A team of scientists from the University of Minnesota has combined satellite imagery, artificial intelligence, and crowd-sourced observations to track the spread of leafy spurge, a noxious weed toxic to livestock, over a 20-year period. Their findings, published in the journal Ecological Applications, reveal that the plant's abundance increased by 570% between 2000 and 2021, with the most dramatic growth occurring around the Fargo-Moorhead and Minneapolis-Saint Paul areas.
Leafy spurge was accidentally introduced to North America from Europe in the late 19th century. It poses a particular threat to farmers and ranchers because when damaged, it secretes a sap that is toxic to livestock. According to the University of Minnesota, the plant caused an estimated $1 billion in economic losses from 1960 to 2000.
The research team, based in the Department of Plant and Microbial Biology within the College of Biological Sciences, trained a deep-learning computer model on real-world observations of leafy spurge. These observations included data from the iNaturalist app, a platform where community members log plant and animal sightings. The model then combined those records with open-source Landsat imagery provided by the U.S. Geological Survey with assistance from NASA satellites.
“When leafy spurge flowers, it's like someone took a yellow highlighter to the landscape. It's so bright yellow, satellites can see it too,” said Thomas Lake, a University of Minnesota PhD graduate who is now a postdoctoral researcher at the Center for Geospatial Analytics at North Carolina State University.
The study found that while many plant ranges shift primarily due to climate, the massive changes in leafy spurge abundance over the two-decade period occurred mainly near major roadways. This suggests that proactive roadway management strategies could help slow the plant's invasion. The researchers also noted that during periods of intense drought, leafy spurge was less likely to be detected because it produced fewer flowers.
Crucially, the detection of newly invaded areas from satellite imagery improved the model's ability to forecast future spread. This capability, the researchers say, can inform state and federal agencies as they work to manage the invasive species.
“Having a fuller picture of how plants invaded and where they actually occur will help us to make better predictions about where it might eventually end-up,” said Ryan Runquist, a senior research associate in the College of Biological Sciences.
“Tracking invasive species from satellites vastly improves our capacity to dynamically monitor invasive spread and prioritize limited resources for management,” said David Moeller, a professor in the College of Biological Sciences. “Remote sensing significantly reduces bias by allowing detection in areas where observers often miss.”
With the leafy spurge map for Minnesota now complete, the research team plans to expand the project to cover a seven-state area. They also intend to study how climate change, insects, and rapid plant evolution affect the weed's spread.
The project was funded by the University of Minnesota's Minnesota Invasive Terrestrial Plants and Pests Center, supported by the Minnesota Environment and Natural Resources Trust Fund, as well as the USDA National Institute of Food and Agriculture.
For university students studying environmental science, ecology, or data science, the research offers a concrete example of how remote sensing and machine learning are being applied to real-world conservation challenges. The method developed by the team could eventually be adapted to track other invasive species, opening up new avenues for research and land management.