Kevin Fries is the lead agronomic modeler at InnerPlant, a California-based biotech company turning plants into living sensors to better communicate their needs to farmers. Dr. Christopher Seifert is head of software product at InnerPlant.
The views expressed in this article are the authors’ own and do not necessarily represent those of AgFunderNews.
For American soy farmers, pathogens are a $14.6 billion problem, causing $33 in losses for every acre they plant—and that’s not counting the cost of chemical inputs required to deal with infections. Fungal outbreaks such as white mold and frogeye leaf spot (FLS) are an especially big headache for soy farmers: the typical yield loss from FLS is now around 35%, for instance, and can reach as much as 60%, putting a big dent in farmers’ productivity.
The problem has drawn a host of new technologies aiming to improve how farmers spot disease so they can treat it more effectively. Sadly, most of these concepts are doomed to failure because they’re all based on imaging/seeing symptoms, which offers insights too late to help farmers. What farmers actually need is a way to detect the “invisible” indicators of disease before symptoms are visible, so they can optimize treatments.
Fungicides work
In theory, fungicides should solve that problem: a well-timed application reduces FLS incidence by up to 50%. But to deliver results, fungicides must be applied at precisely the right moment. Spraying before disease is present achieves nothing; spraying too late means a farmer may not be able to contain the outbreak.
And of course, fungicides aren’t free. To recoup the cost of a single fungicide application, farmers need to boost yields by as much as 7.5 bushels per acre—but according to the Iowa Soybean Association, blanket fungicide applications only increase yields by about 1 bushel per acre. For farmers, these numbers don’t add up, and they need a smarter approach to actually beat pathogens and generate real ROI.
Stop flying blind
Part of the problem is that most farmers rely on “spray-and-pray” approaches, applying fungicides preventatively during key growth windows regardless of whether pathogens are actually present. One recent survey found that 65% of Missouri’s soybean acres were sprayed with fungicides, mostly to control FLS, but just 10% of those acres were scouted before the fields were sprayed.
Flying blind costs farmers money, because unless they get very lucky, they’re either spraying fields with no active infections, or fields where outbreaks are already well-established. So why don’t farmers take a closer look before spraying? Simple: scouting is labor-intensive and doesn’t usually produce much useful information. Walking the field and spotting the tell-tale black dots that reveal an FLS outbreak only indicates you’ve missed your window to spray.
Pathogens like FLS begin spreading within 48 hours of becoming symptomatic, so by the time a farmer sees signs of disease, it’s too late to take useful action.
Some new technologies aim to make scouting more effective, but most fall into the same basic trap. Fitting cameras to drones might feel more efficient than walking the rows in person. But it’s time consuming and expensive, and tells you the same thing: either there’s no visible infection, in which case you’re left guessing, or there is visible infection, in which case you’ve missed your chance to spray.
What’s more, most drone scouting companies offer just two or three overflights during the roughly 45-day growth period when fungicides are most useful. If an FLS outbreak becomes symptomatic after a drone inspection, the fungi could be happily pumping out spores for weeks before it’s finally detected.
Another emerging method uses AI to spot pathogens. Models trained on photos of healthy and diseased plants can often detect infection more efficiently and accurately than human observers. But at the end of the day, they’re still only detecting visual cues, and there’s no way for AI to magically uncover a signal that isn’t there. If you’re only surveying the fields every few weeks, you’re still going to miss most infections until it’s too late.
Time for a closer look
Clearly, waiting for symptoms isn’t the answer. We’re used to thinking of fungal infections as specks on leaves or grime and fuzz on stems. But by the time those symptoms are present, the infection has already been brewing inside the plant for weeks, and during that time, the soy plant’s immune system has been fighting back. Normally, that happens silently and invisibly. But biotech innovators—including my own company—are now turning that self-defense process into an alarm bell, and creating soy plants capable of alerting farmers as soon as they’re attacked by a fungal infection.
Here’s how it works: when the plant’s immune system responds to a pathogen, it triggers a gene that causes the plants to fluoresce. Viewed through a special camera, you can literally see the plants begin to shine as soon as they detect a fungal infection. Effectively, the plants themselves become sensors: instead of looking for external indicators of infection, we look inside the plant, detecting outbreaks around 10 days before there are visible symptoms.
The benefits are obvious: using sentinel plots of fluorescing plants, farmers can get a head start on spraying their fields weeks before symptoms become visible. With small numbers of plots, continual monitoring becomes feasible, eliminating the chance of infections going unnoticed. And with precise data showing where and when infections begin, farmers can determine exactly which fields—or even which parts of fields—need spraying, requiring less fungicide to get the job done.
The USDA has already shown the power of sentinel plots: their soybean rust surveillance trial found that regular monitoring delivered up to $299 million in annual savings for growers. And that was based on visual inspections. Combining those with pre-symptomatic detection would create a real game-changer for America’s farmers.
In fact, our team recently calculated that with pre-symptomatic alerts, it’s possible to improve total yield by at least 6 percentage points relative to intermittent visual scouting. Assuming average yields, that’s a net benefit of nearly $40 per acre—and some farmers are getting even better results, boosting yields by as much as 25 bushels per acre relative to unmonitored fields.
The stakes are high
For farmers, the threat from fungicides is too big to ignore. But spraying too early or too late increases the chance that you’ll need to spray again later, which increases input costs, erodes profitability, and ultimately increases the cost to consumers. Repeated and indiscriminate spraying also means more chemicals in our soil, and more chance of pathogens developing resistance to our frontline fungicides.
To safeguard America’s farms and its food supply, we need a new approach. Smarter scouting will need to be part of the solution, but it will only work if it’s paired with new sensing technologies that go beyond visual detection. It’s time to stop relying on guesswork. We need to empower farmers to respond to fungal infections as soon as they begin, instead of crossing their fingers and waiting around to deal with the consequences.
