same biophysical or ecological variables usually included in wildlife-habitat models. According to Cree participants, the category “Habitat Features,” focused mainly on forest composition and topography, had the highest positive influence overall on habitat. Participants also described the importance of the connection between key habitats and the quality of “Moose Forage” (Fig. 3), specifically the presence of willows, birch, alder, balsam fir, and aquatic plants in key habitats. The connection between the “Habitat Features” and “Moose Forage” categories in the present study helps describe how these ecological variables were observed by Cree land users to impact moose habitat at a regional scale in Eeyou Istchee; key information when we consider that published telemetry data shows fairly variable moose habitat selection across regions. For example, telemetry studies suggest that proxies for snow depth, including elevation and solar insolation (i.e., south and west-facing slopes), are key drivers of winter range distribution across North America (Peek 1997, Poole and Stuart-Smith 2006). However, the importance of mature coniferous stands in winter is debated. Some studies show selection for these habitats in winter (Balsom et al. 1996, Thompson and Stewart 1997, Tyers 2003) and others show more use of riparian areas and lower elevations (Poole and Stuart-Smith 2006). Habitat preference for calving areas also appears to be highly variable (citations in Poole et al. 2007). In the present study, Cree land user’s observations highlight the specifics of local habitat selection, with the two habitat types most frequently identified being winter habitat (of mature mixed forests in upland areas and the adjacent valley bottoms) and spring calving areas (in swampy and riparian areas, occasionally on islands). This emphasis on the importance of winter habitats and riparian areas in this region has previously been stated by the Cree in studies from decades ago (Feit 2004) and from the early years of the AFR (Jacqmain et al. 2008, 2012).

In contrast to most wildlife-habitat models described in wildlife science literature, the ecological knowledge described by the Cree emphasize influential social variables, specifically the categories “Hunting & Predation,” “Noise & Disturbance,” “Forestry & Access,” and “Other Resource Development.” As seen in Fig. 3, “Habitat Features” helped mitigate the negative effects of “Hunting & Predation” on moose habitat; mountains, islands, and calving areas created shelter from predators and hunters, and both hunting pressure and “Forestry & Access” were observed to be lower where terrain was steep and inaccessible and where habitats were well-connected. “Habitat Features” was also related to “Forestry & Access,” where forestry cuts were perceived to reduce the availability of moose habitat, especially in critical winter habitat, as well as reducing habitat quality via the increase in windthrow and debris on the landscape, which restricted moose mobility. As described in more detail below, the categories “Habitat Features,” “Moose Forage,” and “Forestry & Access” were also closely connected and highly influential for the overall quality of moose habitat.

Social variables were predominant within the FCMs in this study. From participants’ perspectives, multiple and interconnected drivers negatively influence moose habitat quality in this region. There was consensus across all communities that “Hunting & Predation” plays a major role in reducing the quality of moose habitat. The presence of hunters and poachers within a territory who were hunting too much or too frequently, hunting the wrong moose (by sex or age), or hunting “the wrong way” (e.g., at night), were considered by the Cree to negatively affect habitat quality. “Forestry & Access” was strongly connected to “Hunting & Predation” on the collective FCM (Fig. 3, weight of +1.00) showing that many Cree participants perceived that forestry practices led to increased hunting pressure:

Why can’t they leave a stand of trees along some forestry roads, so you don’t see that far away? Because it gets very simple to kill moose, and that’s why there’s a lot of overharvesting. Not only that, the road access promotes a lot of hunting in the territory, not only for natives but also for non-native people (Participant from Waswanipi).

Snowmobiles were seen as especially problematic for hunting pressure and access, as users could get to remote areas using smaller trails or one-season winter roads. Many participants also felt that the erosion of the authority by the traditional land managers, or tallymen, was a significant contributor to the negative impacts of “Hunting & Predation.” Other studies have also observed that increased access via extensive road networks and the reduction in forest cover alongside roads increases moose vulnerability to hunting and predation (Rempel et al. 1997, Jacqmain et al. 2005).

During project meetings and workshops, some partners with natural science training questioned the relationship between “Hunting & Predation” and moose habitat quality, wondering if there was confusion among Cree knowledge experts causing them to conflate factors affecting population status and habitat quality. A close reading of the interview transcripts does indicate that Cree participants distinguished between the concepts of habitat quality (central node) and population status (included in the category “Moose Population” along with variables related to moose health). For many participants, the strong negative connection between “Hunting & Predation” and “Good Moose Habitat” reflected the reality that high quality moose habitat was safe, undisturbed, and quiet for the moose, with “no forestry roads, no camps nearby,” a reality informed by their own observations that moose select habitats in part to avoid risk or disturbance from hunters (as seen in Brown et al. 2018). Indeed, participants described this disturbance-avoidance behavior as related to a nexus of social factors, including both “Hunting & Predation” and “Forestry & Access,” but also the “Noise & Disturbance” and “Other Resource Development” categories. Noise disturbance, due to the presence of motorized vehicles, aircraft, hunting equipment, and non-Cree settlements was perceived as having a negative influence on habitat quality (as seen in Shanley and Pyare 2011, Ploughe and Fraser 2022). For some of the communities, especially Nemaska, the cumulative effects of “Other Resource Development,” including mining, hydroelectric development, and powerlines, contributed significantly to the negative effects of disturbances from human activities, forestry operations, and overhunting. One participant observed:

Noise and the presence of people and trucks going back and forth [is disturbing, and the moose] want a quiet place. If there is more development, the moose may move further and further away to get away from disturbance (Participant from Mistissini).

