Guiding fish consumption advisories for Lake Ontario: A Bayesian hierarchical approach

Article history: Received 20 July 2015 Accepted 13 November 2015 Available online 8 January 2016 Communicated by Craig Stow Weevaluate the temporal trends of totalmercury (THg) andpolychlorinated biphenyls (PCBs) inwalleye (Sander vitreus) and lake trout (Salvelinus namaycush) based on approximately 40 years of contaminant data from different locations in Lake Ontario. Bayesian inference techniques are employed to parameterize four hierarchical models. Our analysis provides evidence of distinctly declining trajectories for the two contaminants in lake trout. Likewise, walleye demonstrate a decreasing PCB trend, whereas no distinct temporal shifts were found in their THg rates of change.We illustrate the capacity of our statistical framework to aid in formulating fish consumption advisories by generating customizable probability of exceedance of THg and PCB threshold human exposure levels, based on their tolerable daily intake values. Walleye consumption results in 30% lakewide exceedance frequencies of the THg threshold for the sensitive demographic group of children less than 15 years old with an average body weight of 50 kg. Lake trout PCB threshold is frequently exceeded (N80%) in all of the study sites, whereas exposure to THg through lake trout consumption appears to be within the acceptable levels for human health. The overall trends indicate that the reduced contaminant emissions have brought about positive changes in the fish contamination levels in Lake Ontario. © 2015 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved. Index words: Organochlorines Mercury Fish consumption advisories Lake Ontario Bayesian hierarchical modeling