Estimating the onset of natal dispersal for a large diurnal raptor: A methodological comparison

We estimated the onset of natal dispersal for a large diurnal raptor with high propensity towards large-scaled exploratory movements during the post-fledging period, the White-tailed Sea Eagle (Haliaeetus albicilla). We analysed GPS tracking data of 21 juveniles with respect to the onset of natal dispersal comparing six methods available from the recent literature. While none of the methods significantly differed from the visual method, the Distance Threshold method underestimated the dispersal onset for some individuals. Likewise, coefficient of variation methods overestimated the dispersal onset in few cases, presumably because the temporal scale of available GPS fixes did not correspond to the scale of discrete dispersal movements. We conclude that all tested methods are generally suitable to estimate the dispersal onset, specifically if the research question does not depend on an exact but rather a rough estimate. A visual determination might increase flexibility to account for individual behavior and yields consistent results across individuals, but highly reduces the comparability across observers and studies. For research questions relying on exact estimates, we propose using a combination of an automated method and a visual determination as a back-up method for single individuals with clear under- or overestimation. An exploratory comparison showed that the temporal resolution of the GPS may further affect the accuracy of natal dispersal estimates. For individuals with clear movement patterns, high-resolution movement data could increase the accuracy of Coefficient of Variation methods. We underline the necessity for further investigation on the effects of temporal resolution on dispersal onset estimates.