Moving wildlife tracking forward under forested conditions with the SWIFT GPS algorithm

Abstract

The remote collection of animal location data has proliferated in recent decades, and higher-frequency data are increasingly available with battery-saving optimisations such as ‘snapshot’ algorithms that acquire GPS satellite data and post-process locations off-board. This is the first study to assess the effects of vegetation and topography on the fix success rate and location error of global positioning system (GPS) devices that use the SWIFT fix algorithm, developed by Lotek. To assess fix success rate (FSR—the proportion of successful fixes compared to the total number of attempts) and location error (LE), we conducted a stationary test at a predominately forested site on the South Island of New Zealand. The overall FSR was 83% (± 15.3% SD), which was affected strongly by canopy closure above 90%. Half of the locations were within 8.65 m of the true location, 79.7% were within 30 m, and 95% of locations were within 271 m. When 6 or more satellites were used, this reduced to 4.92 m and 18.6 m for 50% and 95%, respectively. Horizontal dilution of precision (HDOP), the number of satellites, and canopy closure all influenced location error. To field test the fix success rate of SWIFT GPS devices, we deployed them on forest-dwelling parrots with 2 and 3-h fix intervals, which showed similar FSR results to the stationary test when cavity-nesting individuals were removed (FSR mean ± SD = 81.6 ± 5.0%). The devices lasted an average of 147 days before depleting the battery, resulting in an average of 1087 successful fixes per individual at an average time of 9.38 s to acquire the GPS ephemeris, resulting in an average of 3.73 attempted locations per mAh of battery for PinPoint 350 devices. Our study provides a baseline for fix success rates and location errors under forested conditions that can be used for future SWIFT GPS tracking studies.