ASsessing sexual segregation in deer

Sexual segregation in temperate and arctic ruminants is defined as the differential use of space by the sexes outside the mating season. This phenomenon is widespread among taxa, and is especially prevalent among sexually dimorphic deer (Cervidae). Understanding how different genders are distributed across the landscape and how to assess these spatial patterns is of theoretical and applied importance. We developed a simple model to evaluate effects of landscape grain (i.e., patch characteristics), sampling scale, and population density of deer on detection of sexual segregation. We created landscape maps of 2 areas in which landscape grain was changed while other landscape metrics (e.g., area, shape, and diversity) were held relatively constant. We created a high-density population of deer to emulate conditions near ecological carrying capacity (K), and a low-density population at <K/2. Sexes of deer were assigned to 4 habitats based on differences in habitat selection derived from an ideal-free distribution, which created spatial separation of the sexes similar to observations in empirical studies. We sampled this pattern of sexual segregation for both areas at large and small scales of measurement using quadrats arranged systematically We also compared the degree of sexual segregation for the coarse-grained. landscape, where the sexes used habitats differently, with a null model in which habitat preferences were identical for each gender of deer. The null model emulated conditions during rut when the sexes were aggregated. Sexual segregation was significantly greater where habitat use differed between sexes, indicating that our model was correct and that the degree of spatial segregation was not an artifact of patch configuration, sampling scale, or population density. Logistic regression revealed that population density and the size of the sample unit significantly affected our ability to assess differences in the spatial distributions of male and female deer where differential use of habitats occurred. Variation in landscape grain, however, did not influence the detection of sexual segregation. Results from our model emphasize the importance of evaluating effects of population density and especially sampling scale on assessing spatial separation of the sexes. Failure to do so may result in not recognizing patterns of sexual segregation on the landscape, or in misinterpreting that phenomenon, which clearly holds consequences for those managing large herbivores or their habitats.