Epiphyte sensitivity to a cross-scale interaction between habitat quality and macroclimate: an opportunity for range-edge conservation

Bioclimatic envelope models are frequently used to project the species response to climate change scenarios. Development and improvement of bioclimatic models has focussed on data properties and statistical tools, while significant criticism continues to challenge the ecological framework of model assumptions. We hypothesised that a potential for model improvement emerges from linkage across scales, between macroclimate and variation in local habitat quality: i.e. a species’ habitat specificity may shift along macroclimatic gradients. We first sampled two test-case epiphytic lichen species across a steep climatic gradient, and second developed standard bioclimatic models accompanied by a threshold likelihood value for discriminating presences and absences. We used the difference between predicted model values and the threshold as a response variable (D _thr): we show that values for D _thr are explained by an interaction between the climatic setting and habitat quality. A potential error in bioclimatic models is then quantified as the region of false absences or presences, which would be incurred as a consequence of sensitivity to variable habitat. This signature habitat effect occurs at a species’ range-edge, and, as a corollary, provides quantification in support of conservation: i.e. information is provided on how a habitat may be managed in marginal climatic regions (leading or trailing range-edge boundaries) in order to promote species protection.