Home Range Estimator Method and GPS Sampling Schedule Affect Habitat Selection Inferences for Wild Turkeys

Understanding patterns in the spatial distribution of individuals in a population is a central question in ecology. Concurrent with advances in biotelemetry devices, development of home range estimator methods incorporating the temporal component of locational fixes are increasingly used to investigate these patterns at finer scales. However, these methods may necessitate sampling schedules that limit battery life and study period length. Practically, evaluating how home range estimator methods affect calculations of space use and habitat selection prior to deployment of biotelemetry devices could help researchers optimize data acquisition schedules. We quantified spatial overlap between a home range estimator using temporal information (dynamic Brownian bridge movement model [dBBMM]) and home range estimators not incorporating the temporal component of fixes (ad hoc and href kernel density estimator [KDE]) across differing sample schedules, and the resulting error in habitat selection ratios using data collected from wild turkeys (Meleagris gallopavo) equipped with Global Positioning Systems units in Texas, Georgia, South Carolina, and Louisiana, USA, during February-May 2015. When comparing ranges created from KDEs to dBBMM, commission errors were large (20-80%) and did not diminish with increased sampling rates. In contrast, omission error rate declined quicker and improvements were minimal when fix rates increased beyond 4/day. Compared with ranges estimated with dBBMM, KDEs poorly defined the spatial bearings of an individual's range, overestimated areas of use, underestimated areas avoided, and showed different patterns of habitat selection. Our results suggest home range estimator methods incorporating temporal information seem capable of estimating ranges encompassing nearly all area used by an individual and should be used even at relatively low-frequency collection schedules to assess home ranges of wild turkeys. If researchers are interested in describing habitat selection of wild turkeys, we recommend a sampling schedule of <= 1 location/hour during daytime and dBBMM for range estimation. (C) 2018 The Wildlife Society.