Proximity data and social networks: a validation study of two models of mobile dataloggers in the semi-fossorial and social rodent, Octodon degus

The use of mobile dataloggers to gather proximity data has generated novel insights into the social network structure of animals. The performance of mobile dataloggers may be impacted by different environmental and social conditions. Thus, validation studies that account for species-specific habitat use and behavior are essential. We conducted a series of validation studies of two brands of mobile dataloggers (Gipsy6, Technosmart, Italy; ProxLog, IoSA, Belgium) on the semi-fossorial rodent, Octodon degus. First, we determined the sensitivity of both types of dataloggers to track position and closeness in different habitats (Experiments 1 and 2). To determine if the structure of the network is influenced by the kind of data collected, we compared social networks generated by the two datalogger types and a passive monitoring RFID (radio frequency identification) system previously used in our study system (Experiment 3). The Gipsy6 units were more sensitive in open areas, within shrubs, and when into a burrow opening than ProxLog units. Mantel tests revealed that networks based on Gipsy6 and ProxLog units were weakly correlated with those from the RFID system. Both the group-level (weight, density, and eigenvector) and node-level (eigenvector) network metrics generated by the Gipsy6 and ProxLog dataloggers were not correlated with RFID. Degus carrying both types of dataloggers lost 7% body mass, even though the dataloggers weighed < 5% of body mass. Gipsy6 dataloggers were more effective in our study system, though lighter units are required to ensure animal wellbeing. Our observations suggest that the two dataloggers capture different aspects of social dynamics compared to the RFID system. We make recommendations for the design of validation studies of mobile dataloggers.