Northern Prairie Wildlife Research Center
Kevin M. Podruzny and Jay J. Rotella, Department of Biology, Montana State University, Bozeman, MT 59717 USA
Radiotelemetry triangulation is routinely used to estimate locations of animals for determining habitat use. However, radiotelemetry locations have associated error and only provide estimates of actual locations. Therefore, we compared MLE point-estimates and 95% error ellipses by examining their effects on classification of habitats, accuracy of observed habitat-use patterns, and efficiency of tests for detecting habitat selection. We then evaluated effects of precision of location estimates (error-ellipse size) on correct classification of habitats, observed habitat-use patterns, and efficiency of tests for habitat selection, using each location-estimation method. We placed transmitters at known locations throughout 2 study sites and crew members estimated bearings to the transmitters without knowing their actual locations. We then estimated locations to the transmitters from the recorded bearings and examined effects of subsequent radiotelemetry triangulation error on determining habitat use and detecting habitat selection using MLE point-estimates and 95% error ellipses. In each study site, point estimates correctly classified more habitats (70.1% and 68.7%) than error ellipses (62.2% and 64.6%). Point estimates were more accurate, but less precise, at determining patterns of habitat use, and more efficient, but less precise, at determining habitat selection, than error ellipses in both study sites. The ability of each method to correctly classify habitats, to correctly identify habitat-use patterns, and to correctly identify habitat selection increased as ellipse sizes decreased relative to habitat-patch sizes.