The 2008 Joint Meeting of the Society for Range Management and the America Forage and Grassland Council. |
Wednesday, January 30, 2008
13
Too Much Data? Optimizing GPS Collar Data Collection Schedules
Humberto Perotto-Baldivieso, Susan M. Cooper, Manuel Figueroa-Pagan, and Jessica Romo. Uvalde, Texas Agricultural Experiment Station, 1619 Garner Field Rd., Uvalde, TX 78801
Monitoring behavior of livestock and wildlife using global positioning systems (GPS) is relatively new and provides a unique opportunity to test spatial distribution under different management systems 24 hours a day with a high degree of accuracy. This technology also allows the assessment of habitat and resource selection without the confounding influence of disturbance by human observers. However, GPS technology generates the problem of acquiring too much or too little data and running into spatial and temporal autocorrelation issues that can hinder statistical analyses. Our main goal was to examine the most suitable time intervals for GPS collar data collection and their potential use for livestock and wildlife studies. Our specific objectives were (1) to analyze the degree of direction and spatial correlation using different time intervals with GPS data collected from white-tailed deer and rangeland cattle (2) determine the best time-frequency intervals for home range analysis, path analysis, habitat use and habitat selection. We collected GPS collar data at 5 minute-intervals for 12 consecutive days, for 4 periods during 1 year in South Texas rangelands. Data was then thinned to 10 min, 20 min, 30 min, 60 min, and 120 min intervals. Direction and spatial correlation analyses were conducted as well as fractal dimension and path analysis for each cow and deer. Small time intervals showed a highly significant degree of correlation between direction and distance. Time intervals > 1hour provide non-significant correlation for home range analysis. Home ranges increased significantly in size as GPS data collection time intervals decreased. On the other hand, very frequent time intervals may be required to conduct meaningful analysis for path use and animal movement studies.