TY - JOUR
T1 - Otolith-Chemistry-Based Discrimination of Wild- and Hatchery-Origin Steelhead across the Lake Michigan Basin
AU - Watson, Nicole M.
AU - Prichard, Carson G.
AU - Jonas, Jory L.
AU - Student, James J.
AU - Pangle, Kevin L.
N1 - Funding Information:
We thank Kyle Brumm and Kieran Elder for help with field sampling and preparation of otolith samples, as well as the many Michigan Department of Natural Resources biologists and technicians who helped collect samples. Funding for this research was provided by Central Michigan University and the Great Lakes Fishery Trust (project 1298). There is no conflict of interest declared in this article.
Publisher Copyright:
© 2018 American Fisheries Society
PY - 2018/8
Y1 - 2018/8
N2 - Effective conservation and management of mixed-stock fisheries depends upon accurate stock identification of individual fish. We evaluated the utility of otolith chemical analysis as an approach to classify juvenile, premigratory steelhead Oncorhynchus mykiss collected in tributaries of Lake Michigan as either wild or hatchery origin. Two primary hypotheses proposed were that otolith chemistry can be used (1) to classify juvenile, hatchery-origin steelhead to the correct hatchery of origin and (2) to correctly classify stream-collected juvenile steelhead as wild or hatchery origin. Using laser-ablation inductively coupled plasma mass spectrometry, we analyzed the concentrations of seven elements (using 25Mg, 55Mn, 65Cu, 66Zn, 88Sr, 137Ba, 208Pb) in the otoliths of juvenile steelhead collected from five hatcheries and 25 streams in the Lake Michigan basin. When discriminating among hatcheries, only 4.41% of fish misclassified to an alternate hatchery when subjected to the best random forest classification algorithm that included all elements as predictor variables. Distinct chemical signatures between fish of wild and hatchery origin supported 100% classification accuracies of known-wild, age-0 steelhead as wild origin for 19 of the 25 streams sampled. Misclassification of wild, age-0, and hatchery-origin fish, which tended to occur for streams that were located in close proximity to the hatchery, never exceeded 3.4% for a given stream. Our findings demonstrate highly successful discrimination of hatchery-origin and wild juvenile steelhead across a broad geographic range. Applying the classification algorithms developed herein to unknown-origin steelhead provides the ability to infer survival of year-classes from specific hatcheries. Further, the ability to differentiate hatchery- and wild-origin fish will assist in stock assessment efforts allowing for increased effectiveness of conservation and management of the species.
AB - Effective conservation and management of mixed-stock fisheries depends upon accurate stock identification of individual fish. We evaluated the utility of otolith chemical analysis as an approach to classify juvenile, premigratory steelhead Oncorhynchus mykiss collected in tributaries of Lake Michigan as either wild or hatchery origin. Two primary hypotheses proposed were that otolith chemistry can be used (1) to classify juvenile, hatchery-origin steelhead to the correct hatchery of origin and (2) to correctly classify stream-collected juvenile steelhead as wild or hatchery origin. Using laser-ablation inductively coupled plasma mass spectrometry, we analyzed the concentrations of seven elements (using 25Mg, 55Mn, 65Cu, 66Zn, 88Sr, 137Ba, 208Pb) in the otoliths of juvenile steelhead collected from five hatcheries and 25 streams in the Lake Michigan basin. When discriminating among hatcheries, only 4.41% of fish misclassified to an alternate hatchery when subjected to the best random forest classification algorithm that included all elements as predictor variables. Distinct chemical signatures between fish of wild and hatchery origin supported 100% classification accuracies of known-wild, age-0 steelhead as wild origin for 19 of the 25 streams sampled. Misclassification of wild, age-0, and hatchery-origin fish, which tended to occur for streams that were located in close proximity to the hatchery, never exceeded 3.4% for a given stream. Our findings demonstrate highly successful discrimination of hatchery-origin and wild juvenile steelhead across a broad geographic range. Applying the classification algorithms developed herein to unknown-origin steelhead provides the ability to infer survival of year-classes from specific hatcheries. Further, the ability to differentiate hatchery- and wild-origin fish will assist in stock assessment efforts allowing for increased effectiveness of conservation and management of the species.
UR - http://www.scopus.com/inward/record.url?scp=85051665320&partnerID=8YFLogxK
U2 - 10.1002/nafm.10178
DO - 10.1002/nafm.10178
M3 - Article
AN - SCOPUS:85051665320
SN - 0275-5947
VL - 38
SP - 820
EP - 832
JO - North American Journal of Fisheries Management
JF - North American Journal of Fisheries Management
IS - 4
ER -