A team at the University of Colorado Boulder is testing a new type of water sensor in Boulder Creek that is helping to monitor drinking and recreational water quality in Colorado and around the world.

The sensors, which the team has developed and commercialized, can accurately measure and predict risk levels of microbial concentrations in bodies of water.

In addition to Boulder Creek, the sensors have been or soon will be deployed in the Chicago River, the Seine River in Paris and the River Thames in London. The goal of the sensors is to provide cities and companies with more frequent assessments of water quality.

“We want to help make sure people are recreating and swimming and tubing in clean water, we want to make sure people have clean water to drink,” CU Boulder professor and team lead Evan Thomas said.

The sensors will also be tested in Denver, Kenya and Rwanda. NASA is also interested in using the sensors for projects related to recycling the water astronauts use in space.The sensors were developed in partnership with the Fort Collins-based sensor company, In-Situ, and tested in the Poudre and Yampa Rivers in Colorado for the past two years. The sensors provide ongoing, in-stream water quality measurements that allow the CU Boulder team to track any changes. Cities and water utility companies are utilizing the sensors but not relying on them entirely yet, since the sensors are still in the validation phase.

In Boulder, the city is required to monitor for E. coli in Boulder Creek because part of the creek is not in compliance with regulations set out by the Environmental Protection Agency. Boulder Creek has a total maximum daily load for E. coli, which is the maximum amount of a pollutant the creek can have while still meeting water quality standards. If the amount of E. coli in the creek exceeds a certain level, it’s not safe to swim and tube in.

Boulder’s Urban Water Quality Program Coordinator Michael Lawlor said the city has been monitoring bacteria in Boulder Creek for about 30 years. He said the city was interested in the CU Boulder team’s sensors because they give Boulder a real-time look at what’s happening in the stream.

Traditional water quality monitoring methods take 24 hours to come back with results, Lawlor said. The new sensors, which can provide data every hour, were installed in Boulder Creek this spring.

The fluorescence-based sensors use a protein that’s highly correlated to E.coli in the water, and the team looks for strength correlated to E.coli based on the fluorescence. This, combined with machine learning analytics, allows for estimations of E. coli contamination with remote reporting.

The CU Boulder team is testing the sensors in various locations to ensure the readings are accurate. In Boulder Creek, the team has four sensors: one near Eben G. Fine park, another adjacent to CU Boulder and one at 30th Street and at 55th Street. These are four of the same six locations the city is obligated to measure.

Once the technology has been trained and verified, Lawlor said, he could see the sensors being used as a tool to keep the public informed about E. coli levels in the creek.

“Having that type of technology here in our backyard would be really helpful to help our public make informed decisions about when they want to recreate in Boulder Creek,” Lawlor said. “And this sensor is able to accommodate that at a much lower price (compared to other similar sensors on the market).”

Whitney Knopp is an environmental engineering doctoral student whose degree program is focused on applying and validating the sensors. In early June, she flew to Paris to install the sensors in the Seine River.

“A big problem with microbial contamination is that the way you get data is from sampling, and that sampling often happens pretty infrequently, and then you need to take the sample back to the lab and do analysis on it,” Knopp said. “So you’re getting a single data point at one point in time retroactively, so there’s not much you can do when you see contamination in your water, because it might have already passed or those samples are often not necessarily representative of what water quality is like.”

She works on developing the machine learning model, determines the level of sensor accuracy and evaluates and investigates different use cases.

“The technology itself is very promising and I think it’s fascinating,” Knopp said.

“The thing that I think is the most valuable and important to me is that the motivation behind this is really to improve water quality both for drinking water applications … and then surface water applications for recreation and the health of the environment.”

In Kenya and Rwanda, the sensors are placed in water filters and pumps owned by nongovernmental water providers to help make sure the water is safe to drink. Denver Water, a utility company, utilizes the sensors amid concerns about algae blooms that can be risky for dogs and people swimming in reservoirs.

Thomas said everyone deserves clean water, whether they live in Boulder or Rwanda.

“We need more monitoring, we need more technology and we need more modeling to help the public decision-making and the public’s demand for clean water,” Thomas said.