Multivariate data analysis of organochlorines and brominated flame retardants in Baltic Sea guillemot (Uria aalge) egg and muscle

Concentrations of organochlorines (OCs) and brominated flame-retardants (BFRs) were determined in guillemot (Uria aalge) eggs from the island of Stora Karls? in the Baltic Sea where 10 eggs/year were collected in the years 2000, 2001 and 2002. The dominating contaminant in egg was p,p'-dichlorodiphenyldichloroethylene (p,p'-DDE) with a geometric mean (GM) concentration of 18200 ng/g lipid weight (lw). The GM concentration in egg of hexabromocyclododecane (HBCD) of 140 ng/glw, was significantly higher than that of polybrominated diphenyl ether (summation operatorPBDE) of 77 ng/glw. For the evaluation of the data multivariate data analysis techniques namely principal components analysis (PCA), partial least squares regression (PLS), soft independent modelling of class analogy (SIMCA classification), and PLS discriminant analysis (PLS-DA), were used. We investigated whether the eggs' biological characteristics co-varied with egg concentrations of OCs and BFRs, and found e.g., significant negative correlations between egg weight and concentrations of HCB and p,p'-DDE. A PLS model with analyzed BFRs as the Y matrix and OCs as the X matrix could, with varying accuracy, calculate the concentrations of BFRs in the individual egg from their concentrations of OCs (e.g., R(2)Y of 0.89 for BDE47, and of 0.50 for HBCD). Lastly, we compared the contaminant concentrations in the eggs to those in previously analyzed pectoral muscles from adult guillemots from Stora Karls?, from the year 2000. A PLS-DA model, showed that some of the contaminants (e.g., HBCD and CB28) had significantly higher concentrations in egg than in muscle, although 7 of the 14 contaminants showed no difference in concentrations between the two matrices.