How to conduct your own on-farm precision research with DIFM

Kathryn: 00:08

Hello, and welcome to the Talking Crop podcast. My name is Kathryn Seebruck, and I'm a commercial agriculture educator with University of Illinois Extension serving Jo Daviess, Stephenson, and Winnebago Counties. Talking Crop is a row crop production podcast with episodes occurring every other week during the growing season between the months of May and September. In each episode, I bring on a guest speaker to discuss topics related to their areas of expertise and any projects that they might be working on.

Kathryn: 00:35

In today's episode, I talk with Dr. David Bullock with the University of Illinois Department of Agricultural and Consumer Economics about the Data Intensive Farm Management Program, or DIFM, for which Dr. Bullock is the program leader. This is an incredible program that enables farmers to easily conduct their own on farm research.

Kathryn: 00:54

And you may be thinking, well, I already can do my own on farm research, but DIFM helps to turn it up a notch. It enables farmers to seamlessly conduct multi treatment level, multi replication research across entire fields of their own farm ground, all to help them understand with statistical certainty how these treatments affect yield and ultimately their bottom line.

Kathryn: 01:17

And the best part is that it requires little to no effort on your end. As long as you have the right precision equipment, all you have to do is drive. And you can get compensated to do all of this. But I will stop raving and let Dr. Bullock give you all the details in the episode. But before we get to it, I would be remiss if I didn't remind you to please reach out to me with any comments or suggestions on who you want to hear from or what you want to hear about on Talking Crop.

Kathryn: 01:43

I can be reached directly via email at seebruck@illinois.edu, which is s e e b r u c k at illinois dot edu, or you can fill out a brief survey at go.illinois.edu/tcsurvey. And the survey will also be linked in the episode description, which is also where you can find Dr. Bullock's contact information as well as some DIFM links where you can go to sign up or to learn more. On the next episode of Talking Crop, I'll be speaking with Dr. John Jones, an Extension soil fertility specialist with the University of Illinois Department of Crop Sciences about updates to corn nitrogen rate recommendations and more.

Kathryn: 02:23

And that episode will air on Wednesday, June 11. So without further ado, please enjoy this episode of Talking Crop, "How to Conduct Your Own On Farm Research with DIFM," with Dr. David Bullock.

Kathryn: 02:36

Dr. Bullock, thank you for being here today and welcome to the Talking Crop podcast. I am very excited to talk with you today about the Data Intensive Farm Management Program.

David: 02:46

You're welcome, Kathryn.

Kathryn: 02:48

So we will just jump right in with a very easy question. Can you please just give us a general overall description of what the Data Intensive Farm Management Program is?

David: 02:58

It's, DIFM is a USDA funded research project where the research is conducted with farmers on their own farms. It's agronomic research that also involves statistics, computer science, economics, all sorts of fancy stuff. But in the end, we are trying very hard to bring very practical recommendations, management recommendations, to growers and capitalize on their knowledge of their own fields, their agronomic knowledge, maybe the agronomic knowledge of crop consultants they work with, et cetera. So the centerpiece of what we do is, there are two center parts. One is just the idea of on farm precision experimentation.

David: 04:01

And on farm precision experimentation is a methodology where you design agronomic field trials, but on a much bigger scale than what is traditional. People have been running small plot trials for hundreds and hundreds of years. A small plot trial is when you take part of a field and you know, you take various parts of it and maybe you put one amount of nitrogen on there. And then in other places you put another amount. And so maybe you have five, six amounts and you might have three or four replications.

David: 04:43

And then you see how the nitrogen rate affects yield basically, which is a very important piece of information. Those small plot trials are usually less than an acre, you know, often maybe two basketball courts. But they've provided a great deal of the agronomic knowledge we know of, and they're important and they'll continue to be important. The problem with small plot trials is they're small. Okay.

David: 05:14

So if you run an experiment on piece of land, the size of two basketball courts, it will tell you something about how the crop grows and what affects crop growth on that little piece of land. But it may not say a whole lot for something across the field, let alone across the county or the state. And so what we are doing is we're using precision agriculture technology and a lot of stuff we've developed to help farmers and crop consultants that they're working with and Extension people actually make these trial designs on our website. It's not hard. It's that, you know, in a few minutes you can make a trial design, something that used to take people hours if they tried to do it on SMS.

