The Era of Big Ag Data Is Here
Can you automate farmers' intuition?
Ron Moore had gotten a late start that morning. The two and a half inches of rain his farm received in the last two days had kept him from harvesting his cornfields, and now he was even more behind in a season that had already gone on too long.
From the seat of the harvesting combine the world is a sea of tall brittle stalks. Moore comes to the end of a pass and deftly steers the machine into the next row. The mowers grab the corn stalks and pull them up into the machine and out of sight, where a rotor spins them around a big cage, stripping the kernels from the ears of corn. A blast of air pushes the lightweight leafy material up and the heavy corn falls through a screen and into the bottom of the combine. The corn pours out of a shoot that extends from the combine and hovers over the back of the grain truck. He does this row after row.
In 1977, when he began farming with his dad and older brothers, Moore’s eight-row combine didn’t exist – the largest on the market only had six rows. Back then large farms in their area near Roseville Il were around 360 acres. Today Moore farms 2,200, some of which he rents. Farms in this area with a diversity of crops 30 years ago have mostly transitioned to corn and soybeans. Moore and his two farm hands plant and harvest 1,300 acres of corn, most of which gets used for livestock feed and ethanol production.
Still, Moore can get a late start to his day and not worry too much. Over the last thirty years advances in farming technology, like bigger, faster combines, have made 3 people farming 2,200 acres a manageable endeavor. And even more recently technological developments have made farming not just faster, but also more efficient.
The inside of the cab of Moore’s combine is downright high-tech. If he lets the combine’s GPS system run the machine, he hardly needs to use the steering apparatus.
The inside of the cab of Moore’s combine is downright high-tech. The usual steering apparatus is there, but if Moore lets the combine’s GPS system run the machine, he hardly needs to use it. To his right three screens are mounted, several flashing numbers and one displaying a crude 2D map that tracks Moore’s progress in the field.
“It’s basically what I call a super computer,” says Moore. “You can make it do anything.” The supercomputer is actually a yield monitor. Whether it’s a potato farm or a corn farm, yield is how much at the end of the season a farmer has produced, and the monitor is part of a system that records what the yield is at any given location in the field.
“When I started farming you didn’t have that technology,” Moore says. He points to an eight-foot square black spot in the field. This is where water stood in the springtime, he explains. In the past Moore would have no idea what that kind of damage cost him in yield. Now the yield monitor can tell Moore exactly how much corn grew in this spot compared to the field around it – it overlays the yield totals on a map of the field, giving a farmer a bird’s eye view of how productive his field has been, parcel by parcel.
This is important to know, because Moore is throwing a lot of money into his fields these days. He spends around $900 an acre on input costs for corn and around $700 for soybeans, compared to the $350 and $250 he spent in 1977. And as the cost has gone up so has the risk. “If you make a mistake you have more to lose now than you did thirty years ago,” Moore says.
But when Moore looks at his yield maps, he will find other areas where yield was low, but the reason will not be as obvious as the water spot.
“Then the next series of questions is why did that part of the field yield less?” he says. “Was it because of a lack of fertilizer? Was it because of poor quality varieties of corn that we planted?”
Those are questions that yield mapping cannot answer on its own.
But over the last few years several big seed manufacturers have invested millions into developing products meant to help farmers make better use of yield data. One of the major players in this space is Monsanto, which just released its new crop management tool, FieldScripts, this spring.
“One of my goals is to be able to go to the FieldScripts program,” says Moore. “That’s what I’m working towards.”
And Moore isn’t the only farmer interested in making better use of yield data. Many farmers were drawn to the promise of yield mapping when it arrived on the scene over two decades ago, only to realize they didn’t have the tools or knowledge to analyze the data on their own. Now, if Monsanto and others can give farmers a way to do that, it could transform the industry.
Farming is a science. Planting, emergence and harvest are all dictated by numbers – seed depth and population density, rainfall and temperature.
Farming is a science. Planting, emergence and harvest are all dictated by numbers – seed depth and population density, rainfall and temperature. They fall into categories of things that the farmer can control and things they can’t.
“By the time we get to harvest everything’s out of our control,” says Bob Strand, a seed and equipment dealer who also farms 2,000 acres in Hinckley, Il. “Planting till the time the corn is 3 or 4 inches tall, a lot of that’s in our control.”
