New tool to help farmers make nitrogen input decisions

“With these numbers in mind, farmers can make more informed decisions on nitrogen mitigation that not only saves them money, but also significantly reduces pollutants to the environment,” said Zhenong Jin, who led the research and is an assistant professor in the University of Minnesota’s Department of Bioproducts and Biosystems Engineering (BBE) in the College of Food, Agricultural and Natural Resource Sciences (CFANS).

According to the researchers, previous tools did not allow for customised predictions for every field in the U.S. corn belt, as the computational and storage costs of running these crop models at large scale would be very expensive.

Mimic crop model at much faster speeds

The research team built a series of machine-learning-based metamodels that can almost perfectly mimic a well-tested crop model at much faster speeds. Using the metamodels, they generated millions of scenario simulations and investigated two fundamental sustainability questions: where are the mitigation hotspots, and how much mitigation can be expected under different management scenarios.

The study, conducted in the U.S. Midwest corn belt, found that:

• Reducing nitrogen fertiliser by 10% leads to 9.8% less N2O emissions and 9.6% less nitrogen leaching, at the cost of 4.9% more soil organic carbon depletion, but only a 0.6% yield reduction over the study region.
• The estimated net total annual social benefits are worth $ 395 million (uncertainty ranges from $ 114 million to nearly $ 1.3 billion), including a savings of $ 334 million by avoiding GHG emissions and water pollution, $ 100 million using less fertiliser, and a negative $ 40 million due to yield losses.
• More than 50% of the net social benefits come from 20% of the study areas, thus can be viewed as hot spots where actions should be prioritised.

Bundle cover cropping with nitrogen management

“Our analysis revealed hot spots where excessive nitrogen fertiliser can be cut without yield penalty,” said Jin. “We noticed in some places that reducing nitrogen-related pollution comes at a cost of depleting organic carbon in soil, suggesting that other regenerative practices, such as cover cropping, need to be bundled with nitrogen management.”

In the future, the team aims to develop more advanced and accurate carbon qualification models through a combination of process-based models, artificial intelligence and remote sensing.