Water and carbon are key resources people rely on every day, from the availability of drinking water and preventing floods to storing carbon in forests and wetlands.
The National Science Foundation has awarded a $1.79 million grant to a University of Florida research team to develop artificial intelligence-powered tools that could help communities better understand these connections between water and carbon.
The project will use AI to predict how changes in land use, water availability and climate could affect these systems. Researchers will leverage UF’s HiPerGator supercomputer to process large amounts of environmental data and run the models needed to generate these forecasts.
The goal is to help land and water managers, policymakers and other decision-makers better understand the potential outcomes of their choices before making land-use decisions that could affect communities and the environment for years to come.
“Carbon and water are closely connected, and together they influence vital benefits people rely on every day. Our project will use AI to better predict how these important carbon and water functions and their interlinkages change from one place and scale to another,” said Jiangxiao Qiu, principal investigator on the project and an associate professor of landscape ecology at UF/IFAS Fort Lauderdale Research and Education Center and faculty at the UF/IFAS School of Forest, Fisheries, and Geomatics Sciences.
He is joined by project co-PI Matt Cohen, director of the UF Water Institute. The Water Institute’s involvement brings together expertise in water resources with researchers working across disciplines to the table.
“Natural resource problems are inherently interdisciplinary. Water Institute leadership brought together a team that spans computer science, remote sensing and ecology, and that connects long-standing resource management questions to the modern tools of machine learning,” he said. “In doing so, this project is a hallmark example of why UF is so well positioned to lead the nation in science to support solutions to complex shared natural resource challenges.”
Additional team members are Joel B. Harley, associate professor in the Department of Electrical and Computer Engineering, Emmanuel J. Dorley, assistant professor in the Department of Computer & Information Science & Engineering and Chang Zhao, assistant professor in the Department of Agronomy.
The researchers will also develop an interactive online platform that will allow users to explore different “what if” scenarios and see how different choices could affect a landscape.
For example, users could explore what might happen if an area is reforested, a wetland is restored, water flows are changed or land is developed. The tool could show how those changes may affect water resources, carbon storage and other benefits that healthy landscapes provide.
The goal is to help decision-makers compare different options and identify potential problems before they happen.
Another key part of the project is something researchers call physics-informed AI. In simple terms, this means teaching AI to learn from large amounts of environmental data while also following the basic physical rules that govern how the natural world works.
For example, the AI will learn from observations and existing computer models of water and carbon while also accounting for principles such as the conservation of water and energy. This can help make predictions generated by AI easier for researchers and decision-makers to understand.
“Better predictions can help reduce environmental risks related to water shortages or excess water, loss of carbon storage, and changes in how well landscapes continue to provide these critical benefits,” said Qiu. “Economically, that can mean using limited resources more efficiently and effectively, avoiding costly management mistakes and making better-informed investments in resilient landscapes and sustainable communities.”
The researchers plan to make several resources publicly available, including datasets on water and carbon, computer modeling tools and an interactive decision-making platform.
These data and tools will be available to researchers, land and water managers, policymakers, Extension professionals and students. The project also will provide training opportunities for students and researchers working at the intersection of artificial intelligence and environmental science.
##
By Lourdes Mederos, rodriguezl@ufl.edu
ABOUT UF/IFAS
The mission of the University of Florida Institute of Food and Agricultural Sciences (UF/IFAS) is to develop knowledge relevant to agricultural, human and natural resources and to make that knowledge available to sustain and enhance the quality of human life. With more than a dozen research facilities, 67 county Extension offices, and award-winning students and faculty in the UF College of Agricultural and Life Sciences, UF/IFAS brings science-based solutions to the state’s agricultural and natural resources industries, and all Florida residents.
2
