Inferring connectivity and molt patterns of a migratory shorebird using extrinsic and intrinsic lines of evidence

Abstract Tracking birds to identify migratory linkages in space and time is important for managing and conserving populations, but data derived from extrinsic markers, such as field-readable bands and tracking devices can be biased by unequal distribution of capture and resight effort, and by potential consequences of carrying tracking devices. We used stable isotope profiles, in tandem with resight records and tracking data, to examine patterns of connectivity in Calidris canutus (Red Knot) occurring in the northern Gulf of Mexico. While Texas, Louisiana, and Florida all support terminal nonbreeding populations of C. canutus, some individuals marked in those states in other seasons have been reported from nonbreeding areas along the Pacific coast from southern Mexico to Chiloé Island, Chile. Linear discriminant functions performed well at correctly assigning C. canutus to each of the northern Gulf states and to Chile, but poorly for those at lower latitude sites on the Pacific. Examination of resight records and geolocator tracks confirm flight feather molt does not always take place on the terminal nonbreeding area. Applying discriminant functions on a set of samples from spring captures in Texas and Louisiana lacking additional extrinsic information confirmed that most C. canutus present in spring molted there in the fall, with smaller proportions deriving from the Pacific locations. Approximately 15% of these unknown samples from spring captures were outliers, implying that there is either significant unsampled variability in isotope profiles at the known sites, or there are additional yet-unknown terminal nonbreeding or molt stopover sites used by C. canutus that use northern Gulf sites primarily in spring. We found support for fine-scale differences in isotope profiles even within states, and explore hypotheses of potential explanations for these distinctions.