•  
  •  
 

Keywords

Satellite Transmitter, Black Bear, GPS Accuracy, Wildlife Telemetry, Collar Performance, Telemetry Error, Transmitter Error

Abstract

The Arkansas Game and Fish Commission tracks American black bears (Ursus americanus) using Global Positioning System (GPS) satellite telemetry collars in southeastern Arkansas to assess habitat use and behaviors. Companies test the accuracy of their devices, but researchers have found that natural environments, such as the forested habitat of black bears, can cause an increase in spatial error rates. We therefore tested 1 GPS collar in realistic conditions to help refine data interpretation. We placed the collar at 20 random locations in Columbia County, AR, stratifying habitat type based on data collected from black bears. We determined the ‘true’ location of collar placement using a Trimble Geo7X and measured vegetation height and percentage canopy openness at each point. Across 400 points (20 per location), real error (distance from true location) averaged 16.5 m, while the collar-reported errors averaged 23.6 m. Error types were modestly correlated (r = 0.54). Models grouped errors by open versus closed canopy (F3,387 = 5.6, p < 0.001). Open-canopy sites were on average 96% likely to report a location with high accuracy, while forested sites were 23% likely. Vegetation category explained the most variation in models, and data had patterns of increasing error (both accuracy and precision) with increasing vegetation height and decreasing canopy openness. Given that 81% of data had greater accuracy than collar-reported errors, we concluded that the technology has overall good accuracy and reports conservative error estimates. The fact that 86% of black bear locations are under forested cover will inevitably decrease accuracy and precision of the collected data. Therefore, filtering data using collar-reported error or building polygons to incorporate error into analyses may be useful.

Download

Share

COinS