FCM mapping of Cree knowledge emphasizes the importance of considering both social and ecological relationships that can influence habitat quality. On the collective map, bringing together the observations of 56 participants from 4 Cree communities, the two most influential categories directly related to good moose habitat were “Habitat Features” (+0.87 relative weight) and “Hunting & Predation” (-0.75 relative weight). The strongest overall relationship was the positive connection between “Forestry & Access” and “Hunting & Predation” (Fig. 3, weight of +1.00). These results point to how local knowledge can help inform wildlife management decisions by highlighting key social variables (e.g., hunting activity, disturbance) affecting wildlife and their habitat. These results reflect the worldview of Cree participants who see humans as part of (and not apart from) the ecosystem, emphasizing relationships of respect for both the land and all living beings (Berkes and Berkes 2009).

Current forestry practices and moose habitat

More than 15 years after the implementation of the Adapted Forestry Regime (AFR) in Eeyou Istchee, our study aimed to provide an updated picture of how Cree experts understand the current effects of forestry practices in this region of the boreal forest of Northern Québec. In 2002, the Cree and the Government of Québec established a co-management agreement that sought a more collaborative approach to forest and other resource management in Eeyou Istchee. Based on consultations and the results of previous studies in the region (e.g., Feit 1987, Jacqmain et al. 2005), the AFR was created to allow for adaptations to forestry practices that better take into account the traditional Cree Way of Life (Eeyou Pimatseewin) and allow for more consultations with Cree regarding forestry operations and management of wildlife areas. Our study provides a replicable assessment of Cree knowledge regarding the efficacy of many of the AFR adaptions related to wildlife habitat. Despite more than a decade under this co-management agreement, where “specific management standards are applied to maintain or improve the habitat of important wildlife species” (Governments of Québec and the Cree Nation 2002), the majority of Cree participants in the present study found that forestry practices continue to have a significant and negative impact on the overall quality of moose habitat.

Outdegree centrality was highest for “Forestry & Access” indicating that this category most strongly affected all the other categories within the mapped system. “Forestry & Access” was closely connected to “Habitat Features” and “Moose Forage” on the collective map (Fig. 3) and all categories were highly influential for the quality of moose habitat. All types of forestry cuts (clearcuts, mosaic, partial) and the construction of extensive road networks under the AFR were perceived to directly reduce the availability of moose habitat, especially in critical winter habitat (or moose yards). According to the Cree participants, the influence of “Forestry & Access” on “Moose Forage” was mixed and this relationship had a neutral weight on the collective map. Standard forest management guidelines focus on the benefit of harvesting mature stands to create moose forage through the regeneration of young forest stands and studies have shown increased moose use and population increase after both fires and cuts (Collins and Schwartz 1998, Courtois et al. 2002, Potvin et al. 2005). Most Cree participants indicated that forest fires promoted good forage habitat, while only some indicated that forestry cuts led to forage regrowth. Older cuts (> 30 years approx.) were generally thought to provide good quality regrowth for forage, whereas opinions were split on whether this was also true for the more recent mosaic cutting techniques (i.e., variable retention harvesting). Differences in opinion were related to the degree of impact that land users attributed to post-logging silviculture practices on forest regrowth in the more recent cuts (category “Forest Management”).

In the “Forest Management” category, some land users considered that current techniques of scarification, replanting, and brush cutting were bad for regrowth and moose forage, while others observed that moose often returned to recent cuts to forage. Soil scarification, or the striping away of the organic soil layer with heavy machinery, is typically considered beneficial for deciduous regrowth by managers (Collins and Schwartz 1998). However, some Cree land users in the more southerly communities observed that scarification led to disturbed soils, hence less vegetation and lower quality regrowth.

Many do not agree with scarification after forestry. It ruins the earth. Back when they used selective cutting, it was much better. The big machines ruin the ground. It would be better if they just left it the way it was after cutting (Participant from Oujé-Bougoumou).

Replanting was also seen as negative for moose forage, and therefore for moose habitat quality, as the majority of replanted trees in the region are commercially valuable species (e.g., black spruce, jack pine). Participants were concerned about areas that once had poplar and birch being replanted with conifers, and that replanted trees were often different, non-native, or low quality. One participant from Waswanipi commented that “Replanted trees aren’t good trees. It’s plant junk food.” Pre-commercial thinning of regenerating stands, or brush cutting, was also perceived as negative for moose forage because it removed deciduous and balsam fir regrowth and reduced tree density, and therefore cover for moose. In line with some of these observations, published studies in other regions have found that moose avoid recently thinned stands (McLaren et al. 2000) and that natural regeneration of clear-cut areas supports higher abundance of moose forage when compared to stands with post-harvest silviculture (and post-fire stands; Thompson et al. 2003, Boan et al. 2011, McKay and Finnegan 2023). One participant stated that,

Silviculture work is done for the purpose of forestry and forestry only. Areas are affected three times: cut, scarification, brush cut. The same moose yard can be disturbed three times, and it doesn’t grow back properly (Participant from Waswanipi).

