In integrated weed management, farmers employ a diversity of weed-killing techniques, including tillage, cultural practices, and methods for depleting the weed seed bank, rather than depending solely on the spraying of Roundup or another single tool.
But diversification isn’t the same as true integration, cautions University of Nebraska-Lincoln weed ecologist, Steve Young. Why? Because most integrated weed management practices still manage every weed the same, regardless of location or season, and they’re usually deployed one at a time rather than together. Young is now hoping to change that by developing a fundamentally different approach to weed control.
In the automated systems he envisions, sensor and computer technologies onboard a tractor would first categorize each plant in a farmer’s field as either weed or crop, and then go on to identify the species of weed. Once those identifications were made, one of several weed fighting tools located on the tractor could be applied to individual plants based on their biology. If the system identified a weed that’s resistant to Roundup, for example, it could be sprayed with a different herbicide. Or an onboard cutting or flaming tool could be used to kill the plant instead.
In other words, the system could target different weed-killers to specific weeds, Young says—similar to how variable rates of nitrogen are applied in different sections of fields today based on variations in crop nitrogen status.
“This is all about trying to get more precise with our weed control practices,” Young says, and in fact he began developing his ideas while working as a postdoc at the Center for Precision Agricultural Systems at Washington State University. His advisor, Fran Pierce, was “really into automation,” he says, and it got Young thinking about ways to apply automation and robotics in weed science, as well.
Young has since been studying the micro-rates of chemicals and mechanical treatments needed to kill single plants. “If we can figure out these parameters, then they’ll be ready for the engineering part of it,” he says.
As with other precision agriculture technologies, there could be many benefits. Applying micro-doses of herbicide to the leaves of single plants would eliminate spray drift and the leaching of chemicals into groundwater. Weeds would also be less likely to evolve resistance to herbicides, because less chemical would be required overall and only targeted weeds would receive the dose.
Young is not advocating the system as yet another “silver bullet,” but as an entirely new approach in cropping systems where multiple weed-fighting strategies are available for use at the same time. “It’s revolutionary because all the tools for controlling weeds are put on a level playing field; none is regarded or relied on more than another,” he says. “And because the tools are interchangeable, this system can be used in conventional, organic, or any other cropping system.”
If it all sounds a bit futuristic, the future isn’t too far off, he adds. Development of advanced spray nozzles and micro-application technologies is well underway, and the engineering literature is already stuffed with robotics and sensing technologies for automated weed control. The big hurdle now is to identify weeds very precisely in the field, especially under rainy or windy conditions when plants are swaying all over the place.
“One of the challenges to really moving this along,” Young says, “is to account for that variability in the field, not only with identification, but getting things on the target plant.”
Fortunately, weed scientists and other biologists understand variability very well, and Young is now trying to encourage more collaboration between them and engineers. Engineers necessarily focus on machinery—how it’s built, how fast it goes, how well it detects a target. Meanwhile, weed scientists understand plants and their responses to different weed- killing treatments, or the “application side of things,” Young says. “The engineers are bringing biology to their work, but in my opinion they could get there faster if they worked with weed scientists.”
And weed scientists? Young says they need to broaden their thinking beyond traditional practices and chemicals, and toward automation—especially given today’s advances in technology.
“If you go back 10, 15 years, no way—this would not be possible,” Young says. “But because we have computers now that go faster, and things like weather proof cameras that can sense things down to the micro-level, we now have the ability. So let’s use it.”
A seminar series on automated weed control will be held at University of Nebraska-Lincoln this winter, with two sessions streamed live, online. Learn more.