Modelling space use and dispersal of mammals in real landscapes: A tool for conservation

Aim: To explore the usefulness of Spatially Explicit Population Models (SEPMs), incorporating dispersal, as tools for animal conservation, as illustrated by the contrasting cases of four British mammals. Methods: For each of the four species (American mink, Mustela vison, pine marten, Martes martes, dormouse, Muscardinus avellanarius and water vole, Arvicola terrestris) a spatial dynamics model was developed based on an integrated geographical information system (GIS) population model that linked space use to the incidence of the species. Each model had, first, a GIS, which stored environmental, habitat and animal population information, and secondly, an individual-based population dynamics module, which simulated home range formation, individual life histories and dispersal within the GIS-held landscape. Results: The four models illustrated different interactions between species life-history variables and the landscape, particularly with respect to dispersal. As water voles and dormice occupy home ranges that are small relative to blocks of their habitat, they were most effectively modelled in terms of the dynamics of local populations within habitat blocks but linked by dispersal. In contrast, because the home ranges of American mink and pine marten are large relative to blocks of habitat, they were best modelled as individuals moving through a landscape of more or less useful patches of habitat. For the water vole, the most significant predictors of population size were the carrying capacity of each habitat and the annual number of litters. For the dormouse, the likelihood of catastrophe and the upper limit to dispersal movement were the key variables determining persistence. Adult mortality and home-range size were the only significant partial correlates of total population size for the American mink. Adult mortality was also a significant correlate of total population size in the pine marten, as were litter size and juvenile mortality. In neither the marten nor the mink was dispersal distance a significant factor in determining their persistence in the landscape. Main conclusions: At a landscape scale it is difficult to measure animal distributions directly and yet conservation planning often necessitates knowledge of where, and in what numbers, animals are found, and how their distributions will be affected by interventions. SEPMs offer a useful tool for predicting this, and for refining conservation plans before irreversible decisions are taken in practice.