Premium
Balancing Number of Locations with Number of Individuals in Telemetry Studies
Author(s) -
GIRARD IRÈNE,
DUSSAULT CHRISTIAN,
OUELLET JEANPIERRE,
COURTOIS RÉHAUME,
CARON ALAIN
Publication year - 2006
Publication title -
the journal of wildlife management
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.94
H-Index - 111
eISSN - 1937-2817
pISSN - 0022-541X
DOI - 10.2193/0022-541x(2006)70[1249:bnolwn]2.0.co;2
Subject(s) - telemetry , habitat , selection (genetic algorithm) , ungulate , range (aeronautics) , home range , global positioning system , sample (material) , sample size determination , tracking (education) , environmental science , ecology , sampling (signal processing) , statistics , computer science , biology , mathematics , telecommunications , engineering , chromatography , artificial intelligence , detector , aerospace engineering , psychology , pedagogy , chemistry
The study of habitat selection usually compares assessments of habitat use to habitat availability. To investigate habitat selection of large mammals today, researchers must choose between a few very expensive Global Positioning System (GPS) telemetry collars that can provide many locations and several inexpensive very high frequency telemetry collars that will provide few numbers of locations (unless substantial resources are spent in the field). We investigated the effects of number of locations and sampled animals on the outcome of habitat‐selection analyses. We evaluated whether tracking frequency and sample size of individuals influenced our ability to detect habitat selection. We used data obtained from adult female moose fitted with GPS collars to generate data sets simulating various sampling frequencies and sample sizes of individuals. Tracking schedules conformed to those commonly used in ungulate telemetry studies (1 location every 14, 7, or 3 d and 1 or 3 locations per d) as did animal sample sizes (between 8 and 20 individuals). We determined habitat use and availability at the landscape and home‐range scales during summer–autumn and winter. Precision of habitat use and availability estimates did not improve markedly with increasing tracking frequency. Only results obtained with the least‐intensive tracking schedule (1 location every 14 d) differed from those obtained with the other schedules and only in 25% of the cases. Above this threshold in tracking frequency, number of sampled animals was clearly more important than number of locations in detection of habitat selection. Our results indicated that habitat‐selection analyses were more sensitive to inter‐ than intra‐individual variability. Depending on study objectives, it may be more profitable to prioritize number of sampled individuals rather than number of locations per individual. We suggest methods allowing researchers to assess inter‐individual variability while studying habitat selection.