PLoS ONE (Jan 2013)

Explaining geographic gradients in winter selection of landscapes by boreal caribou with implications under global changes in Eastern Canada.

  • Julien Beguin,
  • Eliot J B McIntire,
  • Daniel Fortin,
  • Steven G Cumming,
  • Frédéric Raulier,
  • Pierre Racine,
  • Claude Dussault

DOI
https://doi.org/10.1371/journal.pone.0078510
Journal volume & issue
Vol. 8, no. 10
p. e78510

Abstract

Read online

Many animal species exhibit broad-scale latitudinal or longitudinal gradients in their response to biotic and abiotic components of their habitat. Although knowing the underlying mechanism of these patterns can be critical to the development of sound measures for the preservation or recovery of endangered species, few studies have yet identified which processes drive the existence of geographical gradients in habitat selection. Using extensive spatial data of broad latitudinal and longitudinal extent, we tested three hypotheses that could explain the presence of geographical gradients in landscape selection of the endangered boreal woodland caribou (Rangifer tarandus caribou) during winter in Eastern Canadian boreal forests: 1) climate-driven selection, which postulates that geographic gradients are surrogates for climatic gradients; 2) road-driven selection, which proposes that boreal caribou adjust their selection for certain habitat classes as a function of proximity to roads; and 3) an additive effect of both roads and climate. Our data strongly supported road-driven selection over climate influences. Thus, direct human alteration of landscapes drives boreal caribou distribution and should likely remain so until the climate changes sufficiently from present conditions. Boreal caribou avoided logged areas two-fold more strongly than burnt areas. Limiting the spread of road networks and accounting for the uneven impact of logging compared to wildfire should therefore be integral parts of any habitat management plan and conservation measures within the range of the endangered boreal caribou. The use of hierarchical spatial models allowed us to explore the distribution of spatially-structured errors in our models, which in turn provided valuable insights for generating alternative hypotheses about processes responsible for boreal caribou distribution.