Abstract
Mountain caribou, a behaviourally and genetically distinct set of ecotypes of the Woodland caribou (Rangifer tarandus caribou) restricted to the mountains of western Canada, have undergone severe population declines in recent decades. Although a broad consensus exists that the ultimate driver of these declines has been the reduction of habitat upon which mountain caribou depend, research and policy attention has increasingly focused on predation. Recently, Serrouya et al. (Proc Nat Acad Sci USA 116:6181–6186, 2019) analysed population dynamics data from 18 subpopulations in British Columbia and Alberta, Canada, subject to different treatments and ‘controls’, and concluded that lethal wolf control and maternal caribou penning provide the most effective ways to stabilize population declines. Here we show that this inference was based on an unbalanced analytical approach that omitted a null scenario, excluded potentially confounding variables and employed irreproducible habitat alteration metrics. Our reanalysis of available data shows that ecotype identity is a better predictor of population trends than any adaptive management treatments considered by Serrouya et al. Disparate behavioural characteristics and responses to industrial disturbance among ecotypes suggest it may be incorrect to assume that adaptive management strategies that might benefit one ecotype are transferable to another.