Performance of three GPS collars to monitor goats' grazing itineraries on mountain pastures

Despite more than 15 years of GPS use in studies with domesticated animals, surprisingly little high-resolution data was collected on grazing itineraries of livestock. It seems as if each research group has its preferred GPS tracking equipment, but little comparative data about the reliability of different types of collars are available. This study provides such data for three very different GPS collars that were tested on a human observer's back and on herded goats in the rugged Hajar Mountains of northern Oman. At a set logging interval of 15 s, the obtained number of position fixes per minute varied from 2.3 to 3.8 and differed significantly (P < 0.001) between the three devices in obstructed terrain while differences were negligible (P > 0.05) in open terrain. The large variations between the devices in the obtained latitude, longitude and particularly the altitude data were likely due to differences in the factory-made basic setup of the GPS receivers which placed specific weights on signal reliability and trigonometric properties. In the topographically disrupted study environment, recorded values of the position dilution of precision (PDOP) proved to be of little use as indicators of position quality as they were poorly related to the precision of latitude, longitude and altitude values obtained. If accurate altitude data are required, such as for studies on animals' energy expenditure, separate recording of variations in barometric pressure at the same time intervals as those of the GPS collars is recommended. (C) 2008 Elsevier B.V. All rights reserved.