Determining optimal sampling strategies for monitoring mercury and reproductive success in common loons in the Adirondacks of New York.

The common loon (Gavia immer), a top predator in the freshwater food web, has been recognized as an important bioindicator of aquatic mercury (Hg) pollution. Because capturing loons can be difficult, statistical approaches are needed to evaluate the efficiency of Hg monitoring. Using data from 1998 to 2016 collected in New York's Adirondack Park, we calculated the power to detect temporal changes in loon Hg concentrations and fledging success as a function of sampling intensity. There is a tradeoff between the number of lakes per year and the number of years needed to detect a particular rate of change. For example, a 5% year change in Hg concentration could be detected with a sampling effort of either 15 lakes per year for 10 years, or 5 lakes per year for 15 years, given two loons sampled per lake per year. A 2% year change in fledging success could be detected with a sampling effort of either 40 lakes per year for 15 years, or 30 lakes per year for 20 years. We found that more acidic lakes required greater sampling intensity than less acidic lakes for monitoring Hg concentrations but not for fledging success. Power analysis provides a means to optimize the sampling designs for monitoring loon Hg concentrations and reproductive success. This approach is applicable to other monitoring schemes where cost is an issue.