Evaluation of GPS telemetry collar performance for habitat studies in the boreal forest

Telemetry collars using Global Positioning System (GPS) technology may provide biased locations, and their capability under field conditions should be evaluated before the units are used in habitat studies. We assessed influence of vegetation features, terrain slope, and weather conditions on performance of GPS collars. We used a GPS collar that attempted 5 fixes, separated by 5-minute intervals, in 45 sample sites equally distributed among 3 forest types (deciduous, mixed, and coniferous stands) and 3 age classes (10, 30-50, and greater than or equal to 70 years old). We visited all sites during late September and mid-December. Mean location success was significantly less in fall (69.3+/-5.3%) than in winter (89.8+/-3.0%, P<0.05). Mean number of satellites tracked by the CPS antenna was related negatively to tree height. Location success within tree-height classes (<10 m, 10-13 m, greater than or equal to 13 m) did not differ among habitat types by season. Results from free-ranging moose (Alces alces) fitted with GPS collars indicated that location success varied significantly among months, being least from June to September and greatest during March and April. Location success differed between day and night during summer, which could be explained by changes in habitat use. These results indicated that CPS collars may underestimate use of mature stands relative to other habitat types. This potential bias must be addressed when analyzing frequency distributions of habitats used at location points.