David: 06:19

But literally they can design trials once they figure out what's going on in five minutes. Okay. And it's being done now, quite a bit. We had one crop consultant that conducted 25 field trials this year, 2025. That crop consultant is down in South Indianapolis a little bit.

David: 06:50

Another one is up in the northeast corner Of Indiana and the northwest corner Of Ohio. And he ran, I think 24 field trials with seven, eight different farmers. So we've created this system, I'm going to call it a cyber infrastructure, that makes it really sort of pretty easy to run great big trials now on whole fields. So you can imagine how does that work? Well, let's say you wanted to try look at a seed rate trial for soybeans.

David: 07:27

A lot of times a farmer does little things like this and maybe they have one strip that is one rate and another strip that is another rate. They do a little, farmers are natural experimenters, they're always things and seeing what happens. Well, with our tool, they can do that, but probably in better ways because there's a lot of science and statistics, et cetera, behind it. It's an easy way to design really good trials. And we like to call them checkerboard trials.

David: 07:58

Okay. Because what it does is it puts a bunch of plots, often the plots will be as wide as the applicator, so let's say 60 feet, or the planter, let's say 60 feet, and they might be 250 feet long. Okay. Well, a few hundred of those will fit on 80 acres.

David: 08:17

Okay. So all of a sudden you have many, many replicates of if you have five, six rates, can have dozens and dozens of replicates of the same rate. And so all of a sudden you are, as a farmer, you're just driving. You got yourself hooked up to GPS. And then on your machine, it says, look, there's a computer file that says, look, when you get to this longitude and latitude, put down a hundred thousand seed per acre for a while, but then 250 feet later, put down 80,000. 250 foot later, put down 140,000. Okay. And you do that all over the field. Okay. So it's a gridded trial, not a strip trial, but a gridded trial, which is a lot better statistically because you get a lot more changes in small areas.

David: 09:14

So if you're trying to learn how to do site specific management, strip trials are not near as good as these because strip trials don't have a lot of variance in the rates in small areas. And so you can learn more about crop growth in smaller areas, which is what site specific management is all about. So basically farmers can make these trials on a website. There's a lot of guidance, we provide some training, it doesn't take that much. Or crop consultants might work with farmers and do lots of fields for several farmers.

David: 09:55

And then the farmer uploads that what is often called a prescription and they just drive. Using technology that anybody who's ever put in, you know, a commercial prescription can do. Most Illinois farmers or lots and lots of Illinois farmers are very familiar with the process and the equipment. So all of a sudden instead of running small plot trials that take a lot of work, it takes, you know, professors and graduate students out in a field in March in the rain, putting out little flags, sticking little flags in the ground and measuring tapes and doing a whole lot of work to get data about on a piece of land the size of two basketball courts. All of a sudden, we can have a farmer easy put in a whole field just in the amount of time that it takes them to run the machine through the field.

David: 10:53

And so what that means is we get tons and tons of data and it's really labor saving. And so we can do very big things instead of smaller experiments. So that's sort of the main thing that on farm precision experimentation is one of the principal components of what we're doing. The other principal component is, we call it the DIFM.farm cyber infrastructure or system.

David: 11:25

So anyone who wants to can just put DIFM.farm into their search engine and they'll see our website and the website will pop up, they can create an account easily and then do all sorts of things. It does much more than just design trials. The cyber infrastructure takes in data, allow farmers to not just experimental data, but they can bring in electrical conductivity data, they can bring in soil sampling data, all these kinds of data, but we have a way to process that data. So it all comes together in a nice way so you can do things with it. And so this data processing can take a long, long, long, long time if you're just trying to do it, even if you know what you're doing, if you're just trying to do it yourself.

David: 12:21

But we've written computer programs that do that, you know, right away with the push of a button. And then that analysis then, we've also got we've written a lot of code, some artificial intelligence stuff, some statistics stuff, and that code will take that data and analyze it, okay, and think about what it's saying. And then also at the push of a button our system will write a report that is readable, very readable by farmers, and it'll say, look, this is what the data seem to be saying. Sometimes the data aren't telling the farmer much. Sometimes the data are telling the farmer a lot.

