I guess they are not using bioinformatics. To monitor drinking water quality, for example, I cannot think of an analysis where bioinformatics methods could be used.

Here is an interesting case where bioinformatics could have been useful in hindsight: http://uni.no/en/uni-health/research-unit-for-general-practice-in-bergen/the-giardia-outbreak-in-bergen-and-long-term-consequences/

In Bergen, Norway there was a huge outbreak of diarrhoea caused by the parasite Giardia lamblia in 2004. The water reservoir “Svartediket” was the source, supplying central parts of the city with drinking water. This was the first major outbreak with this parasite in Norway, and it is among the largest outbreaks in the world. Around 1250 individuals had laboratory confirmed infection in stool samples after seeing their general practitioner.

You would not necessarily expect bad water quality in Norway. Anyway, if you coul get any significant amount of DNA sequence for a meta-genome, something must have gone really really wrong.