Using Machine Learning to Decipher Controls Water Quality

Water in the United States is a valuable resource. It is used from everything from drinking water to tourism. However, by the 1970s, many rivers and coastal areas were severely degraded from decades of using water to assimilate pollution. After billions of dollars spent and decades of effort on all levels of government, the increasing intensification of the agricultural system and the growing population is still putting U.S. water quality at risk.

Specifically, in many areas, nitrogen concentration trends are not responding to mitigation attempts and are remaining flat or continuing to increase. The overall objective of this work is to quantify water quality trends in rivers across the U.S. and determine how the landscape, climate, or land management may be controlling river nitrogen concentrations. We will be using novel methods in machine learning to find the patterns in river concentrations across the U.S. and determine what is controlling the concentration trends.

This information is crucial for those trying to improve water quality in the U.S. at a national or local level. The first step to addressing water quality issues is to understand how and why water quality has changed over time. TBC

Faculty Supervisor:

Nandita Basu

Student:

Partner:

University of Illinois Chicago

Discipline:

Engineering

Sector:

Environmental Science and Technology; Water; Other

University:

University of Waterloo

Program: