Research HighlightsNCSRP Project Explores Prescriptions for Best Management Practices
By Carol Brown
Imagine if a farmer could enter information into a computer program about a particular field and it would respond with recommendations for the best productivity and profitability. This is the ultimate goal of a research group involved in a long-term project funded by the North Central Soybean Research Program.
The project is comprised of several years’ worth of data collection, number crunching and communication between farmers and researchers, going back to 2015.
“This is the third iteration, or grandchild, of what began as a benchmarking project,” says Shawn Conley, University of Wisconsin soybean specialist and lead principal investigator on the current project. “We gathered a lot of data over the first two projects to get where we are today.”
The current project includes analysis and evaluation of the amassed data to ultimately create an online decision tool that will provide a farmer with best management recommendations. The recommendations will be based on factors affecting crop yield including input and environmental information. But it had to start with data collection and identifying where gaps lied between current and potential yields.
“In the first project, we conducted agronomic surveys across the North Central region, collecting details on soybean inputs and production from more than 8,000 farm fields and 600,000 acres,” explains Conley. “We created a huge database over a three-year time frame.”
The second project involved pooling all that data into distinct geographical areas with similar climate and soil conditions to analyze the data more closely. The team identified these areas as Technology Extrapolation Domains, or TEDs. The team then tested the best management recommendation outcomes predicted by the data analysis. They conducted more than 100 replicated, large-plot, on-farm field trials, comparing traditional farming practices with the best management practice recommendations according to their statistics.
“We told farmers to continue with what they usually did in their fields, and in nearby strips, they tried our recommendations over three crop seasons,” Conley explains. “Farmers gained 3.2 to 5.5 bushels per acre on average and between $30 to $50 per acre in profit by integrating the best management practice recommendations based on the benchmarking project.”
All of this has led to the current NCSRP project. The research team is developing an online decision tool to generate best management practices for any field.
“Ultimately, our goal is to develop a program where a farmer can drop a pin in a field location, run our algorithm, and develop a prescription for that given field,” he comments. “So far, we have developed the artificial intelligence algorithms and we’ve integrated the platform. We are working on the beta version of the program that is developing site-specific recommendations.”
The team is aware of farmer hesitations when supplying their farm’s information, and Conley says they understand that farmers are very security-conscious with their data. They are making sure everyone is clear about who has access and, more specifically, who doesn’t have access to their private data.
Setting This Decision Tool Apart
The research team knows other decision tools exist similar to the one they are developing. But they intend to make sure this one is different. Other tools create recommendations based on simulated or synthetic data, which may not align with a farmer’s experience in the field. The team is using actual field-level information such as soil type, tillage method, crop variety and rotation as well as inputs such as fertilizer and herbicides and their associated costs. They also include weather and pest presence information.
“I’m excited about this project. I know our data is good. It’s robust and scientifically proven and our techniques are proven,” he comments. “I’m confident that when we roll this out, it will be a really cool, useful tool to help farmers optimize their management practices for improved productivity and profitability.”
Part of the current project includes ground truthing the decision tool. The team continues to survey farmers and scout their fields to align issues, yield monitor data and satellite imagery to hone their information and ensure the tool is as accurate as possible.
“We’re working on enriching the system by entering data from this past crop year as well as this coming crop season,” says Conley. “Hopefully, funding will continue so we can launch this to farmers in 2024.”
Published: Jul 3, 2023
The materials on SRIN were funded with checkoff dollars from United Soybean Board and the North Central Soybean Research Program. To find checkoff funded research related to this research highlight or to see other checkoff research projects, please visit the National Soybean Checkoff Research Database.