Assigning harvested waterfowl to geographic origin using feather δ(2)H isoscapes: What is the best analytical approach?

Establishing links between breeding, stopover, and wintering sites for migratory species is important for their effective conservation and management. Isotopic assignment methods used to create these connections rely on the use of predictable, established relationships between the isotopic composition of environmental hydrogen and that of the non-exchangeable hydrogen in animal tissues, often in the form of a calibration equation relating feather (δ(2)H(f)) values derived from known-origin individuals and amount-weighted long-term precipitation (δ(2)H(p)) data. The efficacy of assigning waterfowl to moult origin using stable isotopes depends on the accuracy of these relationships and their statistical uncertainty. Most current calibrations for terrestrial species in North America are done using amount-weighted mean growing-season δ(2)H(p) values, but the calibration relationship is less clear for aquatic and semi-aquatic species. Our objective was to critically evaluate current methods used to calibrate δ(2)H(p) isoscapes to predicted δ(2)H(f) values for waterfowl. Specifically, we evaluated the strength of the relationships between δ(2)H(p) values from three commonly used isoscapes and known-origin δ(2)H(f) values three published datasets and one collected as part of this study, also grouping these data into foraging guilds (dabbling vs diving ducks). We then evaluated the performance of assignments using these calibrations by applying a cross-validation procedure. It remains unclear if any of the tested δ(2)H(p) isoscapes better predict surface water inputs into food webs for foraging waterfowl. We found only marginal differences in the performance of the tested known-origin datasets, where the combined foraging-guild-specific datasets showed lower assignment precision and model fit compared to data for individual species. We recommend the use of the more conservative combined foraging-guild-specific datasets to assign geographic origin for all dabbling duck species. Refining these relationships is important for improved waterfowl management and contributes to a better understanding of the limitations of assignment methods when using the isotope approach.