Characterizing the Prey Landscape of Endangered Cook Inlet Beluga Whales Using an eDNA Time Series

Our ability to understand habitat utilization in the context of foraging, prey dynamics, and anthropogenic disturbance is critical for the effective management of protected species globally. Here we examine the occurrence of the endangered Cook Inlet beluga whale (Delphinapterus leucas) in the context of seasonally changing prey fields by leveraging quantitative eDNA metabarcoding approaches. Specifically, eDNA approaches were used to generate estimates of fish abundances using a joint modeling framework that derives total DNA concentration from a coho (Oncorhynchus kisutch) qPCR assay and bias-corrected proportional abundance of multiple salmonid species using MiFish 12S metabarcoding of environmental and mock community samples. Temporal patterns of beluga whale occurrence were documented in two estuarine systems using visual observations, metabarcoding, and a novel beluga whale qPCR assay. We demonstrate that weekly time series before, during, and after beluga whale occurrence provide sufficient resolution to resolve changes in predator occurrence and prey field shifts across Cook Inlet. Specifically, we observed Cook Inlet beluga whales co-occurring with increases in coho DNA concentrations in the vicinity of Twentymile River at the terminus of Turnagain Arm. Monthly sampling on the Kenai River captured known seasonal anadromous fish migration patterns and identified a clear mismatch between Cook Inlet beluga whale habitat utilization and spring eulachon (Thaleichthys pacificus) runs. Our results also provide valuable insights to improve quantitative metabarcoding approaches through informed selection of qPCR targets and mock communities to support future eDNA-based studies in the highly turbid, dynamic habitats of Cook Inlet. Specifically, our results highlight the need for increased temporal sampling frequency as opposed to additional spatial and depth sampling in systems with extreme tides, as well as the use of larger pore sizes (> 5um) to facilitate filtration in samples with high suspended glacial silt loads. We also document the capacity for eDNA signatures to integrate across 18 km spatial scales during exceptionally large tidal bores (> 10 m) that characterize Turnagain Arm. Ultimately, this study demonstrates the value of pairing visual surveys of endangered Cook Inlet beluga whales with strategic shore-based eDNA sampling to generate information on the co-occurrence of Cook Inlet beluga whales and putative prey fields.