The difficulty of using species distribution modelling for the conservation of refugee species - the example of European bison

Refugee species have been confined to suboptimal habitat through historic anthropogenic factors. If this is unknown, management might actively conserve these species in suboptimal habitat assuming it represents optimal habitat. Similarly, species distribution modelling (SDM) might misguide conservation management of refugee species by only using presence data from suboptimal habitats. We illustrate this by commenting on a recent SDM for European bison that reconstructed the historic distribution of the species. We challenge the interpretation of this model by suggesting an alternative historic biogeography based on the refugee species concept. We argue that, in the case of refugee species, historic reconstructions using SDM cannot be used as a template for conservation management. Rather, experimental re-introduction programmes should provide us with population performance and life history data from a range of suboptimal to optimal habitats. Such data could be used in mechanistic niche modelling to predict potential distribution of refugee species.