For instance: A farmer seeking to maximize the planting process will perform regular soil sampling, which provides a sense of how much fertilizer to use in each field. Seed depth should be consistent so that every plant emerges within 24 or 36 hours of the one next to it. And the farmer also wants to optimize each field’s plant population, or the spacing of crops, which will differ depending on the type of seed.
But for all its numeric nuances, farmers would historically apply the same amount of fertilizer and seed to an entire field. “You just did it, it was blanket coverage,” says Strand.
Then, in the early 1990’s, yield monitoring came on the scene, giving farmers access to important quantitative information about their fields. “We’re learning now that this little spot and this little spot respond very differently,” says Strand. “And that’s what we’re learning to micromanage.”
Yield mapping caught on with farmers who wanted to maximize their fields’ potential. But those farmers quickly grew frustrated. Yield mapping provided them with unprecedented amounts of data about the productivity of their fields, but they still didn’t know how to act on that information. The yield mapping tools didn’t help them interpret it, and it certainly didn’t offer them any suggestions.
Ted Crosbie was hearing this same story from a lot of farmers. Crosbie was the Director of Global Wheat Breeding with Monsanto until he retired at the end of 2013. “A very progressive younger guy in Iowa said ‘You know, I have enough yield maps to wallpaper my whole machine shed,” says Crosbie. “‘And it doesn’t mean anything because I don’t know how to turn it into decision.’”
Crosbie was having similar issues managing his own 1,000 acres. “At the end of the day my decisions for next year were based on my gut level view of what worked and what didn’t,” he says. “And I thought to myself, ‘you know, this is nuts, I’m a scientist. I ought to be able to make some sense out of this data.’”
Monsanto’s crop management tool, FieldScripts, is designed to help farmers do just that. . FieldScripts doesn’t just churn out data—it takes that data and turns it into useful feedback. It’s too soon to say whether or not the tool is a success, but even an attempt at prescriptive crop management has been a game changer, sending other major seed dealers scrambling to come up with something like it.
When you boil it down to basics, FieldScripts matches the conditions in a farmer’s individual field to one of the many hybrid strains of crop that Monsanto has developed. Using the GIS coordinates (basically, spacial geographic data) of a farmer’s field, two years of yield data, and soil sampling data, it combs through a database of information on Monsanto hybrids to find the ones that respond best to that field environment. Then the farmer sits down with his Monsanto seed dealer to consider the 3-6 hybrid recommendations the algorithm has spit out. When the farmer chooses a prescription, it’s wirelessly transferred to the farmer’s FieldScripts account in the cloud, ready to be used the next time the farmer connects his iPad to the planter.
Monsanto first tested FieldScripts on a small number of farms in 2013 in what it called “Ground Breakers” trials. This spring the product is commercially available in four states – Minnesota, Iowa, Illinois, and Indiana. To use it, farmers need a few pieces of equipment, including a 20/20 seed sense planting monitor, a part added to the planter called RowFlow, which ensures the seed is planted accurately, and the iPad app FieldView – all purchased through Precision Planting, the aftermarket equipment company Monsanto bought in 2012 for $210 million.
Crosbie insists that most farmers already have at least one of these items, and they will see a one-year return on investment for whatever they do have to purchase. The prescription itself is priced at $10 an acre in addition to the cost of the seed the farmer purchases.
Farmers are a skeptical bunch, which is why success in FieldScripts’ first year could be huge.
“I haven’t run into very many farmers who were not interested in this,” says Crosbie. “They’ve been waiting for someone to figure this out.”
That said, farmers are a skeptical bunch, which is why success in FieldScripts’ first year could be huge. Justin Martz was a Ground Breakers farmer who had been creating yield maps of his farm for almost 15 years. But he was in the same position as a lot of other farmers before he entered the Monsanto trials. “I was basically just taking all my data and throwing it on top of each other and saying ‘This area is better,’” he says.
When asked, Martz, a former Monsanto seed dealer, can’t say enough good things about FieldScripts. “It could change the way we farm.”
But he still won’t make a preemptive guess at how successful it’s been for him.
“Until I get it on a computer and am able to sit down and draw out exactly the spots, I can’t tell you yes or no,” says Martz. “My gut tells me it should be better.”
