Designing monitoring programmes for detecting temporal trends in contaminants in fish and shellfish

We present some statistical tools for designing programmes for monitoring contaminant levels in fish and shellfish. In particular, we review the theory, and provide tables, for calculating the power of a programme to detect a linear trend in an annual contaminant index observed in T successive years. We suggest some simple methods for exploring how power is affected by changes in the sampling design, and in the variability of the contaminant data. These are demonstrated using mercury data from various sources.