Predicting root zone moisture

Soil moisture plays a vital role in agriculture, drought and flood forecasting and water supply management. Adequate moisture in the soil allows for plants to grow, while too much moisture can create conditions ripe for flooding. Understanding the levels of moisture in the soil is important for many different groups.
Soil moisture can be broken up into two different areas: surface soil moisture and root zone moisture.
Root zone moisture, considered to be the upper three to six feet of soil, contains the water available for plants. Understanding root zone moisture levels helps farmers ensure their plants have enough water for growth and allows them to manage water supplies in order to ensure yields remain optimal.

In a new project through the Jerome J. Lohr College of Engineering's Future Innovators of America program, Matthew Croke, an incoming South à£à£Ö±²¥Ðã State University senior from Faribault, Minnesota, will work with assistant professor Aritra Banerjee to develop a system that predicts root zone moisture using artificial intelligence and remote sensing tools.
"This is part of a larger project that aims to determine the variation of soil moisture using remote sensing and deep learning," Banerjee said. "Matthew’s work includes the use of satellite data and weather data for different sites in the U.S. to predict the soil moisture and train and validate them with the ground sensor data."
Satellite images will be taken from different sites in the U.S., and drones from SDSU's Department of Geography and Geospatial Sciences will retrieve surface roughness values for some local sites. The drones will also retrieve dielectric material properties from the soil. These properties are widely used to estimate water content in soils through remote sensing and soil-level sensors.
Using electromagnetic techniques, Banerjee and Croke can determine the soil water content from the measurement of soil dielectric permittivity. The dielectric constant, which is a measure of a soil’s ability to store electrical energy in an electric field, is obtained from analyzing satellite data that is needed to obtain soil moisture, Banerjee said.
"A model will be developed by using a soil water retention curve, hydraulic properties of unsaturated soils and machine learning to accurately predict the root zone moisture from the surface moisture for known climatic conditions," Croke said.
Other students in Banerjee's lab — the BioGem Lab — are working to determine the root zone moisture using advanced machine learning techniques coupled with atmospheric conditions and soil water retention properties. Croke's project will be integrated into this larger effort.
Selections for the third class of Future Innovators of America Fellowships were announced last December by the Lohr College of Engineering. Recipients of the fellowship are awarded $5,000, with $4,5000 serving as a stipend and an additional $500 to cover the cost of lab supplies or travel to disseminate the results of the project. The 12 recipients and the departments which selected them are:
• Kamal Albousafi and Samara Overvaag, mathematics and statistics
• Caden Fischer and John Akujobi, computer science
• Matthew Croke, civil engineering
• Andrew Sternhagen, electrical engineering
• Random Nisia and Turner Marr, mechanical engineering
• Lydia Loken and Amee Parmar, ag engineering
• Jakob Burckhard and Levi Minion, construction and concrete industry management
The fellowships were created to provide unique research opportunities for undergraduate students in the college. Any student is eligible to apply as long as they are attending full time and have a GPA of 3.0 or higher. Each student worked with a potential project mentor, who must be a faculty or research staff member, to develop and submit a research plan that entails learning by doing.
Republishing
You may republish SDSU News Center articles for free, online or in print. Questions? Contact us at sdsu.news@sdstate.edu or 605-688-6161.