Moreover, Cree land users felt that most of the adaptations currently implemented under the AFR, including 20 m riparian buffers adjacent to waterbodies and the inclusion of winter moose yards within the 25% wildlife areas, were largely insufficient to protect these two essential habitats. Participants stated that riparian buffers were too narrow and were therefore vulnerable to windthrow, increasing predation risk and impeding travel routes for both moose and land users. Despite the special measures within the 25% areas, participants continued to feel that moose winter yards were being overharvested, especially in more recent years. Some participants had understood that these areas would be protected under the AFR, but were now finding that forestry companies were increasing mosaic cutting in the 25% areas because they were “running out of wood” on the rest of the trapline. Forestry consultations (also included in the “Forest Management” category) were seen by most as a good development under the regime, with a strong potential to promote good moose habitat; however, many land users complained that current consultations were ineffective because their input (e.g., on the location of residual forests and access roads) were not always taken into account. Many participants stated that they were consulted on forestry but not always “listened to,” for example, one participant reported,

The logging company thinks we forget about the areas that we already mapped out [in the 25% areas]. So they keep coming, every year, with the same areas they want to cut, and we keep telling them, no, no, no. And they just keep asking and asking. When somebody tells me no, they just have to tell me once. Why can’t they understand a no? (Participant from Oujé-Bougoumou)

These perceptions from Cree land users help explain why the category “Forest Management,” which included regulations on residual forests, buffers, silviculture, 25% areas, and forestry consultations, had almost no positive influence on the quality of good moose habitat on the collective map.

Overall, our study highlights that many Cree land users in Eeyou Istchee find that current forestry practices and management continue to have a significant and negative impact on moose habitat quality, despite all the adaptions under the AFR. After the signing of the AFR in 2002, an implementation period was established where the parties were to ensure that the AFR adaptations were incorporated gradually into the annual timber harvesting program. Although many changes to forestry practices were implemented before the end of this period in 2008, certain measures stipulated under the AFR have yet to be carried out. Directives for the protection and development of wildlife habitats are still outstanding, yet were supposed to be drafted and introduced into the forest management planning process in 2003. The creation of a mixed-forest management strategy (another stipulation under the AFR) was also much delayed and was only published by the provincial government in 2020 (Dallaire et al. 2020). Dissatisfaction of Cree land users with the current AFR adaptations (silviculture, buffers, 25% areas), as well as the delay of implementation for key wildlife-related measures, can help to explain the generally negative perception of forestry practices and management under this co-management region.

CONCLUSIONS

Our findings highlight how fuzzy cognitive mapping, situated within a knowledge co-production approach, can bring together individual Cree understandings into a collective knowledge account of relevance to natural resource policy, assessments, co-management, and decision making. Rather than confining or reducing Cree knowledge of moose habitat quality to a narrower set of variables and concepts typically included in wildlife habitat models, the FCM approach successfully identified a broader range of social and ecological determinants of moose habitat quality and positioned these Cree-identified drivers alongside and in relation to variables included in more reductionist models. The next stage of the project is to combine the results of the Cree knowledge mapping presented here with our analysis of moose telemetry data from GPS collars to arrive at a consensus habitat quality model inclusive of Cree knowledge and quantitative wildlife science. We hope that these co-produced models can contribute to portraying the plurality of Cree understandings of moose habitat quality to better inform the protection of wildlife habitats supporting the traditional Cree Way of Life, Eeyou Pimatseewin.

LITERATURE CITED

Andersson, N., M. Beauchamp, E. Nava-Aguilera, S. Paredes-Solís, and M. Šajna. 2017. The women made it work: fuzzy transitive closure of the results chain in a dengue prevention trial in Mexico. BMC Public Health 17(S1):408. https://doi.org/10.1186/s12889-017-4301-0

Andersson, N., and H. Silver. 2019. Fuzzy cognitive mapping: an old tool with new uses in nursing research. Journal of Advanced Nursing 75(12):3823-3830. https://doi.org/10.1111/jan.14192

Awashish, P. 2018. A brief introduction to the Eeyou traditional system of governance of hunting territories (Traditional Eeyou Indoh-hoh Istchee Governance). Anthropologica 60(1):1-4. https://doi.org/10.3138/anth.60.1.t01

Badry, N. A., G. A. MacMillan, E. R. Stern, M. Landry-Cuerrie, G. M. Hickey, and M. M. Humphries. 2024. Boundary spanning methodological approaches for collaborative moose governance in Eeyou Istchee. Environmental Management 74:132-147. https://doi.org/10.1007/s00267-023-01918-6

Balsom, S., W. B. Ballard, and H. A. Whitlaw. 1996. Mature coniferous forest as critical moose habitat. Alces: A Journal Devoted to the Biology and Management of Moose 32:131-140.

