Scientists at the University of Nebraska–Lincoln have harnessed the power of machine learning and big data analytics to improve the Global Yield Gap Atlas, providing benefits to agricultural producers around the world.

“This is very exciting. We can now estimate the yield potential for every single piece of farmland around the world,” said Patricio Grassini, a professor in the Department of Agricultural Sciences and Horticulture at the University of Nebraska–Lincoln and lead researcher on the atlas.

This progress offers producers an opportunity to compare their current productivity and promote sustainable intensification of agricultural systems on a global level, Grassini said.

Fernando Aramburu-Merlos, a research assistant professor of agricultural sciences and horticulture, led the effort to develop a “metamodel” approach that leverages the ability of machine learning to assemble and analyze a complex set of information about soil, climate and cropping systems.

The resulting maps, which focus on corn, wheat and rice, provide extremely detailed and accurate information about how much of a crop can be produced at a particular location under the best conditions.

“With this metamodel, we are taking full advantage of the large amount of data the Atlas has collected over the past 12 years to create a global product that is ready to use and easy to use for scientists, farmers and companies,” said Aramburu-Merlos, an expert in crop simulation modeling.

Aramburu-Merlos and co-authors explain the refined analytical approach in a new article recently published in the journal Nature Food.

“Over the past decade, significant improvements in computing power, spatial information on soil and climate, and advances in the use of machine learning for geospatial analysis have produced new tools that can help overcome the limitations of bottom-up and top-down approaches,” the authors write.

Aramburu-Merlos, who joined Grassini’s lab as a lecturer in 2022, has strengthened the department’s expertise in computational and crop simulation. Aramburu-Merlos and Grassini co-authored the new paper with Martin van Ittersum and Marloes P. van Loon of Wageningen University and Research in the Netherlands, who co-lead the Global Yield Gap Atlas initiative.

The metamodel approach extends the Atlas’ existing bottom-up methodology by incorporating a broader range of datasets and applying an analytical approach that improves accuracy, allowing agricultural producers to access detailed, site-specific data on yield potential.

Using this information, farmers can diagnose their current performance and develop strategies to close the gap between actual and potential production, taking into account local climate and soil conditions.

“The applications extend far beyond the farm level,” says Grassini. “The refined ability to estimate yield potential offers governments, international organizations and charitable foundations a transparent and objective approach to understanding where the greatest opportunities for increasing yields exist.”

“(The metamodel data) provide important information for their investments in agricultural research and development programs.”