NSF grant to support development of cyberinfrastructure tools for precision agriculture

Modern precision agriculture requires an understanding of how climate-related factors such as soil moisture, precipitation, and temperature impact agricultural productivity.

“As we enter an era of growing environmentally relevant data that can, for example, drive water management practices, new cyberinfrastructure tools and big data analytics are needed to extract knowledge and value-added products from the data,” says Michela Taufer, professor of computer science at the University of Delaware.

Taufer, who has already brought her knowledge of data science to the field of medicine through collaborations with clinicians, is now teaming with ecosystem ecologist Rodrigo Vargas, associate professor in UD’s Department of Plant and Soil Sciences. The two recently received a three-year, $500,000 grant from the National Science Foundation to develop cyberinfrastructure tools for precision agriculture in the 21st century.

The work involves combining analytical geospatial approaches, machine learning methods, and high-performance computing techniques to build cyberinfrastructure tools that can transform how ecoinformatics data — that is information on landscapes, soils, climate, organisms, and ecosystems — is analyzed.

“Available environmental data is exponentially increasing by including products derived from remote sensing, models, and ground observations,” Vargas says. “We have entered an era of environmental big data sets.”

The developed tools will be made accessible for field practitioners through lightweight virtualization, mobile devices, and web applications, and the educational components will help train the public and students in using the tools supported by online tutorials — for example, through YouTube videos.

Vargas explains that quantitative accessible information at relevant spatial scale is needed to better understand temporal variability, parameterize models, and accurately represent spatial soil moisture to improve agricultural practices.

Feedback on the tools’ interoperability, usability, manageability, and sustainability will be “crowd-sourced” through input provided by users and collaborators at the United States Department of Agriculture and the International Soil Reference and Information Center in the Netherlands.

The researchers expect the project to help answer a number of important questions, including how ecoinformatics data can be used to develop predictive capabilities for precision agriculture; what algorithms are required to analyze and synthesize ecoinformatics datasets; and what types of training and tools are needed for students, scientists, and field practitioners to use the data in a meaningful way.

“Our project aims to combine knowledge, techniques, and expertise from plant and soil sciences and computer science to build tools for advancing agriculture production,” Vargas says.

The research supports the “Growing Convergence Research at the National Science Foundation,” one of 10 Big Ideas for Future NSF Investments. The agency seeks to highlight the value of convergence as a process for catalyzing new research directions and advancing scientific discovery and innovation.

Funding for the project was awarded by the Office of Advanced Cyberinfrastructure and jointly supported by the Division of Earth Sciences within the NSF Directorate for Geosciences.

Vargas and Taufer also received a University of Delaware Research Foundation seed grant that is complementing the integration across these two disciplines driven by a compelling problem such as precise agriculture.