CSU project uses AI to turn soil data into actionable insights for farmers 

An interdisciplinary research team at Colorado State University is using artificial intelligence to help farmers better understand soil health by turning varied agricultural data into practical, decision-ready insights that are easy to access.

Healthy soil is key to farming success, supporting water retention, root growth and nutrient cycling. However, tracking conditions and understanding how management choices affect them over time remains difficult — especially amid challenges such as drought, increasing weather variability and maximizing crop yields.

The newly funded TerraScope project at CSU addresses that gap. In it, the team plans to use AI to combine on-the-ground measurements and remote sensing data; paired with advanced simulations, this will reveal patterns and relationships farmers can act on that would otherwise be invisible when viewed in isolation.

The $1 million project is funded by the U.S. Department of Agriculture. The CSU team includes computer scientists from the College of Natural Sciences as well as soil and outreach experts from the College of Agricultural Sciences. The researchers hope to connect key data points on a unified platform that is useful, responsive and built on feedback from agricultural producers on their needs for soil management.

A key aspect of the project is organizing data that comes from a variety of sources. Professor Shrideep Pallickara said a challenge will be aligning formats, resolutions and even terminology from different settings as they relate to aspects such as soil moisture.

“The data we will be using to train these AI-informed models is both diverse and voluminous,” he said. “That makes it a challenge to work with, but by combining information from traditional observations or on-the-ground records provided directly by farmers with data collected remotely by satellites, we hope to address gaps in understanding.”

FEEDBACK FROM PRODUCERS

Megan Machmuller is a research scientist at CSU and co-director of university’s  Integrated Rocky Mountain-region Innovation Center for Healthy Soils. She said the project includes a strong outreach element and that the team is dedicated to including feedback from agricultural producers in their work.

“Too often, tools like this are created without considering usability or the reality of what producers need. TerraScope will be crafted in collaboration with the agricultural community, and that guarantees it will be practical, reliable and attuned to the real-world challenges faced by producers in Colorado and beyond,” she said.

Machmuller added that this work embodies a fundamental tenet of the university’s land-grant mission to advance science that directly benefits and serves people.

“Colorado’s diverse agricultural landscape, climate, soils and topography provide an ideal environment for the launch and refinement of this platform,” she said. “The rich legacy of soil research here also provides a robust foundation for advancement, including comprehensive datasets and ecosystem models that will be instrumental in this project. This work will strengthen decision-making capacity across the state and enhance the long-term resilience of agricultural systems.”

The CSU team also includes professors Eugene Kelly and Keith Paustian from the Department of Soil and Crop Sciences, Sangmi Pallickara from Computer Science, and Jay Breidt from Statistics. James Hale, the research director of community food systems and social sustainability in the Department of Sociology, is also part of the project.