Most corn and soybean fields in the U.S. are planted with herbicide-resistant crop varieties. However, the evolution of superweeds that have developed resistance to common herbicides is jeopardizing current weed management strategies. Agricultural robotics for mechanical weeding is an emerging technology that could potentially provide a solution. A new study looks at the types of farmers and fields more likely to adopt weeding robots and at what stage of resistance development.
Most corn and soybean fields in the U.S. are planted with herbicide-resistant crop varieties. However, the evolution of superweeds that have developed resistance to common herbicides is jeopardizing current weed management strategies. Agricultural robotics for mechanical weeding is an emerging technology that could potentially provide a solution. A new study from the University of Illinois Urbana-Champaign looks at the types of farmers and fields more likely to adopt weeding robots and at what stage of resistance development.
“The exclusive reliance on herbicides for weed control has led to the appearance of superweeds, and we don’t have anything in the pipeline in terms of new modes of action. If chemical control methods fail, it could result in millions of dollars per year in crop losses,” said corresponding author Madhu Khanna, a professor of agricultural and consumer economics in the College of Agricultural, Consumer and Environmental Sciences (ACES) and director of the Institute for Sustainability, Energy and Environment at Illinois.
Small, lightweight robots that operate under the canopy are highly efficient, have a low labor intensity, and are environmentally friendly. They work by pulling hoes through the soil, thus disturbing the emergence of weed seeds. The robots — which are not yet commercially available for corn and soybeans — rely on artificial intelligence for automation and navigation.
The study focused on controlling common waterhemp (Amaranthus tuberculatus) in corn crops. Waterhemp is a persistent threat to Midwestern cropland, and the weed has already developed resistance to multiple herbicides.
The researchers examined the effect of two different types of weed management strategies that farmers could deploy: myopic management, which considers one year at a time, and forward-looking management, which accounts for future consequences. They also considered weed seed density, weed resistance level, and economic thresholds that would trigger the adoption of robot weeding at the farm level.
“We found that both seed density and resistance level are important for myopic management. For a forward-looking approach, seed density does not matter, because resistant seeds are likely to spread in the future. This perspective does take resistance level into consideration, but almost any level is sufficient to trigger adoption,” said co-author Shadi Atallah, associate professor in ACE.
“Assuming a robot costs $20,000, farmers with a forward-looking management perspective are likely to adopt if 0.0001% of the seeds are resistant, whereas someone with a year-by-year management approach will wait until resistance levels are above 5%,” Atallah noted.
“Consequently, if you’re managing for the future, don’t even bother to look at seed density, just look at the resistance level. And no matter how low that is, you should go ahead and adopt the robots.”
The researchers also looked at adoption rate and intensity over time. Their calculations showed that farmers with a myopic management perspective would not use robots at all in the first six years. These farmers would keep applying herbicides until they are no longer effective and then shift to 100% robotic control — six robots per acre — in year seven, when they have exhausted chemical options.
In contrast, farmers with a forward-looking perspective would begin adopting the robots much earlier and need fewer of them. They would adopt them gradually and not go beyond four per acre. They would use robots to complement herbicide treatment, thus ensuring its efficacy is not exhausted. In year seven, they would use robots on 75% of their land, while 25% would be treated using herbicides.
“We find that myopic management leads to higher profits initially because they’re not investing in the robots. Forward-looking management appears to be worse off at first because they are buying the robots. But that pays off after year six when their profits become higher,” Atallah said.
“Farmers may take the myopic perspective, for example, if they lease their land and must renew it every year, so they can’t really plan for the future. But even for those who are managing on a yearly basis, there will come a point where it is necessary to adopt the robots because other control options are exhausted,” he added.
The different strategies have implications beyond the farm level because resistant seeds can spread to neighboring fields. A forward-looking approach can help reduce the number of resistant seeds and perhaps contribute to a reversal of resistance.
Atallah cautioned that resistance isn’t reversible for all weed species, but for waterhemp, there is a tradeoff when seeds develop resistance; their reproduction rate becomes smaller. As a result, resistant seeds are likely to be outgrown by non-resistant ones if selection pressure is reduced, he noted.
The researchers focused on maximizing profit at the farm level, but a forthcoming study will consider two neighboring farms to understand the spillover effect of resistant seeds. They also plan to conduct a landscape-level analysis to evaluate the impact on larger areas, which will have further implications for policymakers.
Atallah presented the study findings, as well as results from a survey with farmers, in a farmdoc daily webinar.
Khanna is also a professor in the Center for Advanced Bioenergy and Bioproducts Innovation, the Carl R. Woese Institute for Genomic Biology, the Center for Digital Agriculture, and the National Center for Supercomputing Applications at the U. of I.
This research was funded by AIFARMS, an AI Institute at the U. of I., supported by USDA’s National Institute for Food and Agriculture.
