Boreal small mammals show evidence of density-dependent patterns with area-sensitivity

Abstract

In recent years, habitat amount in fragmented landscapes has been shown to positively influence population size, species occurrence, and species diversity. Quantifying how sensitive bioindicator species respond to the amount of habitat in disturbed landscapes (i.e. area-sensitivity) has become a growing research focus to provide robust guidelines for ecosystem-based management. In this study, we modelled the occurrence of North American boreal small mammals in relation with the total amount of forest surrounding remnant forest patches in disturbed landscapes while controlling for local habitat associations. Over the summers of 2013 and 2014, we conducted four trapping sessions in 60 sites located in old forest remnant patches of old forests in both wildfires and aggregated clearcuts, and in continuous old forest blocks within the black spruce forest of northwestern Quebec, Canada. American red squirrels (Tamiasciurus hudsonicus), southern red-backed voles (Myodes gapperi) and masked shrews (Sorex cinereus) represented 85.5% of our total captures. We measured the amount of habitat (percentage of forest cover) within 100 concentric buffers around each capture grid using digital forest cover maps. Buffers varied in radii from 50m to 5km. We quantified area-sensitivity using dynamic models of single-species occupancy to estimate the probabilities of initial site occupancy, site extinction and site colonisation of each species according to both local habitat variables and surrounding habitat amount. We found no associations between initial site occupancy, site colonisation, or site extinction with local habitat features, possibly in response to habitat structure similarity of our three site types. Species studied had different life histories in terms of population dynamics' and timing of juvenile dispersal, possibly explaining why each species had its individual response to the amount of habitat in the surrounding landscape. For the American red squirrel, we found no evidence of within-year area-sensitivity on initial site occupancy patterns, whereas negative area sensitivity on initial site occupancy between-years was observed for the southern red-backed vole. In contrast, we found positive area-sensitivity on between-years site colonisation for this latter species at small spatial scales. For masked shrews, we detected negative area-sensitivity on initial site occupancy within-year. As populations were sampled at low density, we suspect that the sparse distribution of individuals may influence area-sensitivity patterns. Future studies should consider area-sensitivity with regards to both spatial and temporal scales. We encourage long-term monitoring of animal populations at multiple spatial scales to investigate the underlying ecological mechanisms of positive and negative area-sensitivity.