Evaluation of the usefulness of bird feathers as a non-destructive biomonitoring tool for organic pollutants: a comparative and meta-analytical approach.

In this study, we investigated whether bird feathers can be used as a non-destructive biomonitor for organic pollutants. We analysed the outermost tail feathers of 8 terrestrial and aquatic bird species from Belgium (8 species, n=108) for polychlorinated biphenyls (PCBs), polybrominated diphenyl ethers (PBDEs) and organochlorine pesticides (OCPs). Every compound class could be quantified in one single tail feather of the birds under study (sum PCBs ranging from 5.5 to 510 ng/g feather, sum PBDEs from 0.33 to 53 ng/g feather, sum DDTs from 1.5 to 730 ng/g feather), except for PBDEs in feathers of the common moorhen (Gallinula chloropus). Further, we calculated Pearson correlations between concentrations of organic pollutants in feathers and concentrations in corresponding muscle or liver tissue from the birds. Correlations were found significant in half of the cases of the terrestrial species, but were found not significant for the aquatic species, with the exception of a significant correlation of sum PCBs in the common moorhen. Only for the common buzzard (Buteo buteo) (n=43) all correlations were found significant (0.32<r<0.77). In order to cope for low statistical power, we performed a meta-analysis on all bird species together. This led to significant correlations between levels in feathers and corresponding levels in muscle or liver for all terrestrial birds (p<0.05 in all cases, effect size 0.59 (p,p'-DDE) to 0.71 (Sigma PCB) for levels in feather and muscle). When correlations were recalculated excluding the birds that had died due to starvation, correlation coefficients for the terrestrial birds were found even higher (effect size up to 0.83 (Sigma PCB)). These results have important implications for non-destructive and retrospective biomonitoring. Although our results suggest that exact concentrations in the body cannot be predicted using feathers, bird feathers can give a good estimate of contamination levels in a population and as such are a potential non-destructive biomonitoring tool for organic pollutants. Outermost tail feathers and muscle tissue were also examined for nitrogen (delta15N) and carbon (delta13C) stable isotope content in the different bird species. However, delta15N/delta13C signatures in feather or muscle were not significantly correlated with the corresponding levels of organic pollutants in these tissues. Various confounding factors (such as habitat, condition, age, sex) may have masked a relationship between delta15N values and organic pollutant levels in the birds under study.