Developing alternative approach to assess stream health, use data more efficiently
Pouyan Nejadhashemi, an associate professor in the Department of Biosystems and Agricultural Engineering (BAE) at Michigan State University (MSU), is developing models that help improve stream health in Michigan.
January 19, 2016 - Author: Cameron Rudolph
Roughly 2.5 percent of global water is fresh water. Only 1.2 percent of that is surface water such as rivers, lakes and streams. With more than 7 billion people around the world relying on fresh water for drinking, agriculture, recreation and more, the need to preserve this vital resource is greater than ever. In the United States, the Environmental Protection Agency (EPA) passed the Clean Water Act in 1972 in an effort to “restore andmaintain the chemical, physical and biological integrity of the nation’s waters.” The new law established protocols for regulating
water pollution, and although more than 40 years have passed, water quality improvements are moving slowly.
The 2011 EPA Biological Assessment revealed that nearly 42 percent of U.S. streams are in “poor” biological condition, which is measured by the health of native fish and invertebrate populations. Only 53 percent were determined to be in “fair” or “good” condition. The remaining 5 percent have not been assessed. But there are significant factors of stream health that are ignored in these figures. Are enough individual sites being monitored to accurately gauge aquatic ecosystems? How will climate change affect fresh water resources? How can this information be pieced together to inform policy decisions?
Pouyan Nejadhashemi, an associate professor in the Department of Biosystems and Agricultural Engineering (BAE) at Michigan State University (MSU), believes that one of his research projects may hold the answers.
“It’s simply not feasible from cost and resource perspectives to regularly monitor thousands of individual locations on each body of water,” Nejadhashemi said. “That would be a waste of resources, particularly for some areas that may be in good condition. So we needed to develop an alternative approach that takes into account hundreds of variables to assess stream health
condition. Then we can use that data to make better decisions, focusing limited resources on the areas of greatest need.”
Nejadhashemi and his research group — including Matt Einheuser, Matthew Herman and Sean Woznicki — have developed models that factor in aquatic life, soil, land use, climate change, erosion, plant growth and many other variables. Historically, models have done a poor job at considering several variables concurrently, leaving researchers to make generalizations based on
relatively small amounts of data.
“There have been significant knowledge gaps in determining stream health in the past, and we used the wrong criteria to assess stream health,” Nejadhashemi said. “I think of it like going to the doctor for a health checkup. The doctor performs several tests to determine if you’re healthy. You can’t simply evaluate your overall health by checking only your eyesight. That’s what we’ve been doing with stream health historically, so we’ve needed to use new techniques along with several indicators to measure stream health more accurately.”
Through collaborations with the Michigan Department of Natural Resources and the Great Lakes Commission, biological data was collected from streams in Flint, Muskegon and all the way to the Upper Peninsula. These rich datasets were incorporated into the stream health models using fuzzy logic techniques, which are based on “degree of truth” rather than the absolute truth value.
“Stream health is a complex issue that is nonlinear in nature,” Nejadhashemi said. “We can use fuzzy logic to add a linguistic interpretation to stream health conditions, which helps create more easily understandable results.”
Fuzzy logic allows the research team to create if-then statements. For example, if pollutant concentration is high, then stream health is poor. Using thousands of individual data points across hundreds of locations, this simplification proves invaluable. After introducing future climate scenarios into the stream health models, the researchers then develop risk maps showing which sections of water are in greatest need of remediation. Researchers and outreach professionals can use these maps to inform policy that utilizes resources efficiently.
“These are extremely valuable tools for all stakeholders engaging in water resources planning,” Nejadhashemi said. “At this point, we may be spending time and money on areas that may not be in critical need of attention. But with the added information from our models, we have a better chance at making evidence-based recommendations. We would like to expand this project beyond Michigan to be a part of comprehensive national policy for protecting fresh water resources, which not only focuses on water quality but also natural habitats.”
This project is funded by the U.S. Department of Agriculture National Institute of Food and Agriculture and MSU AgBioResearch.