Online microbial monitoring of drinking water: How do different techniques respond to contaminations in practice?

Safeguarding the microbial water quality remains a challenge for drinking water utilities, and because of population growth and climate change, new issues arise regularly. To overcome these problems, biostable drinking water production and water reuse will become increasingly important. In this respect, high-resolution online microbial monitoring during treatment and distribution could prove essential. Here, we present the first scientific and practical comparison of multiple online microbial monitoring techniques in which six commercially available devices were set up in a full-scale drinking water production plant. Both the devices' response towards operational changes and contaminations, as well as their detection limit for different contaminations were evaluated and compared. During normal operation, all devices were able to detect abrupt operational changes such as backwashing of activated carbon filters and interruption of the production process in a fast and sensitive way. To benchmark their response to contaminations, the calculation of a dynamic baseline for sensitive separation between noise and events is proposed. In order of sensitivity, enzymatic analysis, ATP measurement, and flow cytometric fingerprinting were the most performant for detection of rain- and groundwater contaminations (0.01 - 0.1 v%). On the other hand, optical classification and flow cytometric cell counts showed to be more robust techniques, requiring less maintenance and providing direct information about the cell concentration, even though they were still more sensitive than plate counting. The choice for a certain technology will thus depend on the type of application and is a balance between sensitivity, price and maintenance. All things considered, a combination of several devices and use of advanced data analysis such as fingerprinting may be of added value. In general, the strategic implementation of online microbial monitoring as early-warning system will allow for intensive quality control by high-frequency sampling as well as a short event response timeframe.