High-risk contaminants, OCPs and PCBs, were investigated in marine fish from the Adriatic Sea, from which retail fish in Croatia is commonly sourced. The pollutant levels in sardine, horse and chub mackerel, anchovy and round sardinella were analysed based on a two-year sampling and the joint use of generally accepted statistics and advanced clustering methods - self-organizing maps (SOM) and decision tree analysis (DT). Both the SOM and DT suggested fish mass and length rather than fat along with α-HCH, p,p'-DDT, PCB-74 and PCB-189 to cause variable pollutant uptake among species. Main distinctions of sardines occur in coastal and off coast regions rather than in a particular fishing zone and they are associated with both fish characteristics, levels of γ-HCH and PCBs: -60, -105, -150, -170, and -189. The results, mutually compatible or in agreement, could be useful for the design and implementation of the abatement strategies of fish pollution.
Keywords: Advanced classification methods; Marine fish; Organochlorine pesticides (OCPs); Polychlorinated biphenyls (PCBs).
Copyright © 2018 Elsevier Ltd. All rights reserved.