Evaluating the impact of sampling design on drinking water quality monitoring program outcomes

Drinking water suppliers around the world are required by regulations to sample and test water quality in their distribution systems with the intention of generating information that can be used to protect human health. Requirements for where samples must be collected can vary and guidance on how to select locations to meet these requirements is limited. There is a need to better define and evaluate the meaning of representative in the context of selecting locations for collecting grab samples while considering the regulatory context and resources available to suppliers. We develop sampling programs for monitoring disinfectant residual concentrations and compare the results and efficacy of these programs using synthetic water quality data for two example distribution systems. Results showed that achieving accurate estimates of worst quality conditions in a network was more likely than achieving accurate estimates of systemwide conditions, and that using constant locations makes it difficult to accurately represent systemwide conditions. Results also showed that timing of sample collection is likely important depending on factors such as consumer demand patterns, and that it is inappropriate to make conclusions about systemwide conditions when few samples are taken over an evaluation period. This research has implications for developing recommendations that water suppliers and regulators can use to design and evaluate effective sampling programs.