Research Highlights

Developing a Soybean Planting Playbook

In this article you’ll find details on:

• In North Carolina, soybeans are planted from March to July, and farmers use maturity groups 2 through 8, making it challenging to choose the best planting date, maturity group and seeding rate for a given situation. 

• Combining five years of in-field trials with computer science expertise and artificial intelligence, North Carolina State is developing a tool to help farmers make the best play call as they plant soybeans.

---

By Laura Temple

Pitchers don’t throw the same pitch every time. Some situations call for a high fastball, while others call for a change-up low and away. Nor do quarterbacks call the same play every down. Sometimes a run up the middle is most likely to succeed. Other times, the team needs a long pass to a wide receiver.

At the top of their game, these athletes choose the pitch or play that fits the situation. And they rely on teammates, coaches and scouting analytics for information and guidance to make the best call.

The same concept applies to farmers as they select soybean production practices for their unique operations. A given soybean planting date, maturity group and seeding rate doesn’t fit every field and rotation. And to be at the top of their game, they need information and guidance to choose the best option. 

That’s especially true in North Carolina, according to Rachel Vann, soybean extension specialist and assistant professor with North Carolina State University. 

“North Carolina has diverse soil types and cropping systems in different regions of the state,” she says. “Our farmers plant soybeans anytime from March through late July, and they use varieties from maturity groups 2 through 8.”

With so many different situations and options, it can be tough to make the best call.

“Given the advances in soybean genetics and production practices, the need to update fundamental agronomic research to optimize our soybean production recommendations in different situations became clear,” Vann says. “The ultimate goal has been to help farmers figure out the best maturity group and planting population to use for a given planting timing, whatever that might be in their rotation.”

Scouting the Fields

Vann’s first step was to gather data. The North Carolina Soybean Producers Association funded five years of trials at 20 locations throughout the state. The trials compared planting dates, maturity groups and seeding rates.

“To serve diverse planting scenarios, each location contained about 5 acres of small plots,” she says. “We gathered yield and quality data from planting through harvest for a wide range of planting dates and maturity groups, with differing seeding rates across these combinations.”

These complex trials, conducted from 2019 through 2023, generated a massive amount of data. 

“We wanted to make the data useable and dynamic for farmers,” Vann explains. “However, we soon realized that the volume and complexity of the data was more than our initial team could manage.”

She says that during the last couple years of the field trials, NC State was encouraging interdisciplinary projects. She acknowledged that additional expertise would help figure out how best to get the results to farmers in a form that would help them make better planting decisions.

The N.C. Plant Sciences Initiative strongly facilitated connectivity to other disciplines, ultimately leading to securing funding from the NC State Data Science Academy to bring computer science expertise to the team.

Designing the Plays

Vann’s team is now working with Cranos Williams, professor of agricultural analytics and platform director for the N.C. Plant Sciences Initiative; Somshubhra Roy, computer science graduate student; and Katherine Stowe, Ph.D., and director of the U.S. Soybean Research Collaborative. Together, they are developing a dynamic decision support tool that will help farmers make the best call for planting soybeans, given their situation, region and crop rotation. 

“The tool allows farmers to enter their location and get the ideal soybean planting date, maturity group and population,” Vann explains. “They can adjust the date to better fit their rotation and see how that changes maturity group recommendations and optimal seeding rate.”

The tool is being built on the foundation of the five years of extensive field data Vann’s team gathered. That data also is training, or “teaching,” a machine learning program to make sound recommendations for areas beyond the planting trial locations. The system has been in beta testing with soybean farmers and extension agents, and Vann expects it to be released in the fall of 2024, to begin informing decisions for the 2025 growing season.

“The farmers who have beta-tested this new tool believe it has the ability to transform how they make soybean planting decisions,” she reports. “They will be able to make core production decisions based on sound data, customized for their operation.”

She notes that the collaboration has added value to those involved, as well.

“It took some time to figure out how best to work together, because our groups speak different ‘languages’ — agronomy and computer science,” Vann says. “Working through those challenges has been worth it, as we have learned that our groups have to work together to ensure producers have the most robust solutions available.”

As pitchers, quarterbacks, coaches — and farmers — know, using data to design and call the best plays increases chances for success in any field. This research will use data-driven information in a new way to set farmers up for greater success in soybean production. 

• Meet the Principal Investigator on this project: Rachel Vann

Published: Jul 15, 2024