Effect of pooled sampling in the MILFERSK program

Abstract

The MILFERSK monitoring program is studying the occurrence and biomagnification of contaminants in freshwater ecosystems, mainly Lake Mjøsa. For the last 10 years, individual samples of muscle and liver from brown trout have been analyzed for a wide range of environmental contaminants. In this report we are statistically evaluating the effect of converting to pooled samples, i.e., 5 individuals constituting one composite sample. Main results indicate that analyzing pooled samples could mask or hide extreme concentrations, limiting the possibility to detect early occurrence in the environment. The statistical power will be weaker when studying biomagnification potential and time series, as fewer data points are included in the statistical models.