May 14, 2020
NSF grant to assist in improving drainage flow path mapping
CARBONDALE, Ill. — Researchers at Southern Illinois University Carbondale recently garnered a National Science Foundation grant that will advance the modeling and mapping of overland drainage flow paths.
Ruopu Li, an assistant professor, and Guangxing Wang, a professor, both in the School of Earth Systems and Sustainability, connected with Banafsheh Rekabdar, an assistant professor in the School of Computing to create a project that will model and map overland drainage flow paths widely found in agricultural landscapes.
The approach is unique because it steers past previous limitations of evaluating the flow and accumulation of rainfall runoff, said Li, who has been with SIU since 2015.
“Such efforts have been commonly affected by locations of hydraulic drainage structures, such as road culverts and bridges, which were falsely represented as ‘virtual’ flow barriers in digital elevation models, or DEMS,” Li said.
A step forward with ‘deep learning’
“Presently, the common practice is to rely on labor-intensive on-screen digitization to mark those culvert and bridge locations. As a departure, this project will establish a model based on artificial intelligence and deep learning to identify the locations of anthropogenic barriers. The project outcome is expected to benefit a broad scope of natural resources management activities, such as watershed nutrient control, wetland conservation, and aquatic species protection.
Li’s research interests include Geographic Information Systems (GIS), remote sensing, water resources planning and management, social media, and land use. Rekabdar’s research expertise is in artificial intelligence, machine learning, and deep learning.
The $170,718 NSF grant for “Enhancing High-resolution Terrain Data Model for Improving the Delineation of Multi-scale Hydrological Connectivity” came in late March.
Work began a little more than a year ago with Li, who earned his doctorate from the University of Nebraska-Lincoln, studying GIS, water resources, and land use as well as with Rekabdar, who earned her Ph.D. from the University of Nevada-Reno, studying artificial intelligence and data sciences. Wang, a professor in Geography and Environmental Resources, is also involved in the research effort.
Making connections and progress
Rekabdar noted that the research presents numerous opportunities for emerging real world applications and makes fundamental contributions in deep learning algorithm design and the geography areas.
“It’s been a great collaboration,” said Rekabdar. “This project proposes novel deep learning/machine learning methods to improve identifying the locations of man-made barriers. We’ve had some great preliminary results to detect drainage crossing locations as we extended the results from Nebraska to Illinois.”
Li and Rekabdar agree that the project is a success so far due to its interdisciplinary roots.
“These days, collaboration projects are more successful,” Li said.