GPS approach to study fine-scale site use by wild red deer during active and inactive behaviors

Studying fine-scale habitat selection of wild mammals relative to environmental parameters requires 1) correctly sampling the species' activity in time and space and 2) measuring the environmental variables at those locations. We propose a procedure using Differential Global Positioning System (DGPS) technology to achieve these goals. A data set from red deer (Cervus elaphus) was used to demonstrate the method. We validated an individual-based relative method using the count provided by a GPS collar activity sensor to separate active from inactive behaviors. This method allowed the correct classification of locations as active or inactive (82-97% of correct allocations depending on the fix interval used). We analyzed activity sampling using various fix intervals (5, 10, and 15 min) and the resulting spatial sampling when we retained only highly accurate locations (<10-m error radius). We demonstrated that long intervals between fixes could cause a loss of information for the shortest activity bouts, which make up 5 and 11.5% of the daily budget using 10- and 1 5-min fix intervals, respectively. The proposed spatial sampling allowed 65 to 73% of the total fixes to be kept. Active bouts were slightly undersampled, whereas inactive bouts were oversampled. We used a subsample of animal locations that were relocated using hand-held DGPS to measure several environmental variables in situ and examined the variations relative to animal and activity combinations.