Comparing inferences of solar geolocation data against high-precision GPS data: annual movements of a double-tagged black-tailed godwit

Annual routines of migratory birds inferred from archival solar geolocation devices have never before been confirmed using GPS technologies. A female black-tailed godwit Limosa limosa limosa captured on the breeding grounds in the Netherlands in 2013 and recaptured in 2014 was outfitted with both an Intigeo geolocator and an UvA-BiTS GPS-tracker. The GPS positions show that, after its breeding season in 2013, the godwit flew 2035 km nonstop from the Netherlands to southern Spain. It then spent the entire nonbreeding season in the southern part of the Iberian Peninsula before returning to the Netherlands the following spring, stopping for 7 days in the delta of the Ebro River in Spain, and again for a day in central Belgium. To compare the geolocation and GPS data, we analysed the geolocation data with two open-source software packages: one using a threshold method (GeoLight) and the other a template-fit approach (FLightR). Estimates using GeoLight, on average, deviated from the individual's true position by 495.5 +/- 1031.2 km (great circle distance with equinoxes excluded), while FLightR estimates deviated by 43.3 +/- 51.5 km (great circle distance with equinoxes included). Arrival and departure schedules estimated by FLightR were within 12 h of those determined by the GPS tracker, whereas GeoLight's estimates were less precise. For the analysed track, FLightR represents an improvement over GeoLight; if true for other species and conditions, FLightR will hopefully help establish more precise and accurate uses of geolocation data in tracking studies. To aid future improvements in the analysis of solar geolocation data, we also provide the GPS and geolocation data files together with our R scripts as Supplementary material Appendix 1-6.