Episode 561 05.01.18

Sam Eathington Talks Sensors in Agriculture

Sam Eathington, chief science officer of Monsanto and The Climate Corp., joined Peggy Smedley to talk about how having grown up on a farm helped him understand the importance of data and the need to turn it into actionable insights. He explains that quality information is essential and then you need to know how to put it all together to get the most out of the soil every year.

 To hear this interview on The Peggy Smedley Show in its entirety, log onto www.peggysmedleyshow.com, and select 05/01/18 from the archives.

Smedley:
During my research, I found some information out. Not only are you this data guy as the chief scientist, but you’re a farm boy. Is that true?

Eathington:
Yeah that’s correct. I grew up on a farm in west central Illinois, about an hour west of Peoria for those who might know the area. Grain and livestock farm. I still have some brothers that farm there today. So I always had a lot of background in agriculture. And I went on to school to learn about plant breeding, plant genetics, and really that’s what I spent a lot of my career doing was creating better corn hybrids and soybean varieties for farmers here in the United States.

Smedley:
So you’re a big Cat (Caterpillar) fan, because if you grew up in Peoria, you know Caterpillar very well.

Eathington:
Yeah I know them well. Of course, I did all my degrees at the University of Illinois at Champagne Urbana so I’m also a big Illinois fan for sports.

Smedley:
Go Blackhawks. Okay, let’s get back to the topic at hand. Data is like big oil, right? We have to talk about all this information, and ag is an amazing industry. We can’t live without our food, right? It’s seriously growing, constantly growing. How important is it for a farm to have precise information about what’s happening with weather, or just information that you weren’t able to have before?

Eathington:
You’re spot on, Peggy. Today it’s all about the data, and how you take that data and turn it into an actionable insight where you can go ahead and make a different decision on your farm. So if you think about farming, growing up on a farm, we had lots of different sources of data. We would have feed bag tags, sometimes stuck to a nail in the barn. We had notes scribbled down in notepads here and there when we planted or sprayed something. Soil testing, it might be stuck in a three ring binder. And one of the challenges of that is it’s hard to use that data. It’s not connected. It’s not spatially referenced. It’s not there across years very easy. And so that’s the first step in this process is to get all this data connected; bring it all together so farmers can actually start looking at it, and figuring stuff out on their farm.

One of the challenges growing up on a farm is every field is different. There’s a lot of variability in those fields. You watch yield monitors, 50, 80, 100 bushels going across the field and you’re sitting there trying to figure out, “Well what should I do different?” “How do I make a better decision to raise my productivity across the entire field?” So it starts with the data. That’s what we’re making really simple for growers. Make it easy for them to assemble all that data.

Smedley:
But now you have that access of information in realtime, and that’s the ability to make quicker decisions and know to adjust. Is that a part of what helps? Is that the ability to know something on a farm is completely different and the ability to know instant changes that are going to happen, or something in the environment is changing, and the ability to make quicker decisions? Correct?

Eathington:
That is correct. So that realtime connected information, so if you think about it, I got a crop growing out in the field, and I get a new image, whether it’s from satellite or some sort of drone, I need to be able to process that and look at it immediately, because I may have a disease or an insect problem showing up that I want to go ahead and take action on. If you think about how we used to look at harvest, we would harvest the field, and of course, we’d send in truckloads of grain and we’d get weigh wagon tickets from the elevator, and later on we’d calculate what all the yields were. And today you have that realtime with your yield monitor, and you can pull out of the field and know exactly what happened in that field. Did hybrids perform differently, did your fertility treatment perform different, and your realtime understanding what you should do for the next season?

Smedley:
Looking at that, talk to us about The Climate Corp., and exactly what you are doing with FieldView, because that’s giving some real information now. And I say this all the time, you can get a lot of information now about what’s going on; and you have to know to look at the right information to be able to make the right decisions; because you’re just talking about information. If you make a wrong decision, you can ruin that whole field, right? So it’s important to know the right information. Or that can lead to big losses, right?

