Billions of dollars have been spent on programs that encourage American farmers to adopt conservation practices aimed at protecting soil, improving water quality, and building agricultural resilience. But measuring whether those investments are actually changing what happens in individual fields has remained a challenge. Researchers at Virginia Tech, working with the University of Maryland, are now testing whether a new class of artificial intelligence models can fill that information gap.

The project is led by Elinor Benami, an assistant professor in Virginia Tech's College of Agriculture and Life Sciences Department of Agricultural and Applied Economics. Benami was selected as a 2026–2027 Virginia Tech Institute for Society, Culture, and Environment Scholar, which provides seed funding to support the research, according to a university news release.

“Our work connects advances in geospatial technology with the economic questions that matter for conservation decisions,” Benami said. “By applying these tools to economic and policy-relevant questions, we aim to develop more useful information for farmers, policymakers, and others making decisions about conservation investments.”

The team will focus on two specific practices: cover crops and tillage. Currently, the most comprehensive national data on agricultural conservation comes from the U.S. Agricultural Census, which is published every five years at the county level. Virginia Tech noted that these statistics rely heavily on surveys with declining response rates and generally do not show exactly where within a county practices are adopted, making it difficult to track changes at the field level or from year to year.

To address that limitation, the researchers will test a Geospatial Foundation Model (GFM) called OlmoEarth, developed by the Allen Institute for Artificial Intelligence with input from domain experts, including scientists from NASA Harvest and the International Food Policy Research Institute. According to the university, OlmoEarth can learn from changes over time and combine multiple types of satellite information, including optical imagery and radar, allowing it to capture seasonal patterns even when cloud cover limits visibility. Its open design and lightweight options for local analysis also make it easier to adapt for public-interest applications, including those with data confidentiality concerns.

The initial phase of the study will examine field-level tillage and cover crop records from the Chesapeake Bay region, covering parts of Virginia and Maryland. The researchers will evaluate how accurately the models identify conservation practices across different landscapes and farming systems, how well the approach transfers to other locations and years, and whether it can reduce the need for costly additional field data while maintaining performance. Observations from other states may also help assess broader applicability, the university said.

Reliable maps of conservation practices could help researchers better understand how adoption affects farm productivity, land values, and resilience to events like floods and droughts. The information could also support more timely monitoring, targeting of incentives, and evaluation of major conservation investments.

“Farmers, researchers, and policymakers all want to better understand which conservation practices work, where they work, and what benefits they deliver,” said Mark Reiter, director of the Eastern Shore Agricultural Research and Extension Center and a professor at the School of Plant and Environmental Sciences. “Finding ways to more quickly and cost-effectively identify where those practices are being adopted, with or without subsidy programs, is an important step toward more robust impact assessment — and ultimately toward directing resources where they can generate the greatest returns for producers and the public.”

Michael Cecil, a postdoctoral associate with NASA Harvest at the University of Maryland, and Inbal Becker-Reshef, a research professor at the same university who founded and currently directs NASA Harvest while also serving as managing director of Microsoft AI for Good, are additional members of the research team. Reiter additionally holds the position of research director for the Alliance to Advance Climate-Smart Agriculture.

The pilot project is expected to be completed by mid-2027.