Statement of Inclusion
Our camera trap study was conceived by members of the Whitefish Lake First Nation (WLFN), and our work brings together authors from the First Nation, where data were collected, with US and Canada-based scientists. All authors were engaged throughout the research and study design to ensure that their diverse perspectives were considered from the onset. Elders from WLFN were also consulted for their expertise and perspectives on moose ecology and changes in habitat use before and during the writing process.
Abstract
Subsistence hunting, or “country food,” is essential for Indigenous Peoples who face high food insecurity and is critical for Indigenous Food Sovereignty. For many First Nations of Canada, subsistence hunting is also inextricably linked to traditional conservation practices, as hunting is an important way of engaging with nature. In the boreal of Canada, large game such as moose (Alces alces) are a primary source of protein for many First Nations. However, resource extraction, including forestry practices and oil and gas extraction, has shifted large game distributions and affected the availability and abundance of food resources. Here, we used remote camera trap data and generalized linear models to evaluate moose habitat use and spatial-numerical response to possible stressors in north-central Alberta, including fire, harvest, oil and gas extraction, and other disturbances. We also examined the effects of human-caused stressors on habitat use by sex and age class data. The proportion of various land cover types and human land use for resource extraction were important in moose habitat use. Overall, adult moose avoided burned areas and grasslands. Notably, male, female, and young moose all used habitat differently and at different spatial scales. However, young moose (with their mothers) strongly selected natural forest disturbances such as burned areas but avoided human-created disturbances such as petroleum exploration “seismic” lines. Female moose with young attempting to maximize forage opportunities do not use human-disturbed forests in the same ways they use naturally disturbed areas. This also aligns with observations from Indigenous communities, which have linked human disturbance to declines in moose densities and displacement from traditional hunting grounds. Understanding and predicting shifts in large game distributions is critical to supporting Indigenous Food Sovereignty and identifying where industries operating on First Nations lands can better engage responsibly with First Nations.
Keywords
Traditional ecological knowledge, moose, subsistence hunting, resource extraction, country food
Introduction
Indigenous communities globally have relied on subsistence hunting of local species since time immemorial. However, communities most reliant on local ecosystem services, such as subsistence hunting, are the most vulnerable to threats associated with biodiversity and species loss (Díaz et al., 2006). Subsistence hunting sometimes exacerbates the deleterious effects of human resource use on local systems and species (Luz et al., 2017), particularly in regions where hunting is non-selective, or species are already at high risk due to habitat loss, climate change, or pollution (Lindsey et al., 2013; Ripple et al., 2016; Theriault, 2011). However, the assumption that subsistence hunters solely maximizing harvest in the short-term rather than balancing foraging with conservation (i.e., considering the long-term benefits of sustainability) is incorrect (Alvard, 1994; Bodmer et al., 2020; VanStone, 1974).
In North America, subsistence hunting (also referred to as “country food”) is closely linked with Indigenous conservation practices (Feit, 1973; Gottesfeld, 1994). These conservation practices range from limits on the number of individuals harvested to seasonal rotations of hunting grounds. In many parts of North America, ethical subsistence hunting (as determined by local Indigenous communities) is essential for food security (Theriault, 2011), supports Indigenous Food Sovereignty (Cidro et al., 2015), and has additional social, cultural, and spiritual importance (Van Oostdam et al., 2005). Across much of North America, Indigenous harvesting is declining despite the importance of subsistence hunting for Indigenous communities (Gilbert et al., 2021; Shafiee et al., 2022), in part due to cost (i.e., permits, equipment) and concerns about environmental contaminants in hunted food (Skinner et al., 2013), including cadmium, lead, arsenic, mercury, methylmercury, and other persistent organic pollutants (Chan et al., 2021). Many Indigenous communities have expressed resignation at the continued loss of their subsistence landbase (Westman & Joly, 2019).
Industrial resource extraction has resulted in rapid changes in the densities, distributions, and communities of traditionally hunted species across Canada. Energy development, specifically oil and gas extraction, is one of the primary causes of the decline of woodland caribou (Rangifer tarandus; atihk in Cree) across western Canada, (Hebblewhite, 2017). Some species, such as wolves and bears, benefit from and consistently use anthropogenically-created landscape features in Western Canada (Dickie et al., 2020; Dickie et al., 2017), increasing hunting efficiency (McKenzie et al., 2012). However, not all species benefit from these features, and many more actively avoid them (Fisher & Burton, 2018). Increases in predator population size due to landscape development (Latham et al., 2011) and the high prey-kill rates associated with anthropogenic features (Boucher et al., 2022) impact traditionally hunted species across Canada, leading to additional pressures on country food.
