Evaluating vertebrate species risk on an industrial forest landscape

Abstract

We describe a new landscape-specific procedure for ranking vertebrate forest wildlife species for their relative risk of extirpation, and apply it to a ca. 189,000 ha intensively managed industrial-forest landscape in northwestern New Brunswick, Canada (J.D. Irving, Ltd., Black Brook District). A list of 157 vertebrate species that potentially occur on the landbase was compiled based on published range maps and expert consultation. Each species was assessed for risk of extirpation using a categorical ranking system termed the ‘Species-Sorting Algorithm’ (SSA). The SSA places each species into one of four relative risk categories (Class I = ‘potential high risk or management concern’, II, III and Unknown), based on four variables: (1) potential abundance, (2) proportion of the landscape suitable for occupancy, (3) species-specific habitat connectivity, and (4) population growth potential. These variables use a mix of spatial and non-spatial data derived from a spatial analysis of the landscape and a review of published life-history parameters. A total of 26 (17%) species were rated as Class I, 14 of which were due at least partly to low habitat availability. Eight species were Class I due to Provincial or National ‘at risk’ listing, but only two of these would have been Class I based on the four SSA variables alone. An assessment of the robustness of the ranking variables indicated that the percentage of suitable landscape had the largest effect on SSA results, and a logistic regression model showed estimated provincial abundance and body size to be significant predictors of Class I categorization. The SSA is dependent upon stand structure-based species–habitat relationships and species area–sensitivity relationships, which were uncertain for some species. The SSA was useful in highlighting data gaps and in providing a framework for incorporating new knowledge.