Farmland often stands at the forefront in the struggle against climate change. Competing interests, such as the need for renewable energy sources like solar panels and bioenergy crops, often put pressure on available agricultural land.
Farmers face choices between producing food or participating in initiatives that could inadvertently lead to the release of more greenhouse gases by expanding tillable areas.
In addressing these issues, researchers at the University of Wisconsin–Madison have made significant strides. Led by Yanhua Xie and Tyler Lark, a team developed a machine learning tool to create a detailed map of nearly 30 million acres of abandoned farmland across the United States since the 1980s.
This map serves as a critical resource for identifying lands that can be repurposed for sustainable energy and food production without adding pressure to existing farm operations.
The findings, published in the Environmental Research Letters journal, provide insights into these previously cultivated lands. These areas could be repurposed to grow crops like switchgrass or sorghum, which are not only effective in capturing carbon in the soil but also serve as potential biofuels, reducing reliance on petrochemical energies.
Lark emphasizes the importance of pinpointing these lands to fully assess their potential for climate change mitigation. By understanding their specific characteristics, it becomes possible to strategically plan clean energy investments.
Different applications, ranging from solar photovoltaic installations to the restoration of natural ecosystems, could potentially find valuable land among these abandoned fields.
The collaboration between scientists from UW–Madison and Michigan State University has yielded publicly accessible data, hosted in the GLBRC’s interactive atlas of U.S. cropland. This resource not only highlights abandoned lands but also tracks trends in farmland expansion and irrigation, offering a comprehensive view of land use changes over decades.
Traditionally, researchers leaned on datasets like the USDA’s Census of Agriculture, which provides estimates based on county-level data every five years. However, these datasets lacked the granularity needed to accurately locate and date farmland abandonment at the field level.
Recent advancements in cloud computing have allowed for a significant leap forward. By training a computer with existing land cover data, the team could apply algorithms to satellite imagery from 1986 to 2018, effectively mapping out the land use changes.
The algorithm achieved a remarkable accuracy, correctly identifying abandoned croplands nine out of ten times and even pinpointing the specific year of abandonment with about 65% accuracy.
This level of detail revealed that more than 30 million acres of cropland were abandoned over the span of 32 years, with high concentrations in regions such as the Great Plains and along the Mississippi River.
Table: Abandoned Farmland Land Use Changes
| Land Use Conversion | Percentage of Total |
|---|---|
| Pasture or Grassland | >50% |
| Shrubland, Forest, Wetland, or Bare | ~33% |
| Enrolled in Conservation Programs | <20% |
Interestingly, a substantial portion of the abandoned land was not part of formal conservation programs like the USDA’s Conservation Reserve Program. This finding suggests that more land than previously understood might be available for bioenergy crops, presenting new opportunities for sustainable agriculture and energy development.
The data derived from this research can help model potential biomass production and estimate the amount of carbon dioxide that could be sequestered by these lands. Yet, the reasons behind the abandonment of these lands remain unclear, representing the next step in the research.
This investigation may involve combining land use data with socioeconomic factors and tax records to understand the underlying causes and potential future uses for these parcels.
Different scenarios could emerge based on these analyses. For instance, areas currently used for hay might easily transition to bioenergy crops like switchgrass, given the existing farming infrastructure. Conversely, regions with no agricultural activity might be more suited for solar installations or other non-farming uses.
The detailed mapping and data generated also open up the potential for various stakeholders, from policymakers to farmers themselves, to make informed decisions about land use.
By understanding the specific attributes and historical context of these lands, optimized strategies for sustainable development can be crafted.
The study leverages advanced computing and satellite technologies, highlighting how modern tools can uncover valuable information from historical data. This high-resolution analysis represents a significant step forward in sustainable land management and climate change mitigation efforts.
