The power of a contaminant monitoring programme to detect linear trends and incidents

Abstract

Summarizing the temporal variation in an annual series of contaminant levels, for example as a downward or upward trend, is often complicated by the presence of an additional component of random year-to-year variation. This paper describes appropriate statistical tests for detecting two types of systematic variation, a linear trend and an incident in the presence of random between-year variation. Further, the effectiveness of a monitoring programme to detect these changes is quantified in terms of its statistical power. The results are used to assess the power of the International Council for the Exploration of the Sea’s Co-operative Monitoring Programme of contaminant levels in fish muscle. © 1993 Conseil International pour l’Exploration de la Mer.