Natural take-off flight performance is repeatable and scalable in mixed flock tit species

Modelling small bird flights is challenging, as small birds (<20 g) rarely engage in steady-state flights assumed by traditional flight models. Additionally, their flights are difficult to measure in the field owing to the mass of high-resolution GPS loggers and the low resolution in lightweight geolocators. This study aimed to measure and model spontaneous, non-steady-state take-off flights of wild tit species that forage in mixed flocks during winter, despite differences in body sizes and ecological niches. We recorded 1434 spontaneous flights from 49 individuals of great tits (Parus major), blue tits (Cyanistes caeruleus) and marsh tits (Poecile palustris) in the Paridae family using two automated radio frequency identification infrared tunnels. We developed a power margin model to predict in-flight acceleration based on initial velocity, scaled by species-specific wing loading. Lastly, we tested the repeatability of flight metrics to assess whether flight may be interpreted as an individual behavioural trait. We found significant differences in initial velocity and in-flight acceleration between species. The observed difference in the in-flight acceleration aligned with the power margin model prediction. Finally, we found significant individual repeatability in both initial velocity and in-flight acceleration. Our findings suggest that the flight performance of small birds can be consistently measured in the field, modelled beyond steady-state assumptions and interpreted as an individual behavioural trait.