Behr, D. M., A. Ozgul, and G. Cozzi. 2017. Combining human acceptance and habitat suitability in a unified socio-ecological suitability model: a case study of the wolf in Switzerland. Journal of Applied Ecology 54(6):1919-1929. https://doi.org/10.1111/1365-2664.12880

Bergeron, J., P. Grondin, and J. Blouin. 1998. Rapport de classification écologique: pessiere a mousses de l’ouest. Ministère des Ressources naturelles et de la Faune, Direction des inventaires forestiers, Québec, Québec, Canada.

Berkes, F. 2017. Sacred ecology. Routledge, New York, New York, USA. https://doi.org/10.4324/9781315114644

Berkes, F., and M. K. Berkes. 2009. Ecological complexity, fuzzy logic, and holism in indigenous knowledge. Futures 41(1):6-12. https://doi.org/10.1016/j.futures.2008.07.003

Birt, L., S. Scott, D. Cavers, C. Campbell, and F. Walter. 2016. Member checking: a tool to enhance trustworthiness or merely a nod to validation? Qualitative Health Research 26(13):1802-1811. https://doi.org/10.1177/1049732316654870

Boan, J. J., B. E. McLaren, and J. R. Malcolm. 2011. Influence of post-harvest silviculture on understory vegetation: implications for forage in a multi-ungulate system. Forest Ecology and Management 262(9):1704-1712. https://doi.org/10.1016/j.foreco.2011.07.022

Bradshaw, C. J. A., I. G. Warkentin, and N. S. Sodhi. 2009. Urgent preservation of boreal carbon stocks and biodiversity. Trends in Ecology & Evolution 24(10):541-548. https://doi.org/10.1016/j.tree.2009.03.019

Brown, C. L., K. Kielland, T. J. Brinkman, S. L. Gilbert, and E. S. Euskirchen. 2018. Resource selection and movement of male moose in response to varying levels of off-road vehicle access. Ecosphere 9(9):e02405. https://doi.org/10.1002/ecs2.2405

Brown, G. S. 2011. Patterns and causes of demographic variation in a harvested moose population: evidence for the effects of climate and density-dependent drivers. Journal of Animal Ecology 80(6):1288-1298. https://doi.org/10.1111/j.1365-2656.2011.01875.x

Burton, P. J., Y. Bergeron, B. E. C. Bogdanski, G. P. Juday, T. Kuuluvainen, B. J. McAfee, A. Ogden, V. K. Teplyakov, R. I. Alfaro, D. A. Francis, S. Gauthier, and J. Hantula. 2010. Sustainability of boreal forests and forestry in a changing environment. IUFRO World Series 25:249-282.

Bussey, J., M. A. Davenport, M. R. Emery, and C. Carroll. 2016. “A lot of it comes from the heart”: the nature and integration of ecological knowledge in tribal and nontribal forest management. Journal of Forestry 114(2):97-107. https://doi.org/10.5849/jof.14-130

Collins, S. L., S. R. Carpenter, S. M. Swinton, D. E. Orenstein, D. L. Childers, T. L. Gragson, N. B. Grimm, J. M. Grove, S. L. Harlan, J. P. Kaye, A. K. Knapp, G. P. Kofinas, J. J. Magnuson, W. H. McDowell, J. M. Melack, L. A. Ogden, G. P. Robertson, M. D. Smith, and A. C. Whitmer. 2011. An integrated conceptual framework for long‐term social-ecological research. Frontiers in Ecology and the Environment 9(6):351-357. https://doi.org/10.1890/100068

Collins, W. B., and C. C. Schwartz. 1998. Logging in Alaska’s boreal forest: creation of grasslands or enhancement of moose habitat. Alces: A Journal Devoted to the Biology and Management of Moose 34(2):355-374.

Courtois, R., C. Dussault, F. Potvin, and G. Daigle. 2002. Habitat selection by moose (Alces alces) in clear-cut landscapes. Alces: A Journal Devoted to the Biology and Management of Moose 38:177-192.

Cyr, F.-X., S. Wyatt, and M. Hébert. 2022. Power-sharing between the Cree and Québec governments in Eeyou Itschee (Québec, Canada): sovereignties, complexity, and equity under the Adapted Forestry Regime of the Paix des Braves. International Forestry Review 24(3):345-359. https://doi.org/10.1505/146554822835941823

Dallaire, S., C. Dion, and E. Cyr. 2020. Mixedwood stands management strategy: for forested lands covered by the Adapted Forestry Regime of La Paix des Braves. Gouvernement du Québec, Ministère des Forêts, de la Faune et des Parcs, Québec, Québec, Canada.

