Chlorinated dibenzo-p-dioxins and dibenzofurans in the Baltic herring and sprat of Estonian coastal waters

Background, Aims and Scope. The concentration of chlorinated dibenzo-p-dioxins and dibenzofurans in many fish from the Baltic requires monitoring, since it approaches or exceeds the European Union threshold limit value of 4 pg TEQ/g wet weight of fish for human consumption. The concentrations, expressed in TEQs, are important for toxicology and regulations, but hide the concentrations of the individual congeners, which are important for other environmental sciences, source allocation, and for the detection of measurement errors. This report evaluates the results of a survey reported earlier only in the terms of the TEQ concentrations. Methods. Principal Component Analysis (PCA) was used to reduce the dimensions of the data (17 = 7 chlorinated dibenzo-p-dioxin and 10 chlorinated dibenzofuran congeners) to three principal components. This facilitated the interpretation of the congener profiles. Slopes of the congener concentrations as a function of age of the fish were estimated by least squares regression. The results were compared with a large set of data for lake trout from Lake Ontario. Results and Discussion. The congener profiles of Baltic herring are less scattered than those of sprat. The profiles of herring from the central Baltic differ from those of herring from the Gulf of Riga and both appear to be affected relatively minimally by the age of the fish. The congener profiles of herring from the western Gulf of Finland are similar to those from the central Baltic, those from middle Gulf of Finland are similar to those from the Gulf of Riga. Both seem to be more affected by age of the fish than the profiles of the first two groups. The concentrations of several pentachloro- and hexachloro-dibenzo-p-dioxins and dibenzofurans increase with the age of the fish. Conclusion. PCA is a good technique for the evaluation of the congener profiles. The resulting loading and score plots provide a good graphic summary of the multidimensional data. Additional analyses are needed to confirm the observed profile patterns. A comparison with the results of a long-term monitoring from another area shows that the age of the fish is a more important factor than the year of capture. Recommendation and Outlook. The surveys should continue for a number of years and the results should be presented and evaluated both in terms of the TEQs as well as in terms of weight concentrations. Since the concentrations do not appear to change very much from year to year, it would be better to carry out surveys only every 3-4 years and, instead, stratify the sampling according to age and gender of the fish, and to analyze replicate extracts by replicate measurements. The inclusion of unmarked replicate samples would be a good quality assurance measure. It would be desirable to analyze additional parts of the food chain in order to understand the fate of the compounds in the ecosystem.