Working with GNSS collar data. The importance of pre-analysis when setting the sampling interval

Abstract

The use of tracking collars implemented with Global Positioning System (GPS) has allowed increasingly complex analyses of the movement and space use of tagged animals. Fundamental to these studies is the setting of parameters related to the technology itself. In this study we assessed the effect of the fix interval on certain ecological variables used to describe behavior, animal movement, home range size and habitat use through parameters such as path length, sinuosity, core area or minimum convex polygon. The aim of the study was to select the most reliable time interval in the study area in the northwest province of Jaén (Spain) for a subsequent use of the collars on wild ungulates. For this purpose, we designed and developed static and moving tests for evaluating 6 GNSS collars, one of them improved by using multi-costellation and multifrequency module, programmed with 5 different time intervals. The kinematic results suggest using caution when programming the collars since differences in the sampling interval leads to important differences in the real trajectories. Movement descriptors varied according to time interval, being longest and more winding with the shortest intervals. Similarly, home range approximations were closer to reality as more fixes were obtained. We progressively underestimated the use of the central core zone and the distance reached by the individual as we increased the time interval. However, there is a trade-off between time, autocorrelation and battery expenditure that is almost the only limiting factor. The experimental results show that a fixing interval of 30 min is an optimal and reliable value to studies focused on the estimation of path length or home range estimation. This value also involves a trade-off between high-resolution data and battery life.