Dettki, H., R. Löfstrand, and L. Edenius. 2003. Modeling habitat suitability for moose in coastal Northern Sweden: empirical vs process-oriented approaches. AMBIO 32(8):549-556. https://doi.org/10.1579/0044-7447-32.8.549

Dussault, C., R. Courtois, and J.-P. Ouellet. 2006. A habitat suitability index model to assess moose habitat selection at multiple spatial scales. Canadian Journal of Forest Research 36(5):1097-1107. https://doi.org/10.1139/x05-310

Edwards, G. I., and K. Kok. 2021. Building a fuzzy cognitive map from stakeholder knowledge: an episodic, asynchronous approach. Current Research in Environmental Sustainability 3:100053. https://doi.org/10.1016/j.crsust.2021.100053

Feit, H. 2004. James Bay Crees’ life projects and politics: histories of place, animal partners and enduring relationships. Pages 92-110 in M. Blaser, H. A. Feit, G. McRae, editors. In the way of development: Indigenous peoples, life projects and globalization. Zed Books, London, UK. https://doi.org/10.5040/9781350220720.ch-006

Feit, H. A. 1973. The ethno-ecology of the Waswanipi Cree: or how hunters can manage their resources. Pages 115-125 in B. Cox, editor. Cultural ecology: readings on the Canadian Indians and Eskimos. McClelland and Stewart Limited, Toronto, Ontario, Canada.

Feit, H. A. 1987. North American native hunting and management of moose populations. Viltrevy, Swedish Wildlife Research (Supplement 1): 25-42.

Forest Stewardship Council Indigenous Foundation (FSC-IF). 2020. FSC Launches Indigenous Foundation. FSC, Ciudad de Panamá, Panamá. https://www.fscindigenousfoundation.org/fsc-launches-indigenous-foundation/

Government of Canada (GC). 2022. Research Involving the First Nations, Inuit, and Métis Peoples of Canada. Chapter 9 in Canadian Institutes of Health Research, Natural Sciences and Engineering Research Council of Canada, and Social Sciences and Humanities Research Council of Canada, Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans. GC, Ottawa, Ontario, Canada.

Gouvernement de Québec (GQ). 2019. Forêt ouverte. GQ, Québec, Québec, Canada. Forêt ouverte

Governments of Québec and the Cree Nation. 2002. Paix des Braves. Agreement concerning a new relationship between the Government of Québec and the Cree of Québec. Governments of Québec and the Cree Nation, Québec, Canada. https://cdn-contenu.quebec.ca/cdn-contenu/adm/min/conseil-executif/publications-adm/srpni/administratives/ententes/Cris/en/2005-11-02_cris-amendement-2-en.pdf

Gray, S. A., S. Gray, J. L. De Kok, A. E. R. Helfgott, B. O’Dwyer, R. Jordan, and A. Nyaki. 2015. Using fuzzy cognitive mapping as a participatory approach to analyze change, preferred states, and perceived resilience of social-ecological systems. Ecology and Society 20(2):11. https://doi.org/10.5751/ES-07396-200211

Gray, S. A., E. Zanre, and S. R. J. Gray. 2014. Fuzzy cognitive maps as representations of mental models and group beliefs. Pages 29-48 in E. I. Papageorgiou, editor. Fuzzy cognitive maps for applied sciences and engineering. Springer, Berlin, Germany. https://doi.org/10.1007/978-3-642-39739-4_2

Hedelin, B., S. Gray, S. Woehlke, T. K. BenDor, A. Singer, R. Jordan, M. Zellner, P. Giabbanelli, P. Glynn, K. Jenni, A. Jetter, N. Kolagani, B. Laursen, K. M. Leong, L. Schmitt Olabisi, and E. Sterling. 2021. What’s left before participatory modeling can fully support real-world environmental planning processes: a case study review. Environmental Modelling & Software 143:105073. https://doi.org/10.1016/j.envsoft.2021.105073

Hobbs, B. F., S. A. Ludsin, R. L. Knight, P. A. Ryan, J. Biberhofer, and J. J. H. Ciborowski. 2002. Fuzzy cognitive mapping as a tool to define management objectives for complex ecosystems. Ecological Applications 12(5):1548-1565. https://doi.org/10.1890/1051-0761(2002)012[1548:FCMAAT]2.0.CO;2

Huria, T., S. C. Palmer, S. Pitama, L. Beckert, C. Lacey, S. Ewen, and L. T. Smith. 2019. Consolidated criteria for strengthening reporting of health research involving indigenous peoples: the CONSIDER statement. BMC Medical Research Methodology 19(1):173. https://doi.org/10.1186/s12874-019-0815-8

Indigenous Circle of Experts (ICE). 2018. We rise together. Achieving Pathway to Canada Target 1 through the creation of Indigenous Protected and Conserved Areas in the spirit and practice of reconciliation. Indigenous Circle of Experts.

Inuit Tapiriit Kanatami (ITK). 2018. National Inuit strategy on research. ITK, Ottawa, Ontario, Canada.

Jacqmain, H., L. Bélanger, R. Courtois, T. Beckley, C. Dussault, and L. Bouthillier. 2005. Proposal to combine Cree and scientific knowledge for improved moose habitat management on Waswanipi Eeyou Aschee, Northern Québec. Alces 41:147-160.

