Comparison of Seven DNA Metabarcoding Sampling Methods to Assess Diet in a Large Avian Predator

Abstract

ABSTRACTDNA metabarcoding is a rapidly advancing tool for diet assessment in wildlife ecology. Studies have used a variety of field collection methods to evaluate diet; however, there is a pressing need to understand the differences among sampling methods and the downstream inferential consequences they may have on our ability to document diet accurately and efficiently. We evaluated seven DNA metabarcoding sampling methods to assess the diet of a large avian predator: Buteo lagopus (rough‐legged hawk). We collected beak swabs, talon swabs, cheek (buccal) swabs, cloacal swabs, and cloacal loops from captured birds, and collected fecal samples from both captured and uncaptured birds. We described and compared variation in prey recovery within and among the seven sampling methods and identified appropriate analytical methods to compare diet among individuals sampled via different methods. Beak and talon swabs produced the highest prey detection rates, yielded the greatest prey richness per sample, and contributed the most to an individual's total prey richness per sampling occasion compared to other sampling methods. Within individuals sampled using five methods during a single capture occasion, cloacal swabs and cheek swabs positively predicted prey richness and average prey mass, respectively, from fecal samples. While all methods identified similar dominant prey taxa that were consistent with prior diet studies, beak and talon swabs detected greater prey richness at both the individual and population levels. We propose a food residue duration hypothesis whereby methods which sample areas containing food DNA consumed from longer and more continuous pre‐sampling time intervals explain variation among sampling methods in observed prey richness. Choice of sampling method can influence predator diet characterization and is particularly important if researchers wish to quantify uncommon diet items or compare diet metrics using samples collected via different methods.