Eathington:
You’re spot on. You got to have quality information. That’s what we tell people all the time. If you have bad data coming in from your sensors or information that’s not going to help you. And then you need to know how to put it all together. Having one piece of information by itself, say an image or a piece of weather data is useful, but it’s more powerful when you can assemble it all together and put it in the right context, and say, “Oh, I know what’s going on in this part of the field, I know what happened with my rainfall in that area, I can see what it’s really doing to the crops.” So we make it easy to bring all that data together for a farmer, that’s the first thing we’re doing so that they can actually start analyzing those different data layers, and understand what’s going on in their fields.

Smedley:
Do farmers really embrace the idea of using sensors. The Internet of Things right now, and understanding it’s on their equipment … I was just talking in the opening on their livestock now. It’s changing. It’s not only in their field. It’s on their animals. It’s on their equipment. Do they really get that or are we still having to push farmers into it; and they’re still not 100% sold on the idea of digitizing everything?

Eathington:
Yeah, the farmers I’ve had the pleasure of working with very much see the value of having these sensors and data coming from their fields, whether it’s as you said, something mounted on a machine, just collecting the data out of their machines. We’re looking at sensors that go in the ground and do all sorts of measurements about soil and water, and what’s going on with fertility. And that really relates to the growers … when it’s data coming out of their field. They can relate to it. They understand it. They have more believability in it. And they’re very interested in what’s that’s telling them they should do.

And of course again, the challenge here is, we can generate lots of data, and there’s a lot of people creating sensors and connecting stuff. It’s really taking all that information and doing analysis on it, and turning it into an insight that they can actually make a different decision, either realtime or for next year. And that’s the piece we usually see, where growers want some help—Is tell me what to do with this. I’ve seen lots of these before. I always talk about an image of a field. Everybody can create an image of a field through some sort of remote sensing. But what do I do with it? What does it mean to me? What are these different areas of the field? What do I go do? And that’s the piece we’re really focused on, is how you take it all, assemble it, and turn it into an insight for the grower.

Smedley:
That’s an important point. Because when you talk about the imagery part of it, the ability to uncover valuable insights I think is really key here. When you’re analyzing crop performance; when you’re talking about what you want to do for this season versus next season, how do you identify those early issues that you’re talking about? I think when you’re talking about a field zone, or you’re talking about any of that, how do you help a farmer understand, and position them to understand the imagery so they can then use the information to best to their ability, each time they’re actually accessing information, so they make the best decisions?

Eathington:
Yeah, that’s spot on. It really starts with the research we do. We have our own research farms where we’re doing really intensive studies of, okay let’s image this field. Let’s go ahead and characterize every aspect of that field. Some of our research farms have 200-300 different data layers we’ve collected on those fields to really understand exactly what’s going on.

And from that, we’re learning, and in a lot of cases we’re using machine learning algorithms to help accelerate this. We’re figuring out. Well what is that image telling me? And oh, would I put that image information with weather data and soil data, and maybe some information from this kind of sensor, a little bit about the genetics on their farm, and the fertility. Now all of a sudden that pixel tells me a lot more, and I can turn that into the information that a grower needs.

And so instead of sending a grower lots of images and say, here you go, we’ve gotta move to an industry that says, “Hey we’ve scanned your fields this week. For example, the back 80, we’re seeing this problem develop based on all the information, we think it’s this sort of issue, you might want to go and scout that, or talk to your advisor about that and confirm that. And if so, here are actions that we’ve seen based on past data that might be things that would improve your situation or decisions you want to explore and make, either this year or next year, in your farming operation.

So we have to turn it into, what does it mean? And we’re doing a lot of work on that right now at The Climate Corp., to figure that out.

Smedley:
So it’s kind of tricky because you’re using machine learning algorithms, and I assume AI (artificial intelligence). But how do you couple that with the unpredictability of weather? You talk about the Midwest, we have tornadoes, we have crazy snowstorms; and unpredictable weather. I know the entire world has unpredictable weather. But let’s just take a region right now. Unpredictable weather, in general, makes it hard to know planting, to know what they actually have to do using imagery and the things you’re saying. Are there some things that make it hard for a farmer or a cropper to understand and be able to use sensing technology year from year? Or is it those algorithms that are actually playing a critical role in what they can do?