Jacqmain, H., L. Bélanger, R. Courtois, C. Dussault, T. M. Beckley, M. Pelletier, and S. W. Gull. 2012. Aboriginal forestry: development of a socioecologically relevant moose habitat management process using local Cree and scientific knowledge in Eeyou Istchee. Canadian Journal of Forest Research 42(4):631-641. https://doi.org/10.1139/x2012-020

Jacqmain, H., C. Dussault, R. Courtois, and L. Bélanger. 2008. Moose-habitat relationships: integrating local Cree native knowledge and scientific findings in northern Quebec. Canadian Journal of Forest Research 38(12):3120-3132. https://doi.org/10.1139/X08-128

Jetter, A. J., and K. Kok. 2014. Fuzzy cognitive maps for futures studies—a methodological assessment of concepts and methods. Futures 61:45-57. https://doi.org/10.1016/j.futures.2014.05.002

Johnson, J. T., R. Howitt, G. Cajete, F. Berkes, R. P. Louis, and A. Kliskey. 2016. Weaving Indigenous and sustainability sciences to diversify our methods. Sustainability Science 11(1):1-11. https://doi.org/10.1007/s11625-015-0349-x

Kellner, K. F., R. B. Renken, J. J. Millspaugh, P. A. Porneluzi, A. J. Wolf, D. K. Fantz, R. A. Gitzen, J. Faaborg, S. R. Timm, S. Ehlers, M. L. Buchanan, J. M. Haslerig, A. D. George, and C. T. Rota. 2019. Effects of forest management on vertebrates: synthesizing two decades of data from hardwood forests in Missouri, USA. Ecological Applications 29(8):e01993. https://doi.org/10.1002/eap.1993

Khan, M. S., and M. Quaddus. 2004. Group decision support using fuzzy cognitive maps for causal reasoning. Group Decision and Negotiation 13(5):463-480. https://doi.org/10.1023/B:GRUP.0000045748.89201.f3

Kimmerer, R. 2013. Braiding sweetgrass: Indigenous wisdom, scientific knowledge and the teachings of plants. Milkweed Editions, Minneapolis, Minnesota, USA.

Kosko, B. 1986. Fuzzy cognitive maps. International Journal of Man-Machine Studies 24(1):65-75. https://doi.org/10.1016/S0020-7373(86)80040-2

Kovach, M. E. 2009. Indigenous methodologies: characteristics, conversations, and contexts. University of Toronto Press, Toronto, Ontario, Canada.

Kuuluvainen, T., and S. Gauthier. 2018. Young and old forest in the boreal: critical stages of ecosystem dynamics and management under global change. Forest Ecosystems 5:26. https://doi.org/10.1186/s40663-018-0142-2

Lindenmayer, D. B. 2009. Forest wildlife management and conservation. Annals of the New York Academy of Sciences 1162(1):284-310. https://doi.org/10.1111/j.1749-6632.2009.04148.x

Mataira, P. J. 2019. Transforming Indigenous research: collaborative responses to historical research tensions. International Review of Education 65(1):143-161. https://doi.org/10.1007/s11159-019-09766-5

McGregor, D. 2004. Coming full circle: Indigenous knowledge, environment, and our future. American Indian Quarterly 28:385-410. https://doi.org/10.1353/aiq.2004.0101

McGregor, D. 2018. From ‘decolonized’ to reconciliation research in Canada: drawing from Indigenous research paradigms. ACME: An International Journal for Critical Geographies 17(3):810-831.

McKay, T., and L. Finnegan. 2023. Ungulate occurrence in forest harvest blocks is influenced by forage availability, surrounding habitat and silviculture practices. Ecological Solutions and Evidence 4(2):e12226. https://doi.org/10.1002/2688-8319.12226

McLaren, B. E., S. P. Mahoney, T. S. Porter, and S. M. Oosenbrug. 2000. Spatial and temporal patterns of use by moose of pre-commercially thinned, naturally-regenerating stands of balsam fir in central Newfoundland. Forest Ecology and Management 133:179-196. https://doi.org/10.1016/S0378-1127(99)00223-6

Ministère des Forêts, de la Faune et des Parcs (MFFP). 2022. Aerial survey of moose in hunting zone 17 in winter 2021. MFFP, in collaboration with Cree Nation Government, Cree First Nation of Waswanipi and Cree Trappers’ Association. Government of Québec, Québec, Canada.

Niesink, P., K. Poulin, and M. Šajna. 2013. Computing transitive closure of bipolar weighted digraphs. Discrete Applied Mathematics 161(1-2):217-243. https://doi.org/10.1016/j.dam.2012.06.013

Norström, A. V., C. Cvitanovic, M. F. Löf, S. West, C. Wyborn, P. Balvanera, A. T. Bednarek, E. M. Bennett, R. Biggs, A. de Bremond, B. M. Campbell, J. G. Canadell, S. R. Carpenter, C. Folke, E. A. Fulton, O. Gaffney, S. Gelcich, J.-B. Jouffray, M. Leach, M. Le Tissier, B. Martín-López, E. Louder, M.-F. Loutre, A. M. Meadow, H. Nagendra, D. Payne, G. D. Peterson, B. Reyers, R. Scholes, C. I. Speranza, M. Spierenburg, M. Stafford-Smith, M. Tengö, S. van der Hel, I. van Putten, and H. Österblom. 2020. Principles for knowledge co-production in sustainability research. Nature Sustainability 3(3):182-190. https://doi.org/10.1038/s41893-019-0448-2

Özesmi, U., and S. L. Özesmi. 2004. Ecological models based on people’s knowledge: a multi-step fuzzy cognitive mapping approach. Ecological Modelling 176(1-2):43-64. https://doi.org/10.1016/j.ecolmodel.2003.10.027

Peek, J. M. 1997. Habitat relationships. Pages 351-375 in A.W. Franzmann, C.C. Schwartz, editors. Ecology and management of the North American moose. Smithsonian Institute Press, Washington, D.C., USA.

