Same habitat, different species: otolith microchemistry relationships between migratory and resident species support interspecific natal source classification

We tested the hypothesis that otolith trace elemental signatures (microchemistries) of mottled sculpin Cottus bairdi, slimy sculpin C. cognatus, and juvenile coho salmon Oncorhynchus kisutch were predictive of those of juvenile steelhead O. mykiss across many sites within the Lake Michigan basin. Laser ablation inductively coupled plasma mass spectrometry was used to generate otolith microchemistry signatures for each individual fish. For each species pair, statistical correlations of mean otolith concentrations of Mg, Mn, Cu, Zn, Sr, Ba, and Pb for each site were estimated. Linear equations describing these relationships were used to transform juvenile steelhead otolith microchemistry data to those of each of the other species. Transformed otolith microchemistry data were subjected to random forest classifications developed for mottled sculpin, slimy sculpin, and juvenile coho salmon to assess interspecific natal source assignment accuracies. Steelhead otolith concentrations of Sr were significantly correlated with those of each of the other species, whereas otolith concentrations of Ba and Mn were significantly correlated among some species pairs, but not others. Natal source assignment accuracies of juvenile steelhead to site and watershed generally decreased when otolith microchemistry data were transformed to those of mottled sculpin, slimy sculpin, and coho salmon. Miss-assigned fish often classified into nearby watersheds within larger hydrologic units, leading to higher assignment accuracies at coarser geographical resolutions (75-97% correct assignment to hydrologic unit for each species). These findings suggest that applications of otolith microchemistry data may extend beyond the species from which they are collected.