Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Adaptive Cropping Systems Laboratory » Research » Publications at this Location » Publication #429112
Research Project: Developing Practices for Nutrient and Byproducts to Mitigate Climate Change, Improve Nutrient Utilization, and Reduce Effects on Environment (BRIDGE PROJECT)

Location: Adaptive Cropping Systems Laboratory

Title: Machine learning-based prediction of cereal rye cover crop biomass across diverse agroecosystems

Author
item GHIMIRE, UTSAB - University Of Florida
item MITRA, ALAKANANDA - University Of Nebraska
item Fleisher, David
item PARK, JOHN - Oak Ridge Institute For Science And Education (ORISE)
item Barnaby, Jinyoung
item KIM, YONGHYUN - George Washington University
item Han, Eun Jin

Submitted to: Agricultural & Environmental Letters
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/12/2025
Publication Date: 1/9/2026
Citation: Ghimire, U., Mitra, A., Fleisher, D.H., Park, J., Barnaby, J.Y., Kim, Y., Han, E. 2026. Machine learning-based prediction of cereal rye cover crop biomass across diverse agroecosystems. Agricultural & Environmental Letters. https://doi.org/10.1002/ael2.70055.
DOI: https://doi.org/10.1002/ael2.70055

Interpretive Summary: Cover crops like cereal rye provide many benefits to farms, such as improving soil health, reducing erosion, and supporting the next cash crop. To manage these benefits, farmers and advisors need reliable ways to estimate how much cover crop biomass (plant growth) will be produced. In this study, we developed machine learning models that use weather, soil, and basic management information to predict cereal rye growth. We trained these models using data collected from 24 states across the eastern U.S. The models were fairly accurate when using weather data from early spring, and adding later-season weather data sometimes improved the results but also made the models less reliable. We also tested a version of the model that provides a range of possible outcomes instead of a single number, which helps account for uncertainty. Our results show that with just public weather and soil data plus a little management information, it is possible to make useful and practical predictions about cover crop growth.

Technical Abstract: Accurate operational predictions of cereal rye biomass are essential for quantifying the agroecosystem services provided by cover crops and for informing growers’ management decisions for subsequent cash crops. In this study, we developed machine learning (ML)-based biomass prediction models using two advanced gradient-boosted tree algorithms (CatBoost and XGBoost). A comprehensive dataset of cereal rye biomass and management information from 24 states in the eastern United States, combined with publicly available soil and weather data, was used to train the models. Models relying on early spring weather inputs achieved moderate predictive skill (R² ˜ 0.74). Incorporating later-season weather data modestly improved mid-season fits but led to overfitting in late-spring models. Extending CatBoost to quantile regression enabled estimation of 10–90% prediction intervals with moderate pinball loss. Overall, our findings demonstrate that public soil and weather data, supplemented with limited management inputs, can support interpretable, uncertainty-aware biomass predictions suitable for cover crop management.