USDA grant to fund project to develop AI-powered water quality database
The study focuses on the Upper Mississippi River basin, the Ohio River basin, and the Chesapeake Bay basin.
The backbone of the system is a large network of nitrate sensors like this one currently deployed in a small stream in an agricultural area, providing high-frequency measurements of water chemistry. Monitoring is done by computers that use artificial intelligence to detect complex patterns and anomalies that could indicate pollution. (Source: Penn State)
UNIVERSITY PARK, Pa. — Nitrate is a common chemical compound that occurs naturally and is found in plants, water and soil. However, it can break down into molecules that are harmful to human, animal and environmental health and accumulate as a pollutant. Nitrate pollution of rivers, lakes and estuaries is a serious problem in many agricultural watersheds, yet water quality data is limited, making it difficult to monitor water health and make management decisions, according to Penn State researchers. To improve available data, the U.S. Department of Agriculture (USDA) has awarded a four-year, $650,000 grant to a Penn State research team.
The study focuses on the upper Mississippi River basin, the Ohio River basin, and the Chesapeake Bay basin. The award, administered by the USDA’s National Institute for Food and Agriculture, funds a new approach to studying nitrate concentration dynamics. The proposed system will use deep learning – a subset of machine learning and computer science, and a form of artificial intelligence (AI) – to make sense of the massive amount of nitrate data collected.
The AI will identify complex patterns and anomalies in the flood of data and create a comprehensive nitrate database that will allow resource managers to quantify hotspots of nutrient pollution, said team leader Cibin Raj, associate professor of agricultural and biological engineering in the School of Agricultural Sciences. With this knowledge, resource managers can determine exactly where to implement conservation measures.
“By advancing modeling, mapping and measurement using deep learning models and high-frequency sensors, managers can identify nitrogen sources and sinks and determine locations where implementing conservation measures will provide the best return on investment,” he said.
High-frequency sensors are changing the way scientists monitor and manage water quality, notes Jonathan Duncan, associate professor of ecosystem science and management, who is co-leader of the project. These sensors, he explained, provide unprecedented insight into time series of nitrate concentrations, enabling a better understanding of fertilizer and manure application, seasonal and rainfall events, and the size and intensity of storm events.
“Our team will develop a hybrid, integrated modeling framework that includes field data collection, flow sensor data, and machine learning interpretation to understand the nitrate dynamics of a region and continuously generate daily nitrate concentration data,” Duncan said. “The information generated can be used to assess ecosystem health and develop more effective and targeted watershed management strategies.”
The researchers will make the data publicly available and the results will be presented to local decision makers, watershed planners and conservation district staff through a well-developed collaboration with Penn State Extension.
Chaopeng Shen, professor of civil and environmental engineering, will also contribute to the project. The research was initially funded by the Penn State Institute of Energy and the Environment through its seed grant initiative.
–Jeff Mulhollem, Penn State University
