Risk of extirpation for vertebrate species on an industrial forest in New Brunswick, Canada: 1945, 2002, and 2027

Abstract

The risk of extirpation was assessed for 157 vertebrate species for a ca. 190 000 ha forest in New Brunswick, Canada, based on land cover in 1945, 2002, and 2027. Data from 1945, prior to intensive forest management, were derived from detailed spatial 1944–1947 timber-cruise data and maps. Extirpation risk was determined by species, using a categorical system called the species-sorting algorithm whereby each species was assigned to one of four risk classes based on four variables: potential abundance, proportion of the landscape suitable for occupancy, species-specific habitat connectivity, and population growth potential. Data for these variables were derived from species-specific spatial landscape assessments and published life-history parameters. Forest management from 1945 to 2002 decreased the mixed hardwood–softwood forest area from 37% to 18%, increased the area of tolerant hardwoods from 10% to 25%, and decreased the area of forest >70 years old from 86% to 44%. Projections for 2027 showed further declines in old softwood, hardwood, and mixedwood habitats. Twenty-seven vertebrate species were ranked as class I (highest extirpation risk) in 1945 versus 20 in 2002 and 26 in 2027; 35 species (22%) were ranked as class I at least once and 15 species in all 3 years. Sensitivity analyses demonstrated that habitat availability was the most important ranking variable for determining extirpation risk, and that changes in habitat threshold values for assigning risk scores significantly altered results. The forest was less sensitive to habitat thresholds in 1945 than in 2002 or 2027 because of greater homogeneity. Low cover of old-forest habitat, especially mixedwood in large patches with adequate connectivity, resulted from both management and natural disturbances, and was the primary factor determining extirpation risk for vertebrates on the landscape.