1. Migratory birds travel vast distances and the timing of migratory flights can affect survival and the ability to reproduce. For Neotropical migrant songbirds, early spring departure from wintering sites, early arrival to the breeding grounds and higher reproductive success have been related to the use of suitable habitats and environmental conditions during the non-breeding season. However, how migratory strategies are shaped by winter habitat choice is largely unknown due to the general inability to track birds from specific wintering habitats to stopovers or breeding destinations.

2. We assessed how winter habitat (native forest vs. shade-grown coffee plantations) relates to spring departure date and migration pace in Swainson’s Thrush (Catharus ustulatus). We also determined the effect of departure date and total migration duration on the arrival date of birds detected near or within their breeding range.

3. We used a novel application of Motus radio telemetry arrays to track individuals from their wintering grounds in the Andes of South America along their migratory journey to North America.

4. We found variation in migratory strategies between habitats, with birds wintering in native forest departing later than birds in coffee. We present isotopic evidence for native forest being of higher quality than shade coffee for Swainson’s Thrush and hypothesize that moister conditions in forest, as shown by stable isotope (δ¹³C) analysis of thrush whole blood, provides favourable pre-migratory conditions allowing birds to delay departure from wintering grounds. Habitat, between-site and -year variation in departure date, suggests that birds made facultative adjustments to winter habitat quality and environmental conditions. Independent of habitat, birds that departed later migrated faster and this pattern was maintained along the migration route (n = 44). Migrating earlier and slower or later and faster was unlikely to result in significant differences in arrival time to breeding destinations. 

5. Our findings reveal underappreciated complexity in migratory decisions by long-distance migrants that contrast with the current paradigm of earlier departures and arrival from optimal habitats. The next step is to understand the relative fitness benefits of early versus late schedules or whether each strategy is an equally good response to experienced conditions.

We captured Swainson’s Thrush (Catharus ustulatus) between 10 January and 30 March from 2014 to 2018 at two study sites located on the western slope of Colombia’s Eastern Andes (Hacienda La Fragua and Los Vientos). We captured birds in shade-coffee plantations and secondary forests at each site. Birds were banded, processed, and fitted with digitally-coded VHF transmitters using leg-loop harnesses made of elastic thread. We tracked birds locally and during spring migration using an array of over 300 automated radio-telemetry stations spanning key migratory flyways. We used Motus to determine the effect of occupying native forest or shade-grown coffee during the overwintering season on spring migration departure date (Data set 1) and migration pace (Data set 2) of a Nearctic-Neotropical migratory landbird, the Swainson’s Thrush (Catharus ustulatus). We also determined the effect of departure date and total migration duration on the arrival date of birds detected near or in their breeding range (Data set 3). To describe the relative quality of shade coffee versus native forest, we measured stable carbon (δ¹³C) and nitrogen (δ¹⁵N) isotopes in whole blood to assess moisture patterns and the relative contribution of fruit and insects as sources of dietary protein (Data set 4).