Ploughe, L. W., and L. H. Fraser. 2022. Find new roads^TM? A systematic review on the impacts of off-road vehicle activity on soil, vegetation, and wildlife. Frontiers in Ecology and Evolution 9:805707. https://doi.org/10.3389/fevo.2021.805707

Poole, K. G., R. Serrouya, and K. Stuart-Smith. 2007. Moose calving strategies in interior montane ecosystems. Journal of Mammalogy 88(1):139-150. https://doi.org/10.1644/06-MAMM-A-127R1.1

Poole, K. G., and K. Stuart-Smith. 2006. Winter habitat selection by female moose in western interior montane forests. Canadian Journal of Zoology 84(12):1823-1832. https://doi.org/10.1139/z06-184

Popp, J. N., P. Priadka, and C. Kozmik. 2019. The rise of moose co-management and integration of Indigenous knowledge. Human Dimensions of Wildlife 24(2):159-167. https://doi.org/10.1080/10871209.2019.1545953

Potvin, F., L. Breton, and R. Courtois. 2005. Response of beaver, moose, and snowshoe hare to clear-cutting in a Quebec boreal forest: a reassessment 10 years after cut. Canadian Journal of Forest Research 35:151-160. https://doi.org/10.1139/x04-151

Potvin, F., R. Courtois, and L. Bélanger. 1999. Short-term response of wildlife to clear-cutting in Quebec boreal forest: multiscale effects and management implications. Canadian Journal of Forest Research 29:1120-1127. https://doi.org/10.1139/x99-040

Reid, A. J., L. E. Eckert, J.-F. Lane, N. Young, S. G. Hinch, C. T. Darimont, S. J. Cooke, N. C. Ban, and A. Marshall. 2021. “Two-Eyed Seeing”: an Indigenous framework to transform fisheries research and management. Fish and Fisheries 22(2):243-261. https://doi.org/10.1111/faf.12516

Rempel, R. S., P. C. Elkie, A. R. Rodgers, and M. J. Gluck. 1997. Timber-management and natural-disturbance effects on moose habitat: landscape evaluation. Journal of Wildlife Management 61(2):517-524. https://doi.org/10.2307/3802610

Rosenblatt, E., K. Gieder, T. Donovan, J. Murdoch, T. P. L. Smith, M. P. Heaton, T. S. Kalbfleisch, B. M. Murdoch, S. Bhattarai, E. Pacht, E. Verbist, V. Basnayake, and S. McKay. 2023. Genetic diversity and connectivity of moose (Alces americanus americanus) in eastern North America. Conservation Genetics 24(2):235-248. https://doi.org/10.1007/s10592-022-01496-w

Sarmiento, I., A. Cockcroft, A. Dion, S. Paredes-Solís, A. De Jesús-García, D. Melendez, A. Marie Chomat, G. Zuluaga, A. Meneses-Rentería, and N. Andersson. 2022. Combining conceptual frameworks on maternal health in Indigenous communities—fuzzy cognitive mapping using participant and operator-independent weighting. Field Methods 34(3):223-239. https://doi.org/10.1177/1525822X211070463

Schipper, J., J. S. Chanson, F. Chiozza, N. A. Cox, M. Hoffmann, V. Katariya, J. Lamoreux, A. S. L. Rodrigues, S. N. Stuart, H. J. Temple, J. Baillie, L. Boitani, T. E. Lacher, R. A. Mittermeier, A. T. Smith, D. Absolon, J. M. Aguiar, G. Amori, N. Bakkour, R. Baldi, R. J. Berridge, J. Bielby, P. A. Black, J. J. Blanc, T. M. Brooks, J. A. Burton, T. M. Butynski, G. Catullo, R. Chapman, Z. Cokeliss, B. Collen, J. Conroy, J. G. Cooke, G. A. B. da Fonseca, A. E. Derocher, H. T. Dublin, J. W. Duckworth, L. Emmons, R. H. Emslie, M. Festa-Bianchet, M. Foster, S. Foster, D. L. Garshelis, C. Gates, M. Gimenez-Dixon, S. Gonzalez, J. F. Gonzalez-Maya, T. C. Good, G. Hammerson, P. S. Hammond, D. Happold, M. Happold, J. Hare, R. B. Harris, C. E. Hawkins, M. Haywood, L. R. Heaney, S. Hedges, K. M. Helgen, C. Hilton-Taylor, S. A. Hussain, N. Ishii, T. A. Jefferson, R. K. B. Jenkins, C. H. Johnston, M. Keith, J. Kingdon, D. H. Knox, K. M. Kovacs, P. Langhammer, K. Leus, R. Lewison, G. Lichtenstein, L. F. Lowry, Z. Macavoy, G. M. Mace, D. P. Mallon, M. Masi, M. W. McKnight, R. A. MedelliÌn, P. Medici, G. Mills, P. D. Moehlman, S. Molur, A. Mora, K. Nowell, J. F. Oates, W. Olech, W. R. L. Oliver, M. Oprea, B. D. Patterson, W. F. Perrin, B. A. Polidoro, C. Pollock, A. Powel, Y. Protas, P. Racey, J. Ragle, P. Ramani, G. Rathbun, R. R. Reeves, S. B. Reilly, J. E. Reynolds, C. Rondinini, R. G. Rosell-Ambal, M. Rulli, A. B. Rylands, S. Savini, C. J. Schank, W. Sechrest, C. Self-Sullivan, A. Shoemaker, C. Sillero-Zubiri, N. De Silva, D. E. Smith, C. Srinivasulu, P. J. Stephenson, N. van Strien, B. K. Talukdar, B. L. Taylor, R. Timmins, D. G. Tirira, M. F. Tognelli, K. Tsytsulina, L. M. Veiga, J.-C. Vieì, E. A. Williamson, S. A. Wyatt, Y. Xie, and B. E. Young. 2008. The status of the world’s land and marine mammals: diversity, threat, and knowledge. Science 322(5899):225-230. https://doi.org/10.1126/science.1165115

