: Recent discoveries of diamondiferous kimberlite deposits in the Canadian central Arctic led to unprecedented levels of mineral exploration and development. The cumulative effects of such activities are an issue of concern for government regulatory agencies, regional and international conservation organizations, wildlife managers, and indigenous peoples. We investigated the impacts of human activities and associated infrastructure on the distribution of Arctic wildlife in 190,000 km2 of the Taiga Shield and Southern Arctic ecozones 400 km northeast of Yellowknife, Northwest Territories, Canada. We used covariates for vegetation, interspecific interactions, and human disturbance features to develop seasonal resource-selection models for barren-ground caribou (Rangifer tarandus groenlandicus), gray wolves (Canis lupus), grizzly bears (Ursus arctos), and wolverines (Gulo gulo). We used an information-theoretic approach to select 11 seasonal models for the 4 species. Nine models were good predictors of species occurrence and vegetation covariates were important components of all models. Mines and other major developments had the largest negative affect on species occurrence, followed by exploration activities, and outfitter camps. We did not, however, record strong avoidance responses by all species during all seasons to each disturbance type (i.e., major developments, mineral exploration sites, outfitter camps) and for some models carnivores selected for disturbance features (i.e., occurred closer to sites than comparison random locations). We used a geographic information system (GIS) to extrapolate each seasonal resource-selection model to the study area and quantified the reduction in habitat effectiveness as a function of modeled and hypothetical disturbance coefficients. Across all models, grizzly bears and wolves demonstrated the strongest negative response to disturbance and corresponding reduction in habitat effectiveness, followed by caribou and wolverines. The largest seasonal effect was recorded for caribou during the post-calving period, where model coefficients suggested a 37% reduction in the area of the highest quality habitats and an 84% increase in the area of the lowest quality habitats. This is the first study to demonstrate the cumulative effects of multiple sources of human disturbance for caribou, wolves, bears, and wolverines found across the Canadian central Arctic. Resource selection models and corresponding maps of important habitats can be used to guide and evaluate future development proposals and can serve as a component of a regional environmental assessment. However, inferences from large-scale modeling efforts should be carefully evaluated when making detailed prescriptive recommendations. Study design, sample size, reliability of GIS data, and accuracy of model predictions are important considerations when evaluating the strength and scale of inference of correlative resource selection studies such as this. We recommend that regional cumulative effects analyses serve as the coarsest framework for understanding the impacts of human developments on wide-ranging animals. Monitoring and research should be conducted at various behavioral scales leading to a body of knowledge that fully describes the range and strength of impacts resulting from cumulative effects.
Read full abstract