Disentangling the effects of biogeoclimatic factors and local stressors on breeding harlequin duck presence and abundance using field surveys, MaxEnt, and stable isotope turnover models

Harlequin ducks breed at low densities on montane streams across a large geographic region. Variation within and between watersheds can make it difficult to isolate the effects of local stressors within a broader suite of overlapping and interacting environmental factors influencing harlequin presence and abundance. For example, fish introductions may lower harlequin breeding habitat quality, either directly, through competition for food resources, or indirectly, by inducing antipredator behaviours in shared invertebrate prey or attracting predators, such as eagles. However, fish presence can be confounded with stream size, elevation, and reach gradient. To disentangle these effects, I used field surveys and complementary habitat suitability modeling in MaxEnt. I constructed candidate models using suites of biogeographic variables hypothesized to affect breeding harlequin duck presence and abundance. Proportion of grassland and shrubland was the best predictor of true absences in the field surveys, denoting unsuitable harlequin habitat. Annual temperature range contributed the most information to the final MaxEnt model, with low temperature range being most suitable (coast and mountains). Both analyses found habitat suitability increased with stream order, but order alone was insufficient in predicting harlequin distributions, likely because stream orders were fairly evenly distributed. Overall, climate and distance to coast appeared to constrain harlequin distributions at the broadest spatial scale while stream order seemed to be a subsequent environmental "filter." Lastly, I tested the feasibility of making inferences about breeding habitat using body tissue stable isotopes from recent arrivals to the wintering grounds. To account for the incorporation of wintering habitat isotopes, I created models of claw and blood cell isotopic turnover using flight feather regrowth as an index of time since arrival. Claw turnover mirrored blood cell turnover, but with a "lag" due to the time required for "current" claw tissue isotopes to grow from the base to the tip of the claw. In stream invertebrates collected during the field surveys, δ13C decreased with distance from the coast while δ15N increased with river order. As both factors were also associated with habitat suitability, isotope turnover back-calculations may offer a large scale method to examine harlequin duck breeding habitat usage.