David: 13:06

And so it's a way for farmers to run high quality experiments at minimal cost. And it's free, so whatever cost they're bearing, they don't have to pay us. As a matter of fact, we've always been able to use funding to compensate farmers if they have any losses, okay? And we've been paying farmers a thousand dollars just for participating. We're doing this because NRCS, which is the USDA's Natural Resource Conservation Service, has a program which is called their On Farm Trials Program in which they want to subsidize farmers to do research on farm.

David: 13:52

And so that's why we pay people a thousand dollars for participating and we can compensate them if they lose any money by running a trial. So DIFM.farm Farm at the push of the button all of a sudden you can bring in data, process data, analyze data, have reports on data. So something that would have taken dozens and dozens and dozens of hours, we can do it with just the pushes of some buttons. We're extremely proud of this.

David: 14:23

We started thinking about this about ten years ago and I have managed to find some just terrific people to write the computer code and put this all together. I think it's impressive. I think it's great. I think a lot of farmers using it and crop consultants are using it are doing so very enthusiastically.

Kathryn: 14:45

I can only imagine, this sounds like a just a win win across a lot of different areas in terms of it being free, farmers can be compensated for it, being paid to participate as well. And I think too, you know, I speak with farmers and they often talk about how that small plot research, while valuable and interesting and informative, it doesn't always mean a lot to them because they recognize the fact that they're applying different treatments across, like you were saying, a wide swath of area. So that must be different in terms of the results in comparison to the small plot research. Like I said, just a win win across the board. So how can farmers participate then?

Kathryn: 15:31

I think a lot of people, their ears probably pricked up to the fact that they can be compensated or paid. So how can farmers get involved in the first place, but then also that monetary aspect of it as well?

David: 15:43

Well, they go to DIFM.farm, difm.farm , and you can open up an account. It says sign in and you sign in, you open up an account, you put in your farm's address, you put in an email, that kind of thing. And then the tool is open to you, right? It includes making some agreements and by the way, farmers' data is farmers' data.

David: 16:15

We don't sell their data. We don't give away their data, the data is held private. We reserve the right to use data for academic research but the farmers it's all anonymous. Or if you want, go to your consultant and say, Hey, I would really like some help with this. And the consultant might find, Oh yeah, hey, I can do this with one farmer, but I can get pretty good at this and do it with 10 farmers and really help them all run these experiments.

David: 16:43

So you get started by just signing up and getting started. Now you're welcome to email me, David Bullock. My email address is dsbulloc@illinois.edu and I'll answer your email and get you started. Or you can just get started and that's how you do it. And I know it's hard to believe, oh wait, this sounds too good to be true.

David: 17:13

Well, the thing is that tax dollars are going into this, right? The USDA NRCS wants farmers to do these trials, okay, because they see a future in it. And so that's why they give us funding. Really, I don't think there's a lot of downside to it. You know, obviously it takes a little learning.

David: 17:34

It, you know, it takes a little effort, but not that much. The learning curve is not steep.

Kathryn: 17:41

So once a farmer makes that account and like you said, simply gets started, do they have to take any steps to get that compensation?

David: 17:50

Yeah, I mean there's contracts and things that you sign and then after harvest we do the analysis and we, and our system writes a report to help them think about what's going on and what the management implications of the data are. Then it also, in I think very reasonable ways, estimates if they lost any money and then, you know, we write them a check. Now I got to say, this is what we've always done. Okay. As you know, right now, a lot of part of the USDA is frozen.

David: 18:31

Okay. So if they don't give us, if we don't get money next year, then, you know, farmers are still welcome to use our system but we can't, you know, we can't pay out of the money that is not given to us. Right? Mhmm. So a lot of this depends on what the federal government decides to do.

Kathryn: 18:50

So you mentioned the economics of it, which I think is one of the biggest thing that farmers are gonna be looking at with these experiments is, are these changes going to be making me or saving me money? So when it comes to designing their experiments, I guess what are kind of the parameters? Are there limitations as to what they can do? Or are there just kind of general characteristics that you can describe of what these experiments will typically look like?

