By Sahila Beegum, NWC Research Assistant Professor
Sahila Beegum and Alakananda Mitra, Research Assistant Professors at NWC, and Aditya Kapoor, Postdoctoral Research Associate at NWC, closely collaborate with USDA ARS Adaptive Cropping Systems Laboratory (USDA ARS ACSL). Since 2021, this collaboration has consistently improved the process-based crop models developed jointly by USDA ARS ACSL and NWC. Recent improvements and updates include incorporating greenhouse gas (GHG) emission modeling, yield quality modeling, AI-based crop models, biogeochemistry integration, graphical interface development, and crop modeling training.
The crop models have been upgraded to simulate gas production and transport mechanisms and are now integrated with GHG emissions routines for carbon dioxide (CO₂) and nitrous oxide (N₂O) emissions. A recent publication in the European Journal of Agronomy evaluated the CO₂ model in a soybean-maize ecosystem under present and future climatic conditions. The study found that climate change will lead to a long-term decline in soil organic carbon, but increased CO₂ could offset this loss. The N₂O simulation in the model has also been improved and is currently being tested with measured data. The group is now working to incorporate a methane (CH₄) emission routine.
Another focus is the integration of AI into crop models. Using extensive field data and process-based model-generated data, the group developed a Random Forest (RF) regressor-based model. This study, published in IEEE Access, demonstrates the applicability of machine learning models to climate-smart agriculture. The algorithm, initially developed for cotton models, is being expanded to develop an AIbased model for corn simulation.
One limitation of the USDA ARS ACSL crop models is the biogeochemistry component. The PhreeqC model is the most widely used standalone model for biogeochemistry, and the group is working on integrating it into the USDA-ACSL crop models. They will collaborate with the developers of PhreeqC on this effort.
The group also developed the world’s first cotton fiber quality simulation model, which quantifies cotton quality in addition to yield. This study published in Field Crops Research was featured in ScienceDaily, ScienMag, EurekAlert, MSU Newsroom, and Cotton Grower. Using this model, they created a spatial map for the best planting dates for cotton at the county level in the USA cotton belt, helping farmers and decision-makers select the optimal planting dates for high-quality cotton. The same methodology can be applied to other crop models.
Significant advancements have been made in the interface called CLASSIM, the graphical user interface (GUI) for the crop models. A recent addition is an expert system that provides irrigation recommendations based on soil moisture status.
USDA ARS ACSL, in association with NWC, was invited to provide a one-day, hands-on crop modeling workshop at the “Training Program on Building Climate Resilience Through Crop and Hydrological Modeling” conference. This event was coordinated by the National Agriculture and Forestry Research Institute (NAFRI), the Department of Meteorology and Hydrology (DMH) of Lao PDR, and the Economic Research Institute for ASEAN and East Asia (ERIA). The online and in-person event took place on September 1, 2024, with 47 participants. The workshop focused on providing knowledge and practical skills in using USDA-ARS ACSL-developed crop models and the CLASSIM interface. It highlighted process-based modeling approaches to support climate-resilient planning and decision-making. The crop models the research group are currently improving can be accessed at go.unl.edu/ cropmodels.
