Tracing PCB contamination in fish using shadow profiles and self-organizing maps

We traced the origins of contamination with polychlorinated biphenyls (PCB) in fish by mapping PCB profiles, whereby profiles represented congener fractions of the total ICES-7-PCB concentrations. ‘Shadow profiles’ were modelled using published profiles from potential source countries and by representing runoff driven as well as atmospheric transport pathways into fish habitats. Subsequently, fish profiles from two international datasets were matched to shadow profiles using self-organizing maps (SOM) to uncover source to sink relationships. Fish caught in inland waters were predominantly mapped to shadow profiles with a runoff driven contamination history; but shadow profiles representing atmospheric contamination pathways also played a role. In contrast, marine fish from the NE Atlantic matched contamination mainly via atmospheric pathways linked to emissions from major European industrial nations such as Germany, Spain, and United Kingdom. Fish from the Baltic Sea reflected a more complex contamination history, with exposure to both runoff driven and atmospheric pathways from surrounding countries. While most source to sink relationships were consistent, some inconsistencies highlighted the need to refine the modelling of shadow profiles. Overall, the SOM technique was effective in grouping a large number of shadow profiles into a manageable array of units, onto which fish profiles could be easily mapped.