Conventional approaches to water quality testing rely on sophisticated manually-read sensors and methods that sample at a handful of locations. Startup company Varuna instead uses "less costly sensors that measure fewer contaminants but touch more parts of a system." By feeding information about a water system’s architecture and history into artificial intelligence algorithms, the company then seeks to provide a utility with more detailed information on the contaminants in water." (GovTech) This follows a "recent trend in air quality monitoring, the idea is that instead of relying on a few sophisticated sensors, water utilities can rely on many cheap sensors that provide less detailed data. Rather than measuring 25-plus different contaminants, Varuna’s sensors simply tell utilities about the level of organic material versus metal." (GovTech)
This approach points towards a broadly emerging strategy in urban environmental and ecosystem sensing—cheap, often power-harvesting sensors, providing reach streams of lower-quality data that can be synthesized to supplement and even replace older more detailed but less continuous or widely-sampled data streams.