| I studied habitat selection of female white-tailed deer ( Odocoileus virginianus) at 3 spatial scales: landscape, meso-scale, and plots. Results were generally robust to differences in the 3 vegetation maps used in analysis, South Dakota GAP, National Landcover Dataset (NLCD), and Black Hills National Forest, despite poor agreement among them. I attribute this agreement in deer habitat selection among vegetation maps to vegetation structure being more important to white-tailed deer habitat selection than species composition. At the meso-scale, Normalized Difference Vegetation Index (NDVI) and topographic variables were more important than specific vegetation cover types. While specific plant species were not tested in the plot level analysis, variables like number of tall shrubs and saplings, average diameter breast height (DBH), and average basal area were significant. Female deer selected productive habitats year round. However, grass, forb, and shrub cover was less at deer locations than random sites. At the landscape scale, interspersion and juxtaposition of cover types was important. The percentage of the home range in grassland was also important. At the meso scale, vegetation structure was important. Ponderosa pine (Pinus ponderosa) appeared to be used as thermal cover in winter to compensate for poor forage availability. Preference by deer for areas with high productivity but with low coverage by forbs and shrubs suggested that competition by elk (Cervus elaphus ) and cattle was important. Furthermore, deer apparently avoided areas of human disturbance adjacent to roads and houses. At the plot scale, thickets were important to deer. Utility of thickets was reduced by proximity to roads. Deer bedded adjacent to large trees surrounded by smaller trees. Knowledge of habitat selection at all 3 scales (landscape, meso, and plot) was necessary to effectively manage white-tailed deer in the central Black Hills. |
