Testing the functionality and contact error of a GPS-based wildlife tracking network

Telemetry is a fundamentally important tool for studying animal movements. Traditional telemetry systems have provided time-specific information on locations of individuals; however, recent developments in instruments allow for the tracking of networks of interactions among individuals. Currently, these devices rely on very high frequency (VHF) radio technology, and cannot precisely gauge where interactions occurred without on-site triangulation. Furthermore, although commercially available devices can log successful communication attempts between collars as little as 1 m apart, researchers cannot accurately determine distances within the pre-set detection (contact) threshold of these units. Data loss from proximity-logging devices, and even traditional telemetry devices that are lost or damaged in the field, poses another obstacle to monitoring wildlife social networks. We have developed a prototype proximity-logging Global Positioning System (GPS) collar that offers greater spatial resolution of social interactions, and reduces probability of data loss. In this study, we used captive bighorn sheep (Ovis canadensis) to test the GPS capabilities, contact rates, and contact distance error of our prototype collars. The GPS fix success rate of our collars was >97.93% (n = 95,041; 95% CI = 97.84-98.02%). The collars were communicating with each other about 98% (n = 22,253; 95% CI = 98.50-98.81%) of the time and reciprocal communication occurred 9% of the time. Contact distance error was 9.5 m, which is what would be expected taking into account a baseline GPS spatial error of ±5 m in open environments. The high GPS fix success, low GPS error, and ability to log accurate social interactions with low contact distance error by our prototype collars suggest that the implementation of GPS with proximity-logging technology has the potential to improve currently available social network data.