Eathington:
We talk about that a lot. Thinking about a world of a model, where I’m trying to predict something based on a model. And where I’m supplementing that with a measurement, something from a sensor or remote sensor, or something on a machine or in the field, and bringing those two worlds together because they give you that context of, how should I think about this new piece of data in the, say, growth of this crop. And that’s where the machine learning models actually are extremely powerful, because we can take massive amounts of historic data, we can assemble all that into these models.

And think about it this way: it’s almost daily, being able to run a new outcome of what can happen on the farm. And here’s all the new information that I’ve accumulated in the last period of time. Maybe I have a new satellite image, maybe I have some data from a sensor, maybe a machine ran through the field and I measured something as the machine was going into the field. And being able to feed that into these algorithms and say, “Okay, based on this new data, how’s the crop growing? What might the crop do next week? How is a pest or disease risk potentially developing, given these weather conditions, and the status of the crop?”

And that’s what we’re trying to assemble, it takes a lot of data to do it, and you’re absolutely right, weather is the big variable. How much rain, how much temperature. Look at the Midwest right now. We’re at about 50% of normal planting in Illinois because of the cool wet spring. So how do you need to be able to account for that?

And the way we think about it is, we’re not going to predict the weather, but we can react to all that new information, and update the models continuously, help the grower understand what’s going on, and what they might do.

We then do a bunch of work, especially in our fertility programs, where we studied the last 30 years of weather, and the variability within that weather, and we used that to forward forecast, and start to give farmers, “Hey here’s the range of outcomes you might expect.” We’re not quite yet to a probability world, but we’re moving towards that space. But you want to start giving the growers a range of outcomes.

For example, “Hey I think a gray leaf spot is developing on your crop here in central Illinois. We know at the rate it’s developing, and what the weather looks like over the next seven to 10 days, this could be a problem for you. It could reach a threshold. If it does reach that threshold, here could be the yield impact. Here are your alternatives if you want to try to control it.”

And to get to a world where we start to put a probability around that is where we’re headed. So it will never be absolute, that’s farming, and that’s the weather. But to start to give growers more insight, so that they can make informed decisions, we can absolutely do that.

Smedley:
Do you, see right now, that the expertise of the farmer, or those that have to help traditional farmers, is changing? Do farmers have to get new people to help them manage this type of technology that you’re talking about, to be able to run a traditional farm?

Eathington:
What we see a lot of it and the skills and experience that farmers have, the knowledge that agronomists have, that stuff doesn’t go away. What you want to do is combine it with this technology, and bring the best of both of those worlds together. You know I think about it as, I was actually trained as an agronomist, as an undergrad. And I’m really now becoming a digital agronomist. I use all this new data and information to help me make the very best agronomy decision that I can for a field. But I still have to use my historical knowledge, my experience, my understanding of those fields today to really make the very best decision. So we think about it as really a hybrid between the worlds, and it’s going to take the best of both right now to really drive the insights that a grower needs.

Smedley:
And when you look at that, when you look at all the things that are happening in this space, and the emerging technologies, how are those helping, as you describe, an agronomist maximize yields right now? Is that helping to get the best out of a field?

Eathington:
We are starting to see great examples of where this digital technology is helping those agronomists make more informed and better decisions. A simple thing that I always point to is think about field scouting. For those who have scouted corn fields in July, it’s rather unpleasant usually. And it’s a difficult thing to do at any sort of scale. You walk into a crop, maybe it’s a 100 acre field. You probably see a very small percentage of that field, even with a good scouting process.

Imagine now that your sensors, your remote sensing, possibly even sensors out in that field are giving you information back about what’s going on in that field. Historical data tells you about the variability in that field. And you can now use it to say, well I think this fields okay. I don’t need to worry about scouting it today. But these other two fields, something is going on that doesn’t look right. I need to go to this specific area of the field and understand what’s happening there.