David: 19:15

Well, typically they would go in and we'd need to know things like, first of all, what's the crop, right? It's soybeans. What input are you looking at? I want to do seed rate, a seed rate experiment. Great.

David: 19:33

What's the size of your planter? It's 60 feet. Does your planter have, you know, individually controlled sections? Yes, it has 24 independently controlled rows. What's the size of your harvester?

David: 19:48

My harvester is 30 feet wide. Okay, we go with that. What's your usual seed rate? Oh, I usually put down 120,000 pounds per acre, 120,000 seeds per acre. So there's various, not tons of them, but there's various pieces of information we got to know to design these trials, especially about equipment sizes and what farmers usually do.

David: 20:16

But that, you know, is just some questions, it's not a long list of questions and so they put those in and then, you know, they can start designing a trial. And we're always around. We have people whose job it is to work with farmers and to answer questions and to troubleshoot, etc, etc. So I think we have a real personal element in this. It's not just some anonymous computer thing.

David: 20:45

I mean, it can be, but there's people involved too and we're around to help.

Kathryn: 20:51

And that's great because I think a lot of people, it might be a little bit of a barrier to them, kind of that overwhelming thought of the data analysis, which when I was mentioning that win win situation, I think that's part of the wins is that you get true statistical data analysis done on these experiments. And I know a lot of farmers will do kind of like you said, their own strip trials where they just do a comparison of treatment to no treatment. So this is more intensive, which can be overwhelming for a farmer but like you said, the program just does it and the farmer just drives and then you get the data analysis at the end of the program as well.

David: 21:27

We've been doing this for almost ten years and we have made a lot of mistakes. All right. And we've learned from those mistakes. And we've presented farmers with things and ideas and statistics that they really couldn't use or didn't find useful. But we've worked on that too.

David: 21:44

And I think we have a very usable system. Now, if you're a researcher and you want to use this, you can get reports that have all kinds of statistical analysis and all kinds of, you know, technical jargon, etc. But we have worked very hard to make what we're putting out understandable, you know, to farmers and crop consultants. Now look, some farmers are super tech savvy and super computer savvy and they're not going to have any trouble working with this system. But if, if you don't, you know, if like me, you can't figure out how your iPhone works, then, you know, get your crop consultant going because a crop consultant has an incentive to work with lots of farmers and a crop consultant can learn this stuff.

David: 22:31

He might attend, you know, some two, three training sessions, but, and then the farmer has to worry even less about it, you know, if somebody else is taking care of much of the administration.

Kathryn: 22:44

And I'll also add as well, I'll put a plug in for Extension too, if you have an Extension educator in your area as well we can we can help as well.

David: 22:53

Absolutely and we are working with a couple of people in University of Illinois Extension. One is Meagan Diss, and she's up in the Northwest Illinois R and D Center. Another one is Talon Becker. He's here on campus, and they're commercial ag specialists and they're working with farmers. So I think that this could really be a way for Extension.

David: 23:22

And this, we're dealing with Extension people in many states, right, to work sometimes directly with farmers, sometimes with crop consultants, but to really facilitate this happening. I mean, I just can't tell you the kind of great data we get that in the past was just, you know, it was a dream. And I don't know, you know, like I said, sometimes we don't learn much from the data. Sometimes we do.

David: 23:52

Had a lot of farmers a little further west where it's drier and they're really over planting soybeans, right? Well, who are they listening to? They're listening to the seed dealer, you know? And sometimes depending on the weather, we're finding, oh, hey, this person, they're putting down 130,000 seed, but really 85 would be fine. So there's just lots of stuff that can happen, right?

David: 24:23

It might be that, you know, when you're applying some fertilizer, it really matters the slope of the terrain, right? So we might say, look, where it's steeply sloped, we think it might be good to try this nitrogen rate, but then if it's a flat part of the field, it might be good to try this nitrogen rate. And so it enables a farmer to see what happens and those farmers are thinking about that anyway, right? They're thinking, okay, this is a steep thing, the nitrogen is going down with the water downhill. What does that mean and where exactly?