For farmers leery of the program’s recommendations, Monsanto includes an option to run comparisons.
The algorithm is mostly automated, and can’t be customized, but if a farmer doubts the prescription he can plant two “benchmark” rows at whatever rate he thinks might be the optimum population for that field. He can then compare the success of those rows to the ones planted according to the prescription.
That, says Martz, will tell you if FieldScripts knows better than the farmer.
“We’re selling a bag of seed that’s worth however many hundreds of dollars,” says Mike Twenhafel, the Monsanto Integrated Farming Specialist who worked with the trial farmers in Illinois. “It’s their reassurance that they spend their money in the right place.”
Currently, farmers depend on experts such as a seed dealer or their agronomist — soil and plant scientists that help them understand their resources and recommend courses of action. A farmer’s livelihood is almost entirely hitched to those recommendations, so they are understandably cautious about trusting an experimental program over longtime advisors.
Moore says every year he plants a test plot with the hybrid and fertilizer recommendations he gets from his agronomist to see whether the agronomist is correct. “If he is, I’ll continue to work with him and trust his recommendations,” says Moore. If not, “then we may go somewhere else.”
Seed dealers may have the most influence over farmers. They provide information about new products that farmers need to keep their operation current, so they play a huge role in helping determine what purchases get made.
“Customer relationships and customer service plays a bigger role in that than anything,” says Strand. “People buy from people.”
Of course, prescriptive crop management tools could change all that. FieldScripts is designed to combine the expertise of an agronomist with the recommendations of the seed dealer to help farmers make quick, efficient planting decisions. Traditional farming practices could go the way of the horse and buggy – if big companies like Monsanto can get farmers to trust them.
Not to be outdone, Dupont Pioneer — Monsanto’s major competitor — last year released its own crop management tool, Field360, and in February introduced Encirca, which it calls, “a suite of whole-farm decision services.” One of the biggest concerns farmers have about products like these is the use of the cloud to store yield data. And it’s not that they are worried either company will misuse their data. They’re worried about their neighbor getting a hold of it.
The doomsday scenario farmers envision is this: Farmer Jones gets his hands on Farmer Smith’s yield data and finds out how much he’s producing. Then Farmer Jones calls up Farmer Smith’s landlord and promises bigger profits if he kicks Farmer Smith off that land and gives it to Farmer Jones.
“It’s hard to come up with an analogy for any other business where that would happen,” says Martz. “But that is a legitimate fear.”
But much of the strength of programs like FieldScripts lies in the power of numbers. The program doesn’t just write prescriptions based on one farmers’ isolated data; both Monsanto and Dupont Pioneer tools use aggregate data from farms in similar geographical locations to give local farmers the best possible recommendations for their soil environments.
“We do put that information in an aggregate database so that we can look at it and develop and improve both products and services,” says Joe Foresman, director of services at Dupont Pioneer But the data isn’t paired with individual farmer information, and what is collected is only accessible to authorized Dupont Pioneer employees.
Most experts agree that precision agriculture is good for farming and industry will continue to be the major driver of innovation in this arena.
“We know that in our business if we don’t protect that information and continue to support and engender trust with that customer, they can go to a different provider the next season,” says Foresman. “So we can very confidently tell a customer, no, their information is not going to go to a third party, or somewhere in Pioneer that it shouldn’t go.”
In the case of Monsanto, FieldScripts launch lead Dave Rhylander says that back when the company was discussing the idea of FieldScripts with farmers, data privacy was their number one concern. “So we knew that when we brought it to market we were gonna have to have something in place that would make the farmer feel that Monsanto was secure with this data, that the only person who could access that data was the farmer who originally gave it to us,” he says.
But downsides and skepticism may have to take a backseat to the march of technology — most experts agree that precision agriculture is good for farming and industry will continue to be the major driver of innovation in this arena.
Lots of farmers across the U.S are benefitting from “investing into precision techniques and technologies,” says Raj Khosla, a professor in the department of soil and crop sciences at Colorado State University. Precision agriculture “makes a very significant difference in the productivity, efficiency, profitability and sustainability” of farming.
“There’s farmers out there that don’t keep track how much bushels are getting off each field, what spots are good,” says Martz. “The people that keep score are the ones that are gonna be the ones farming 20 years from now.”