Simpler methods can outperform more sophisticated ones when assessing bird migration starting date

Abstract

The migration starting date (MSD) of 30 Eurasian Teal and 8 Eurasian Stone-curlews was estimated by processing tracking data with four methods. A significant difference was found for Teal, with methods fitting models on Net Square Displacement postponing MSD compared to distance threshold methods. In Stone-curlews the four methods provided comparable MSD estimates. The reliability of Teal MSD was evaluated comparing the estimated ratio between time at stopover/time in flight with the ratio expected from the optimal migration theory. Threshold methods provided estimates closer to the ratio expected for time-minimizing migrants and therefore seems the most reliable approach, especially for datasets with irregular sampling and variable migratory strategies.