SDSU researchers to develop soil moisture mapping capabilities 

A collaborative project from South Dakota State University, South Dakota Mines and Oglala Lakota College will develop novel algorithms needed for soil moisture mapping — an essential tool for helping farmers with their agricultural management decisions. 

Accurate soil moisture measurements are essential for helping farmers with their agricultural management decisions. Soils that are too wet can lead to nutrient leaching, while soils that are too dry can mean decreased crop yield and quantity. Understanding the soil moisture levels can help farmers make informed decisions that allow for improved crop yields and efficient water-usage.

Despite its importance, widespread mapping of soil moisture levels has been difficult. Since the 1970s, scientists have been experimenting with remote sensing — the measurement of properties on Earth's surface via data from satellites — to develop continuous measurements of soil moisture levels. But developing the needed algorithm for widespread accurate measurements has had its challenges, specifically due to data disruptions from ground-level vegetation, surface roughness and from multiple sensor data acquired at different dates and varying degrees of canopy penetration.

Now, researchers from South Dakota State University, South Dakota Mines and Oglala Lakota College are ready to meet these challenges head on. Over the course of a three-year, $750,000 project funded through the National Aeronautics and Space Administration (NASA), the research team will develop algorithms for mapping soil moisture levels using satellite data in support of agriculture and water management practices.

Hankui Zhang, assistant professor in SDSU's Department of Geography and Geospatial Sciences, will serve as the project's primary science investigator. Colleague Maitiniyazi Maimaitijiang, assistant professor of remote sensing, is a co-investigator. 

"This presents an excellent opportunity for South Dakota researchers, who have strong interactions with crop domain scientists, to develop a soil moisture mapping algorithm for agricultural applications," Zhang said. "I am looking forward to the implementation of our research plans and their success."

Edward Duke, professor of geology at South Dakota Mines, will serve as the project's primary investigator. Randy Hoover, a professor of electrical engineering and computer science at South Dakota Mines, is a co-investigator.

“This proposal is particularly timely because NASA and the Indian Space Research Organization are developing a new satellite-based soil moisture sensor for launch in 2025," Duke said.

Outside of supporting agricultural management decisions, mapping soil moisture levels will provide insights into the planet's water cycle, improve weather forecasting and contribute to the understanding of the changing climate and ecosystems.

"One of the challenges associated with soil moisture mapping from geospatial data is fusing data from different monitoring platforms," Hoover said. "These platforms have very different temporal, spatial and radiometric scales that all need to be integrated to provide a cohesive input-stream for the deep learning algorithms being developed."

To develop the algorithms, Zhang and Maimaitijiang will leverage existing and soon-to-be launched satellite data (Landsat 8/9, Sentinel-1, Sentinel-2 and NISAR) to derive high-resolution soil moisture maps. The research team will work in conjunction with the U.S. Geological Survey Earth Resources Observation and Science Center near Sioux Falls. The algorithm will fuse together the data sources to create a publicly accessible product that can be used by producers, researchers, meteorologists and many other groups.

"Our methods can retrieve soil moisture at any satellite data acquisition date, which advances beyond previous efforts that retrieve soil moisture only when the microwave data are contemporaneous with optical data," Zhang said. "This innovation is achieved through a novel time series deep learning methodology developed by the South Dakota team to model the vegetation seasonal dynamics to better quantify vegetation coverage and soil moisture." 

The maps will be validated by ground-level field measurements from Dana Gehring and Charles Tinant, researchers at Oglala Lakota College.

The project will also involve local industry partners, who will use the derived soil moisture data support their irrigation systems.

"With our collaborators, we aim to not only develop a novel soil moisture mapping algorithm for NASA but also advance STEM workforce development in South Dakota," Zhang added.

Overall, the project is expected to positively impact the state's $32.1 billion agricultural industry, responsible for approximately 30% of South Dakota's GDP.