How much can we trust GPS wildlife tracking? An assessment in semi-free-ranging Crested Ibis Nipponia nippon

: GPS tracking has been increasingly used for wildlife studies in recent decades, but its performance has not been fully assessed, especially for newly developed lightweight transmitters. We assessed the performance of eight GPS transmitters developed in China by attaching them to Crested Ibises Nipponia nippon confined to two acclimation cages mimicking real habitats. We calculated the distance between GPS locations and the centroid of the cages as the positioning error, and used the 95% (95th percentile) positioning errors to define the accuracy. The positioning success averaged 92.0%, which is much higher than that of previous studies. Locations were not evenly distributed by Location Class (LC), with the LC A and B locations accounting for 88.7%. The observed 95% positioning error in the locations of LC A (9-39 m) and B (11-41 m) was quite accurate, while up to 6.9-8.8% of poor-quality locations were detected in LC C and D with >100 m or even >1, 000 m positioning error. Positioning success and accuracy were different between the test sites, probably due to the difference in vegetation structure. Thus, we argue that the tested transmitters could provide a large proportion of high-quality data for fine-scale studies, and a number of poor-quality locations that need attention. We suggest that the HPOD (horizontal dilution of precision) or PDOP (positional dilution of precision) be reported instead of the LC as a measurement of location accuracy for each location to ensure identification and filtering of implausible locations.