And we’ve seen a lot of cases with our agronomists, they’ll get an early read on something, get out there, go to that specific area of the field and say, “We got a disease problem breaking out here, likely going to continue to spread. We need to go ahead and make a decision about how to treat that now.” Versus waiting until maybe you happen to see it from the road, or maybe you happen to see it a week or 10 days later when you’re looking at parts of the field.

So the ability to catch stuff a lot quicker, have all this information assembled in one spot so you can really make a completely informed decision about what’s going on. And the ability to target what fields, what areas I should really go look at. And those are just some pretty simple things, but today are already helping agronomists and advisors out there make better decisions.

Smedley:
Are we looking at other things like nitrogen and things like that? Are those the things that we talk about, or that you’re helping them in the fields?

Eathington:
We have in our tool today information around nitrogen management. It lets the grower determine what do they want to do on a field and help them distribute that nitrogen in the optimal way across the field. But we’re not stopping there. We’re looking at quite a bit in our research program of how to use other sensors to supplement that information. One that we’re really excited about is our nitrate sensor, it’s really a first of its kind in the industry where we can use that sensor in many different situations, whether it’s a handheld, possibly a soil probe, possibly even in liquid, to actually measure realtime what’s going on.

So just imagine a world where you got all this data that tells me what’s the optimal nitrogen I should use in this field? What’s the optimal distribution of that nitrogen across the field? I’m measuring realtime what the weather is doing to the nitrogen that’s out there, both creating more or in some cases, causing me to lose some of that nitrogen. And building all that into how the crop is growing, and then having a sensor too out there that’s giving you a real measurement of, hey here’s where your nitrogen is at, here’s the status of it, where it is in the soil. And all that coming back together, and you digitally see it and make decisions, it’s going to make farming …

Farming has always been a great thing growing up in it. But it’s pretty cool when you start thinking about how all this data is going to come together and let you make a lot of realtime decisions. And that really impacts your farming operations, and agriculture sustainability going forward.

Smedley:
So there’s this real confluence right now of information that we’re seeing in the farming industry right now. And the IoT is really changing it. So what are you seeing now for your customers? Is it cost savings, is it efficiency, is it giving them more time with their family, is it a little bit of everything? What are you seeing right now, looking at this?

Eathington:
I think it’s a little bit of everything right now, when we think about IoT impacting ag today. Some examples: we’ve connected the cab with our FieldView drive system, which makes it really simple for a grower to assemble this data. So that’s an efficiency play. It lets them assemble data they couldn’t before. So they’re learning about their farm operations that they maybe had some insights about, but now they have real data that makes it clear to them.

We’re seeing things with remote sensing, as I talked about, to let you go ahead and prioritize fields that you should take a look at for scouting. Actually prioritize where in that field you might do your scouting, or maybe take a soil sample, or maybe take a leaf tissue sample to send that in to understand what’s happening.

And then I would say we’re just at the beginning of getting a lot of sensors on our machines. If you think about it, just in the U.S., machines are running across all these acres of our crops many times. And today, a planter is really optimal planting, but there’s no reason why it can’t be doing three, four, five other things as it’s out there in that field planting a corn crop.

And we’re really focused on what are those things that would be meaningful? For example, getting a measurement of spatially what the organic matter looks like in a field all the sudden gives you a totally different view about your field. And totally different way to manage your fertility, possibly even your plant density. So how do you go ahead and let the machines that are already out there working start collection and feeding that data back into the systems?

So in that case, it will be new data that unlocks new insight. I’ve seen all sorts of things with the digital handheld tools. People being at their kids basketball game checking what their fields are doing, remote access in to how a planter is doing, watching their weather data impact their crops, efficiencies in their programs of operations, and decision making. So we’re seeing quite a wide range of how it’s impacting the farming operations.

By |2018-05-23T15:40:05+00:005/23/2018|

Leave A Comment