Unexpected migration patterns in a high-latitude breeding songbird: Evidence from multi-sensor geolocators and isotopes

Seasonal migration allows animals to use habitat where conditions are unfavorable for part of the year but may constrain breeding ranges due to the costs of longer migrations as ranges expand poleward. In species with large ranges, high-latitude breeding populations may employ different migration strategies allowing them to persist far from other core nonbreeding areas. The myrtle warbler (Setophaga coronata coronata) has two disjunct nonbreeding ranges in North and Central America—one along the Gulf Coast and the other on the Pacific. Previous work indirectly linked birds breeding in Alaska with the Pacific nonbreeding area, suggesting that high latitude populations evolved a shorter migration route. We directly tested this hypothesis using geolocators measuring both light and atmospheric pressure to track Alaskan myrtle warbler migration in fine detail and inferred nonbreeding areas using hydrogen isotopes for a larger sample of birds breeding in Alaska, British Columbia, and Alberta. We found, contrary to expectations, that all geolocator-tracked birds—and 95% of birds with stable isotope data—migrated to the southeastern United States, a much longer migration than expected for a species commonly considered a “short-distance” migrant. We additionally demonstrate the advantages of pressure geolocation for characterizing migratory behavior at a fine scale.

Methods
We deployed 30 multisensor geolocators (Migrate Tech BARP30Z11-DIP) on myrtle warblers (Setophaga coronata coronata) breeding in Far North Bicentennial Park, Anchorage, AK in June 2022 and recovered 6 of these geolocators in June 2023. The geolocators sampled light intensity every minute, recording the maximum intensity every 5 minutes, and recorded pressure and temperature every 20 minutes.

Light data (.lux) was adjusted for clock drift, and raw geolocator data files (.lux, .deg) were trimmed to remove data recorded after activation of the tag but before deployment on the bird, and after removal from the bird but before deactivation. 

Light data was processed with the R packages TwGeos and GeoLight. Pressure data was processed both alone and in combination with light data using the R package GeoPressureR. 

We inferred wintering areas for 167 myrtle warblers breeding in Anchorage, Alaska (n = 46), northern British Columbia (n = 54), and Alberta (n = 67) using stable hydrogen isotopes from greater covert feathers grown during the pre-alternate molt. We also assessed stable hydrogen isotopes in feathers grown during the pre-basic molt (i.e. on breeding grounds) for myrtle warblers breeding in Alaska (n = 12) and British Columbia (n = 12). Feather sample preparation and hydrogen pyrolysis were performed at the Cornell University Stable Isotope Laboratory. Stable isotope data was processed using the R package assignR.