Shanley, C. S., and S. Pyare. 2011. Evaluating the road-effect zone on wildlife distribution in a rural landscape. Ecosphere 2(2):16. https://doi.org/10.1890/ES10-00093.1

Smith, L. T. 1999. Decolonizing methodologies: research and indigenous peoples. Zed Books, London, UK.

Stern, E. R. 2022. Wildlife science inclusive of local priorities and knowledge co-production: moose habitat selection in the Adapted Forestry Regime of Eeyou Istchee, Northern Quebec. Thesis. McGill University, Montréal, Québec, Canada.

Stern, E. R., and M. M. Humphries. 2022. Interweaving local, expert, and Indigenous knowledge into quantitative wildlife analyses: a systematic review. Biological Conservation 266:109444. https://doi.org/10.1016/j.biocon.2021.109444

Tanner, A. 2018. Afterword: two world views of Eeyou family hunting territories. Anthropologica 60(1):161-166. https://doi.org/10.3138/anth.60.1.t15

Teitelbaum, S., and C. Lussier. 2016. Consultation of Waswanipi tallymen on amendment No. 6 of the Paix des Braves Agreement. Cree First Nation of Waswanipi, Waswanipi, Québec, Canada.

Tengö, M., E. S. Brondizio, T. Elmqvist, P. Malmer, and M. Spierenburg. 2014. Connecting diverse knowledge systems for enhanced ecosystem governance: the multiple evidence base approach. Ambio 43(5):579-591. https://doi.org/10.1007/s13280-014-0501-3

Thompson, I. D., J. A. Baker, and M. Ter-Mikaelian. 2003. A review of the long-term effects of post-harvest silviculture on vertebrate wildlife, and predictive models, with an emphasis on boreal forests in Ontario, Canada. Forest Ecology and Management 177(1):441-469. https://doi.org/10.1016/S0378-1127(02)00453-X

Thompson, I. D., and R. W. Stewart. 1997. Management of moose habitat. Pages 377-401 in A. W. Franzmann, C. C. Schwartz, editors. Ecology and management of the North American moose. Smithsonian Institute Press, Washington, D.C., USA.

Tyers, D. B. 2003. Winter ecology of moose on the Northern Yellowstone Winter Range. Dissertation. Montana State University, Bozeman, Montana, USA.

Voinov, A., K. Jenni, S. Gray, N. Kolagani, P. D. Glynn, P. Bommel, C. Prell, M. Zellner, M. Paolisso, R. Jordan, E. Sterling, L. Schmitt Olabisi, P. J. Giabbanelli, Z. Sun, C. Le Page, S. Elsawah, T. K. BenDor, K. Hubacek, B. K. Laursen, A. Jetter, L. Basco-Carrera, A. Singer, L. Young, J. Brunacini, and A. Smajgl. 2018. Tools and methods in participatory modeling: selecting the right tool for the job. Environmental Modelling & Software 109:232-255. https://doi.org/10.1016/j.envsoft.2018.08.028

Wong, C., K. Ballegooyen, L. Ignace, M. J. (Gùdia) Johnson, and H. Swanson. 2020. Towards reconciliation: 10 calls to action to natural scientists working in Canada. FACETS 5(1):769-783. https://doi.org/10.1139/facets-2020-0005

Zadeh, L. A. 1988. Fuzzy logic. Computer 21(4):83-93. https://doi.org/10.1109/2.53

Zurba, M., M. A. Petriello, C. Madge, P. McCarney, B. Bishop, S. McBeth, M. Denniston, H. Bodwitch, and M. Bailey. 2022. Learning from knowledge co-production research and practice in the twenty-first century: global lessons and what they mean for collaborative research in Nunatsiavut. Sustainability Science 17(2):449-467. https://doi.org/10.1007/s11625-021-00996-x