Influence of different GPS schedules on the detection rate of wolf feeding sites in human-dominated landscapes

GPS collars are commonly used to estimate predation and scavenging rates in large carnivores. However, little information is available on the impact of different schedules on feeding site detection rates. Here, we evaluated the effect of different GPS schedules on the detection rate of wolves' feeding sites in a human-dominated landscape of NW Iberia (Galicia), where the main food sources for wolves were large livestock ungulates (horses and cattle). Combining an intensive GPS schedule of 20 min time intervals between locations, used as reference values, with the field examination of clusters of locations, on average, we observed a 40 and 24 % decrease in clusters and events detection rates from 20 to 40 min, respectively, and a 13 and 15 % decrease from 40 to 60 min, respectively. On the other hand, on a subset of monitoring days, from 10 to 20 min, the proportion of events detected decreased by 6 %. The decrease in detection rates over time was similar across livestock species and age classes. It is worth noting that the decrease in detection rates was higher for scavenging events, which can be common in human-dominated landscapes, compared to predation events. Our results indicate that using long time intervals between locations to study wolf feeding behavior in human-dominated landscapes will underestimate not only predation rates, but also the importance of scavenging events. Since a 10-min schedule reduces the expected battery life of collars notably and the decrease in detection rate was low between 10 and 20 min, compared to 20 min and the longer time intervals explored, we recommend a GPS schedule of 20 min to study the feeding behavior of wolves in human-dominated landscapes.