Statistical Assessment of Tissue Residues in Sport Fish and Indicator Fish Species in the Detroit River to Address Fish and Wildlife Consumption

...

The Delisting Criteria identified in the 2010 Stage 2 RAP report is written as follows:

“When consumption advisories for indicator fish species (e.g. walleye, brown bullhead, and smallmouth bass) given for the sensitive population in the AOC are similar to upstream and downstream non-AOC Great Lakes reference areas” (Green et al. 2010)

For purposes of this analysis, the upstream and downstream non-AOC Great Lakes reference areas were defined as Lake St. Clair and western Lake Erie, respectively. Data on priority contaminants in the three sport fish indicator species from the Detroit River Area of Concern (AOC) and reference areas were compiled using data sets generated by the Ontario Ministry of Environment and Climate Change’s sport fish monitoring program and supplemented, as required, with additional samples generated by the Great Lakes Institute for Environmental Research, University of Windsor. The contaminants of consideration focused on PCBs and mercury in fish tissues which are responsible for the majority of fish consumption restrictions generated for Ontario waters of the AOC.

The main objective of this report was to test the delisting criteria as specified in the Stage II RAP delisting criteria for each indicator species. A secondary objective of the report was to identify data gaps, data inadequacies or confounding factors that may interfere with the evaluation of delisting as stated in the criteria and to provide recommendations for data needs to support a completed delisting assessment.

Methods

Data on priority contaminants in sport fish samples were obtained from the Ontario Ministry of Environment and Climate Change (OMOE CC) sport fish contaminant database used to generate sport fish consumption advisory information for Ontario waters. The priority chemicals responsible for the majority of Ontario fish consumption advice in the Detroit River are listed as Dioxin, furans and dlPCBs (80%) and mercury (20%). However, OMOE CC no longer routinely analyzes for dioxins, furans and dlPCBs in sport fish to generate sum toxic quotients (TEQ) for dioxin like compounds. Instead, OMOE CC has adopted a computational algorithm to estimate dioxin like TEQ values based on measurements of total PCBs generated for fish (Bhavsar et al. 2007). As such, the data query compiled all relevant data on total Hg and total PCBs in sport fish. The query was restricted to sport fish samples collected from Canadian waters of Lake St. Clair, Detroit River (upper and lower fishing zones) and western Lake Erie over the time period of 2005-2012. The time period was restricted to limit temporal trends contributing to complications in the analysis. The collection locations correspond to the normal fishing zone boundaries over which Ontario fish consumption advice is provided. All data collected correspond to skinless dorsal muscle fillet samples. Mercury concentration data are expressed in units of ug/g wet tissue weight. PCB concentrations are expressed in units of ng/g wet tissue weight.

The delisting criteria for sport fish consumption advisories in the Detroit River identifies brown bullhead, smallmouth bass and walleye as indicator species. Thus, for these species, additional datasets (OMOE CC database for the Detroit River 2000-2004) and GLIER (University of Windsor) contaminant data) were compiled to address data gaps or low sample sizes for these species where required to facilitate delisting criteria evaluation. Both sediment and water chemistry in the Detroit River have been shown to remain stable since 1999. Thus supplementation of fish data with earlier data sets was considered sound, whereas sport fish data in both Lake St. Clair and western Lake Erie have shown temporal changes over the 2000-2012 time period.

Fish contaminant concentrations in samples were evaluated against the tolerable daily intake (TDI) criteria for the sensitive sub-population used to specify fish consumption advice by OMOE CC. The TDI values below which fish consumption advice are not warranted are 0.26 ug/g wet weight and 105 ng/g wet weight for mercury and PCBs, respectively. The TDI values resulting in the most restrictive advice information possible (i.e. above which a recommendation of no consumption of fish of that species and length for the sensitive sub-population is issued) are 0.52 µg/g wet weight and 211 ng/g wet weight for mercury and PCBs, respectively. Samples having concentrations between these TDI thresholds would be issued intermediate restrictive advice, specifying between 2 to 8 allowable fish meals per month. It should be noted that OMOE CC has harmonized the TDI criteria between PCBs and dioxin-like compounds. Thus, the TDI’s of 2.7 pg/g wet wt and 5.4 pg/g wet wt. for dioxin-like compounds correspond to total PCB concentration values of 105 ng/g and 211 ng/g, respectively. Throughout the document, summary statistics and statistical analyses were completed on total PCBs instead of dioxin like TEQs. However, as indicated above, the analysis on total PCBs is directly applicable to TEQ data, both in terms of spatial trends and frequency of exceedence of TDI values. For each indicator species, the percentage of samples falling into three categories of advice information: i) unlimited consumption, ii) restricted consumption and iii) most restrictive advice (recommended no consumption)) were assessed.

Statistical analyses were conducted to evaluate spatial and temporal trends in the data for each indicator species as well as determine the effect of biological variables (body length and sex) on contaminant residues. Mercury and PCBs data were log transformed since non-transformed data were found to be non-normally distributed but were normal following log transformation. Statistical tests used included ANOVA, ANCOVAs and linear regressions to uncover differences in contaminant residues between sample locations and to account for differences in biological variables (body size and sex) among samples from the different locations. Where there were non-significant relationships between contaminant residues and body size for a given fish population, the geometric mean concentration and 90% confidence interval was generated to compare with TDI values. In cases where fish length was found to be a significant factor, linear regressions (log concentration vs body length) were used to establish the mean fish length at which TDI criteria were exceeded.

RESULTS

Table 1 provides a summary of available data for PCB and mercury residues in various sport fish species from the OMOE CC database over the query period. Among the three indicator species identified by the Detroit River Fish Consumption Advisory delisting criteria,