Detection of statistically significant trends in atmospheric concentrations of semivolatile compounds

This paper examines the sampling and analytical requirements of observing s tatistically significant temporal trends in atmospheric concentrations of s emivolatile organic pollutants. Data precision and study duration are found to be the most important factors that are subject to experimental control. We define a detectable half-life as the longest half-life for which statis tical significance can be observed for a given length of study and show how this can be calculated. Finally, we show how to predict the length of a st udy that will find a statistically significant half-life. These concepts ar e applied to a recently updated temporal trends study from the Integrated A tmospheric Deposition Network. Updated ha If-lives for gas-phase pesticides at three U.S. sampling sites are given. There are now enough data to obser ve statistically significant temporal trends for most of the gas-phase pest icides except for alpha-chlordane and trans-nonachlor at Eagle Harbor. Usin g the concept of detectable half-life and based on observations of half-liv es at other lADN sites, at least 2 more years of study are required to obse rve statistically significant trends for alpha-chlordane and trans-nonachlo r. The concepts presented in this paper can also be applied to the design o f sampling strategies for measuring long-term temporal trends in media othe r than air.