Assessment of spring habitat selection of red deer (Cervus elapbus L.) based on census data

Data from 11 years of spring census over a 6,325 ha area were used to estimate the habitat preferences of red deer in the West Hertogenwald forest (Ardenne, Belgium). Each year, 1 to 4 censuses were organised to assess deer population density. The census method was based on a systematic coverage of the forest by experienced spotters. Each observation was described (number, sex, behaviour) and located in a GIS. Within the study area, 915 groups of non-antlered deer (hind, yearling and calf), corresponding to 2,775 individuals, were recorded during the 24 censuses of the period from 1990 to 2000. Different environmental variables describing the study area, such as forest cover (tree species and structure), canopy closure, altitude, distance to roads and pastures, were identified and integrated into the GIS. Each observation of a group of non-antlered deer was spatially related to these environmental variables to preform a habitat selection analysis. The results of this analysis showed significant preferences for certain forest types (eg spruce vs beech), and for open canopy areas. Positive correlations to high altitude areas, pastures and spruce thicket stages and a negative correlation to roads were highlighted. A comparison procedure involving GPS-collared animals in the same area was performed and showed a good correlation between census and GPS data. The methodology of this approach and the management implications of the results are discussed