Auburn University faculty member Sanjiv Kumar in the School of Forestry and Wildlife Sciences is leading a research team in connecting forecasts of soil moisture—a key influencer of climate change and a vital component in agriculture productivity and the prediction of flood and drought—through the combined use of big data, artificial intelligence and user interactions.
The study, “FACT: Interactive Deep Learning Platform and Multi-source Data Integration for Improved Soil Moisture Forecasting,” which began in September and will continue through August 2023, is funded by a nearly $500,000 grant from the USDA National Institute of Food and Agriculture, or NIFA. Kumar’s co-principal investigators are Woniun Lee of Yeshiva University in New York City, who oversees the big data portion of the research, and Imtiaz Rangwala of the University of Colorado, Boulder.
This research is one of 12 projects funded by NIFA grants totaling over $7 million that will initiate research on big data analytics, machine learning, artificial intelligence and other cyberinformatics—technology that is now considered necessary to keep the nation’s agriculture on the leading edge of food and agricultural production.
It is part of NIFA’s Food and Agriculture Cyberinformatics and Tools, or FACT, division, which aims to develop and provide data-driven solutions for complex agriculture problems.
Where the soil and the crowd meet the cloud
Kumar said an underlying concept of his team’s research is, “Let the climate model do its best, and let big data do the rest.”
The multidisciplinary approach will combine the expertise of earth system modeling, big-data technology and drought monitoring and forecasting. The objectives are to develop new algorithms for integrating soil moisture data from different sources, building a scalable big-data infrastructure and deep learning analytics platform for real-time interactive soil moisture forecast applications and developing new or improved forecast attributes at the interface of human technology and data interactions.
There are multiple real-world implications, Kumar said.
“The project aims to provide an improved soil moisture forecast that is calibrated using user-provided soil moisture measurements and location information,” Kumar said. “At the front end, the user can access the forecast using a mobile app that is supported by cloud-based computer power and software-defined storage technologies.”
It will provide an increasingly useful tool for climate study by improving overall understanding of the soil-moisture process: Using the crowd-sourced moisture data, the proposed system becomes smarter as the user provides new input, he added.
This artificial intelligence/human/big data research will play an important role in climate study, Kumar said.
School of Forestry and Wildlife Sciences Dean Janaki Alavalapati said Kumar and his team likely will reveal groundbreaking findings.
“The researchers’ implementation of multiple disciplines—from farmers’ observations to artificial intelligence to big data to study the measurement and forecasts of soil moisture—could lead to breakthroughs in this vital area of study within the next few years,” Alavalapati said. “This ambitious project is one more example of Auburn researchers stepping to the forefront of innovations with worldwide implications.”
While the work that Kumar and his team have ahead of them benefits from its multidisciplinary approach to soil moisture science, it also delivers the most high-tech—but easily accessible—data to the people who need it most, when they need it most.
“This project combines fundamentals of climate science with big-data technologies and stakeholder interaction to bring up-to-the-minute science to farmers’ fingertips,” he said.
(Written by Teri Greene)