A interdisciplinary UWM research team, including Michael Nosonovsky, professor, mechanical engineering, collaborated to determine how to predict E. coli outbreaks at beaches more effectively.
Their goal was to identify the key sand conditions linked to E. coli growth and determine the best machine-learning method to predict these conditions.

Their findings were published in an article that appeared in a 2024 volume of the journal Surface Innovations.
E. coli bacteria are often used as indicators of water contamination, and elevated levels are one of the main reasons for beach closures. The researchers focused on how the physical and chemical properties of sand influence the survival and spread of these bacteria.
“Bacterial contamination of beach sand poses a public health problem,” Nosonovsky said. “So, it’s important to establish correlations between surface properties of sand and biocontamination, allowing us to predict and prevent the latter.”
Nosonovsky, whose expertise is tribology – the study of surfaces and friction – said that how sand interacts with water plays an important role in bacterial growth.
The work was conducted using data from Bradford Beach on Milwaukee’s lakefront. The researchers tested five different machine learning techniques and concluded the artificial neural network (ANN) technique outperformed other models in predicting E. coli concentrations.
The ANN model identified three critical factors predicting E. coli levels in sand:
- the state of sand, including parameters such as moisture content, pore size and the zeta potential – the difference in potential between a particle’s surface and the liquid it is suspended in.
- processing temperature
- the contact angle, which measures how easily water spreads across its surface. A special methodology was developed to measure the water contact angle of sand, Nosonovsky said.
The study shows the great potential of “tribo-informatics” – a new field that combines tribology, with data science and machine learning methods – to solve various problems, he said. The technique can be scaled up and applied to other beaches. Other team members were Md Syam Hasan (’22 PhD, mechanical engineering); Marcia Silva, former manager of UWM’s Water Technology Accelerator, and Alma Nunez.