David: 24:58

And we can help farmers think about those things. Now listen, we're not trying to create an answer machine. We're trying to create a decision tool, something that farmers, along with their consultants, they can apply what they know about their fields and they can get data and they can think about things agronomically, you know, and they can see if what they're seeing in the data makes sense to them. So we're trying to help people make decisions. I mean, if somebody wants us to say, we're the experts, here's your map, go do this, we can do that.

David: 25:39

But not that's not really what we're trying to do. We're trying to create something that farmers can use to think with.

Kathryn: 25:47

And I think a lot of farmers would really appreciate that because, like you said, they're natural experimenters. They're going to be, they're always asking these types of questions. To give them more of that autonomy and power to be able to do so easily, I think is just, it's really invaluable. So you mentioned that this started about ten years ago and you mentioned that you've also learned a lot along the way. So can you kind of run us through a little bit of that history and some of the main things that you've learned and how it's kind of evolved in those ten years?

David: 26:18

Actually, the thoughts about this go back about thirty years. And, a lot of the vision came actually from my brother, Don Bullock, who, he's retired now, but he was in the Crop Science Department at University of Illinois. So we were both in the College of ACES. And, we started work together because I'm an economist and he's a crop scientist and precision ag technology came around and there was a lot of bad work being done. You know, agronomists are not good economists and economists are not good agronomists and there wasn't much multidisciplinary stuff going on.

David: 26:58

And so, you know, it frankly, there was a lot of bad research. And so my brother and I started working on this and, you know, when you're doing something multidisciplinary, your first reaction is, boy, are they stupid, you know? But not the people you're working with are not stupid, they just are speaking a different language and seeing things differently. And so the hard part about multidisciplinary work is getting people to communicate. And since you can yell at your brother on Friday and come back and work together on Monday, that really helped.

David: 27:34

So, my brother, Don, he told me when the precision ag technology came out in the nineties, he told me, this is incredible. It's literal space age agriculture, right? Literal space age ag. They're using satellites, he said, but they have no idea what they're doing.

David: 27:55

They don't know where to change nitrogen rates. They don't know where to change seed rates. There just has never been enough data. The data's been coming from strip trials and small plot trials, and nobody knows. And he was able to do that because he was a skeptic, which is good.

David: 28:15

For a long time, especially with like nitrogen application algorithms, recommendations from land grant universities, frankly, people believed in like yield based methods, but everybody believed it just because everybody believed it. Right? And what we're finding out is a lot of things that everybody believed weren't right. Okay. So my brother came up with the idea.

David: 28:45

He says, what we need, we need more information. We need more data. And the beautiful thing is, and if we don't have more data, we cannot do site specific agriculture profitably. He says the beautiful thing is we can use this precision technology to actually create the data that we need to then think about how to use precision technology profitably. So you can get the information you need to make the technology profitable.

David: 29:13

You can get that information with the same technology. So he came to that insight or we came to it together. And so he wrote some computer software, and he thought that researchers would use it. And it was called the Enhanced Farm Research Analyst. And it was kind of a simple Windows based thing.

David: 29:38

And what happened over time is they didn't get more funding, and then they needed to update it to a new Windows. You know, Windows is always changing. And so it kinda fell by the wayside. But I had the idea, you know, we could do this a little differently. We should aim this at farmers and crop consultants, and we need to create a system that makes this easier.

David: 30:06

And I started thinking that about ten years ago. Okay. So my brother came up with the idea and some Australians were doing something similar at the same time. And then I have carried a lot of it out and had that vision and helped make it happen. And that's been more over the last ten years.

David: 30:25

So we thought about it for twenty years and then we finally did something about it.

Kathryn: 30:30

In those ten years then, since this kind of really ramped up, what are some of the main things that you learned? I know you mentioned how some of it was like providing farmers with data that they could actually, you know, utilize. Are there any other things that really stick out to you that were really big learning points for you?

David: 30:47

One thing we wanted to find out was about nitrogen. Right? Are farmers really applying more nitrogen than is profit maximizing? And ten years ago, and still there's a lot of talk about that. Look, a lot of people think there's a win win situation out there.

