Contaminant biomonitoring programs in the GreatLakes region: Review of approaches and criticalfactors

摘要

Since the 1970s, several formal biomonitoring programs have been initiated to monitor temporal and spatialtrends of persistent, bioaccumulative, and (or) toxic chemicals (PBTs) in the Great Lakes. Researchers have used a varietyof approaches to deal with common issues that arise in these programs and there is considerable debate in the literature asto the most appropriate methods to use. Here, we critically review factors of importance in biomonitoring programs andcommon methods used for dealing with them under three main categories: organism-specific factors (lipid, age, size, sex,growth and bioenergetics, diet, and chemical biotransformation), study design (choice of tissue type, compositing, replicationand length of monitoring program, sampling frequency, selection of contaminants to include in monitoring programs, anduse of specimen banks for retrospective studies), and data analysis (accounting for changes in analytical methodology, treatmentof censored data, assessment of compound classes, comparison of empirical models, mechanistic models, and spatialassessments). We use data from the literature as well as longterm measurements of polychlorinated biphenyls (PCBs) in laketrout (Salvelinus namaycush) collected in Lake Ontario as part of monitoring programs run by Environment Canada, the OntarioMinistry of the Environment, and the United States Environmental Protection Agency to illustrate these factors. Wefind that, in general, there are several defensible methods, ranging from simple to complex, to deal with the issues consideredhere, with each having specific advantages and disadvantages. The optimal approach depends largely on the programobjectives, particularly if the results are meant for the public (understandable without a scientific background) or for researchpurposes (balance between complexity versus simplicity). Given that data analysis typically requires fewer resourcescompared to other aspects of monitoring programs, it may be feasible to use more than one data analysis approach to increasecredibility of the results and to improve comparability of data among studies. The importance of conducting preliminarysurveys and (or) pilot studies and regular review of ongoing programs (e.g., through a power analysis) is emphasized.