Using big data to produce bigger return for growers, less environmental impact

Two South Dakota State University researchers are partnering with 20 eastern South Dakota farming operations and GEVO to pursue practices that will reduce greenhouse gas emissions and increase farm profitability.

“Increasing the Adoption and Generation of Climate-Smart Practices to Produce Low Carbon-Intensity and Net Zero Sustainable Products” is designed to quantify the carbon intensity score for jet fuel produced from corn in South Dakota. 

Hossein Moradi, an assistant professor in the Department of Mathematics and Statistics, is co-PI with David Clay, the principal investigator. Clay, a distinguished professor and South Dakota Corn Endowed Chair in Precision Agriculture, has been conducting agricultural research for 30 years. 

Moradi was new to agricultural statistic work when he arrived at SDSU in August 2018. However, the director of SDSU's Statistical Consulting Center has been working  with Clay for several years. 

In these projects, Moradi is using statistical models and machine learning/artificial intelligence techniques to convert information from satellites and drones, combines, soil test results, weather reports, soil surveys and  farmer practices into recommendations that will reduce costs and improve profitability.  

Moradi boils it down: “Are there farming practices that improves soil health while minimizing the impact of agriculture on the soil, water and air?

“We’re also tracking climate data, such as mean, minimum and maximum temperature and moisture, to produce a model that will be useful to South Dakota farmers.”

Project runs through December 2027

The project started in January and finishes up in December 2027, so this is the first growing season to accumulate a full data set. However, the researchers do have onsite climate data for the past two years as well as satellite imagery, which, through the federal government’s Landsat satellite, passes over the sites every 16 days, Moradi said.

He said he is developing statistical models that include farming practices, seed types, and fertilizer data in addition to climate and satellite data.

Moradi said, “We need a better machine learning model to assess the data. I’ve had lots of back-and-forth conversations (with David Clay): ‘Maybe add this or that type of data. Is this variable useful in the model?’ As a statistician, I don’t know enough about agriculture to know what variables should be included in the model, but I do know how to develop models that are useful to answer scientific questions.”

Iowa State, Colorado State partner with SDSU

SDSU has two other academic partners in what overall is a $30-million project financed by the U.S. Department of Agriculture’s Partnership for Climate-Smart Commodities program. Iowa State and Colorado State have similar projects underway. SDSU’s share of the budget is $1.8 million.

Moradi said his challenge as a statistician is, “Data comes in lots of shapes and forms. I have to make sure all the data is at the same level. For example, Landsat data is based on 30-meter pixels. (Farm) combine data comes in 10x10 pixels. I have to develop an appropriate data integration model. My job is to find out what state-of-the-art model is the best to analyze the data or develop my own models to get the job done.”

Naturally, most of the data collecting is done during the growing season with modeling occurring during the off-season.

By the end of the project, Moradi and Clay hope to be able to submit their findings to multiple research publications. The researchers also are working with Gevo, an Englewood, Colorado-based firm with plans to develop sustainable transportation fuels. In early 2021, Gevo bought 240 acres near Lake Preston to develop sustainable aviation fuel using corn.

‘The data always has a story’

Moradi said, “Gevo has a lot of farms in its network, and we hope to help create a structure and roadmap leading farmers to improved soil health.”

Clay said the farming partners are all corn and soybean operations that are reducing their tillage intensity, planting cover crops, using  improved nitrogen management techniques and applying biological products that are designed to improve soil health.

Moradi said, “Farmers are doing most of the work and we are working hard to insure that they will receive most of the project benefits. 

“As a statistician, I believe that data always has a story and I have to find the best way to tell that story.”