Anthropogenic disturbance and trophic interactions mediate population viability of American martens

Anthropogenic disturbance can influence population viability through multiple direct and indirect pathways. Robust management of such systems requires a synthetic modeling approach to compare the impact of multiple sources of variation on vital rates. We derived a population viability model for the American marten ( Martes americana ) based on field studies conducted over 5 years in uncut mature forest landscapes and forest landscapes regenerating from commercial logging replicated at 2 different study locations in northern Ontario, Canada, in 2001–2007. Our central goal was to evaluate the net impact of fur trapping and forest disturbance due to commercial logging on the demography and long‐term viability of the American marten. We found no consistent difference in prey abundance between logged versus uncut landscapes, but all sites exhibited substantial year‐to‐year variation in prey abundance. Mean population density of martens was positively related to population density of their most common prey species, southern red‐backed voles ( Myodes gapperi ). Marten population density was higher and body condition was better in uncut than in logged landscapes. Mortality due to commercial fur trapping averaged 16% per year, compared to 12% per year due to other causes. Loglinear regression models suggested that exponential rate of change in marten populations each year was positively related to mean prey abundance but negatively related to marten population density. There was strong evidence that disturbance due to commercial forest operations 30–60 years previously had a negative effect on mean marten abundance but only marginal evidence that it influenced the population growth rate of marten. Our data suggest that marten populations are shaped by a mixture of density‐dependent, bottom‐up, and top‐down (via fur trapping) trophic interactions. Population viability models parameterized with observed prey‐ and density‐dependent demographic rates suggested that there should be negligible risk of collapse by moderately harvested (<16%) marten populations, but the risk of collapse should rise substantially if fur‐trapping mortality exceeds 25% of the population per year. This risk should be appreciably higher in forest landscapes regenerating from logging than in uncut mature landscapes.