Moose (Alces alces; moswa in Cree) are an important, but declining, subsistence resource for the First Nations of Canada (Kuzyk et al., 2018; Natcher et al., 2021; Priadka et al., 2022; Ross & Mason, 2020), and there is widespread recognition that resource extraction impacts moose population dynamics, distributions, and predation rates in Alberta (Lamy & Finnegan, 2019; Neilson & Boutin, 2017). In the boreal, moose select for habitat that provides security when predator abundance is high (Ethier et al., 2024), and have lower occurrence in areas with pipelines, seismic lines, 3D seismic lines, unpaved roads, and new cutblocks (Dickie et al., 2022; Finnegan et al., 2023; Fisher, Grey, Anderson, Sawan, Anderson, Chai, Nolan, Underwood, Maddison, et al., 2021; McKay & Finnegan, 2023; McKay & Finnegan, 2022), which are often used by predators. This suggests that perceived predation risk is a strong driver of habitat selection, especially in areas with high human use. There are some potential benefits of human land use for moose, as forest cutblocks offer increased moose forage (Francis et al., 2021; Johnson & Rea, 2023). However, the effects of herbicide treatment and predation risks in these cutblocks might outweigh the benefits, as moose in high herbicide-use areas consume fewer forbs (Koetke et al., 2023). Finally, resource roads and trails associated with forestry and petroleum open access to previously remote areas, facilitating poaching, as has been observed by First Nation communities. Overall, the degree to which moose abundance and habitat use have shifted across traditional territories with increasing human land use is unclear, partly due to the challenges of monitoring moose populations in traditional hunting areas.
There are many inherent challenges in assessing moose abundance and distribution changes, primarily those associated with animal detection. While aerial surveys are often used in moose monitoring (Moll et al., 2022), these methods are challenging for Indigenous communities to employ (e.g., cost, time, and human safety risks associated with helicopters (Jones IV et al., 2006; Watts et al., 2010). Instead, remote cameras (also known as “camera traps” or “trail cameras”) offer an effective means of sampling mammal populations when appropriately employed (Burton et al., 2015). Furthermore, camera trap studies are an effective approach for the coproduction of knowledge and Indigenous-led or co-created research (Fisher et al., 2021).
We sought to quantify the effects of human disturbance, fire, and land cover on moose relative abundance and spatial distribution using remote camera data to inform First Nation subsistence hunting. Specifically, our goal was to determine the relative impacts of forest harvest, linear features (e.g., roads and pipelines), oil and gas extraction sites, forest cover types, and age of burned areas on moose distribution. We also sought to compare the relative effects of these features by age and sex, as male and female moose (with and without young) may select different features at different scales when balancing predation risk with forage availability. We generally expected moose with higher nutritional needs (e.g., females with young) might use “riskier” (open) habitat when high-quality forage is available and that all moose would be strongly associated with aquatic features due to their dietary needs (Fraser et al., 1984). We also expected differences in habitat use by sex and age, primarily between young moose (young of the year and young of last year) and males due to different dietary and safety needs. We also expected that the spatial scales at which landscape features explain moose distribution (Holland & Yang, 2016) might differ by sex and age, as male moose may use broader areas, balancing foraging and seeking mates. In contrast, cows with young moose would likely be driven by the need for high-quality forage to support nursing young and offspring growth.
Methods
Study area
Our study area encompassed the Whitefish Lake First Nation (WLFN) traditional territory, a Treaty 8 Territory (Fumoleau, 2004) in north-central Alberta, Canada (Figure 1), characterized by expansive central mixedwood forests interspersed with many small lakes, bogs, wetlands, fescue grasslands, both open and closed conifer stands, and closed shrublands (AMBI 2020). The Indigenous co-authors have been living on this land for millennia, relying on its resources to survive in this cold and relatively nutrient-poor boreal landscape. Recent industrial modification in the form of forestry and oil and gas extraction – which we refer to as “anthropogenic landscape features” and “anthropogenic disturbance” – differs vastly from traditional stewardship techniques and is abundant across the landscape (Figure 1). The area has experienced forest harvest for a few decades. Harvested conifer stands are typically replanted and treated with glyphosate (N-(phosphonomethyl) glycine) via helicopter, resulting in notable changes to plant communities and resources used by the First Nation. Widespread petroleum extraction is a more recent and even more widespread and diverse disturbance (Pickell et al. 2013, Pickell et al. 2014, Pickell et al. 2015). WLFN Elders note that the drastic human-induced landscape changes on their Territory have resulted in precipitous declines of many important mammal and plant species. They also note that, like many other parts of Canada, increasing fire frequency is ‘extreme’ (Gaboriau et al. 2022), with new burn records being set in recent years (CIFFC, 2023; Canadian National Fire Database, 2023; but see Chavardès et al. 2022).
Camera Trapping
WLFN co-authors designed the research study, sampling sites were assigned based on a constrained random stratified design. The landscape was divided into four strata based on dominant canopy cover and hydrological conditions, and sites were randomly selected from these strata, with some constraints based on the logistics of access. Community members deployed 130 ReconyxTM Hyperfire 2 (Holmen, WI, USA) cameras between 2018 and 2023 (Figure 1). Of these cameras, 75 were deployed and active from December 2018 to April/May 2019, 25 more were added in March, and all 100 were active between March and November 2019. WLFN deployed an additional 30 cameras, which were active from June 2022 to July 2023. Cameras were placed ca. 1.5 m above ground at sampling sites. Sampling sites were active wildlife trails and camera sensors were set to “high sensitivity” to record one image with each heat-in-motion detection, with no programmed delays between photographs, adopting techniques used in Fisher and Burton (2018). WLFN staff and volunteers classified images to species by using TimeLapse2 Image Analysis software ( Greenberg et al., 2019). Of the deployed cameras, timelapse data and images were retrievable from 121 cameras (96 of the original 100 and 21 of the subsequent 30). Images were grouped across sampling periods for our analysis to ensure naïve occupancy was sufficiently large for meaningful results. Images were also categorized as male, female, young of the year (YOY), or young of last year (YLY) whenever possible.