David: 31:07

Farmers are putting on more nitrogen than is profit maximizing. So if we can get them to cut their nitrogen rates, then they'll make more money and we'll have less, you know, environmental problems with nitrogen running down the Mississippi River. Okay. That sounds great. And we certainly wanted to look at that.

David: 31:27

But where's the data to tell you that? Well, we've done, I don't know how many hundred nitrogen rate trials over the past many years and what we find is some farmers some years are yeah they're putting on more nitrogen than they need to to maximize profits. But we're also finding farmers underapply. Right? So this tells us something really important.

David: 31:54

And it's important for policy because if we want to clean up, you know, the nutrient loss mess, We need to know how farmers are being affected economically and we can't just sort of say, yeah, farmers are putting too much on, there's a win win situation, they should just put on less. We're not finding that. And that's an important piece of knowledge. It's an important piece of knowledge for policy because it means solving the nitrogen problem is a little harder. Okay? If the government wants to solve the nitrogen farmer, then they're gonna have to, they are, the government already subsidizes farmers to, you know, you can, there's all kinds of USDA programs and NRCS programs that farmers can take advantage of to manage things sort of more, you know, environmentally sustainably.

David: 32:46

It tells us that if the government wants to cut the nutrient loss, it's probably going to have to subsidize, which it already does. But now we're starting to have tools where we can figure out good ways to subsidize, efficient ways to subsidize. So you don't give out way more money than you need to, but you give out enough so that farmers are fairly compensated, you know? So these are things both we're learning and we're aiming towards. So I would say that's a major finding.

David: 33:22

I don't know how you get that finding unless you run all these trials. And I can tell you that no, on average, I don't know about average, but do lots and lots and lots of, you know, do a great majority of Illinois farmers put on more nitrogen than is economically optimal? No, some do, you know, some don't. And so there's one answer. We are finding first of all, just for individual farms, it's like, hey, wow, we think you ought to sort of treat the west part of the field and the north part of the field, you know, a little different from the rest of the field.

David: 34:02

And this is what we're seeing. This is how we see yield response to the input different in these different parts of field. And we can help farmers start to do things site specifically. And I'm telling you, if you just go to your consultant or even if you go to your land grant university and take their recommendations about how to do things site specifically, I have a low view of those algorithms that give farmers advice. I don't think they come from very much data.

David: 34:34

Okay. And so, and so what? I think that we can get data that starts making sense and helps farmers. If you want to do precision ag, you want to do site specific farming, then you need a lot of information and we can help farmers get that information. Now, another thing we're finding is often on flat and black Illinois fields, frankly, there's not enough of differences across the field in soils, etc. to really justify managing site specifically.

David: 35:12

Now, course, John Deere and and Raven and all these other companies that create precision technology equipment, they probably don't want to hear that in a lot of places. You really don't need to manage site specifically. But that's what we're seeing. On the other hand though, that site specific, you know, variable rate applicator, the yield monitor, I think a big value of that is the information you can get from data.

David: 35:51

Okay. And so that's kind of a different use of this technology. Instead of just managing different parts of the field differently, you also use the technology to get information and that's revolutionary.

Kathryn: 36:08

Absolutely. I kind of like how you were saying how it's, you're kind of flipping the script in a sense instead of, you know, going about it, doing the precision ag to get the data, you're kind of getting the data to improve the precision agriculture.

David: 36:24

Yes. But you're using the technology to get the data. And kind of ironically, you need the data to use the technology well and make more money. Now obviously, eventually you're going to stop experimenting and start using what you learn. But there's a million great questions about what's the best way to, how much do you experiment? When do you start putting your knowledge to work to stop experimenting and make more money? A lot of great questions there.

Kathryn: 36:54

Sure. Well, Dr. Bullock, I really appreciate you coming on today and talking about this. Again, I think it's a really great program. It's that win win across the board.

Kathryn: 37:05

And I hope that folks listening garner an interest in this and, look into the program and can get started and hopefully you'll hear from them in the near future.

David: 37:14

I would love to. I look forward to that.