Hello and welcome back to Advancing Nitrogen Smart, the special podcast series from University of Minnesota Extension. I am Jack Wilcox in Extension Communications.

Today, I'm here with Brad Carlson, Extension Educator out of Mankato, and Jeff Vetsch, the lead researcher here at the Southern Research and Outreach Center in Waseca, Minnesota.

Today, we're gonna talk more about Variable Rate Technology. Brad, why don't we start by talking about sensors?

Yeah. So the the the first thing we wanna talk about is is is active sensors, and and the idea here is that these things are emitting a light beam. It's reflecting off the plant, and then they're kind of measuring what's going back. And so, you know, the idea is that they're supposed to be very timely because, for instance, you could potentially mount one on the front of an application unit. It can be sensing the computer, can be digesting the information it's giving, and it can then be translated into a nitrogen rate that's applying right at the back of the machine, which, you know, that certainly is an ideal situation from a timeliness standpoint. However, the research has kinda indicated that there's definitely some drawbacks about this.

And so one of the reasons we wanted to have Jeff here this morning is, you know, Jeff has actually researched these products. And so, Jeff, why don't you tell us a little bit about some of your experience with this?

Yeah, Brad. We started doing some work with colleagues from other states, primarily Oklahoma State, Nebraska, back in the early two thousands, and we used the Green Seeker Sensor fairly extensively, also the Crop Circle, and currently, the Crop Circle rapid scan unit, we have one of those as well in the department and here at the SROC. We've actually done a fair amount of this over the years, and there are certainly are some challenges. These these sensors work. I mean, they can create a normalized difference vegetation index or NDVI of the crop, and they can calculate that very rapidly with today's computing technology.

And as you said, Brad, they can create a as applied map, basically, on the go and make recommendations. Some of the challenges, though, is that these sensors not only were developed for corn, but they were also developed in wheat, especially in the Central Plains, Oklahoma, and that area. And wheat, if you understand how winter wheat grows, the amount of biomass is keyed in the springtime or about this time of the year, late February, early March, to determining how what potential that yield potential that crop has. And a lot of their sensor algorithms that were developed to help determine how much nitrogen should be applied were based on those NDVI or biomass readings. So a sensor like the Green Seeker and others, they can measure the amount of plant material, but at some point, like in a corn crop, especially here in Minnesota, when the corn gets to about v seven, v eight, it's pretty much covered the ground, and the the amount of biomass makes that sen sensor basically reach almost saturation, and then it's very difficult for it to identify the difference between really good corn and really poor corn when the biomass is basically saturated the amount of light that's being emitted.

So there's some other alternatives then that that work better. We go way back to the eighties and nineties, and we use the SPAD meter, but the SPAD meter cannot take real time measurements like that in the field. It's it just doesn't have the practicality because it can't do it. It's not emitting light. It's just basically looking at the greenness of the leaf, and it takes someone to physically do it.

The other things that do work are if you have a sensor, like the rapid scan or some others that can measure NDRE, where they're looking at different wavelengths, and they're focusing more on the greenness of the corn, or the greenness of the plant. So they use not only the amount of biomass that's there, but also the green color to determine how healthy that plant is, and then they use that in their algorithms to determine the amount of nitrogen. But ultimately, Brad, it comes down to and I think this is the next talking point is, when are we gonna make these applications? And when does the plant best tell us how its end status is? We know the uptake pattern of corn is such that it doesn't really take up a lot of end until it gets to that rapid growth phase between v six and r one.

So we need the plant to get going, you know. The plant has to be at least v eight or v seven to start showing some end stress. And then that period of time between that initial stress and r one is so crucial that if we starve the plant just a little bit at that time, we can actually cause a yield reduction and not a yield advantage. Now how they get around this in states like Nebraska and other places where they do this use this technology a lot is irrigated agriculture. I mean, they can spoon feed this crop as the crop is telling them it needs more N.

They can put in on with their irrigation system, and that's very efficient, and it works quite well with these active sensors. But in dry land ag, it's a little bit tougher.

Well, and even in Minnesota, I mean, we've found that in irrigation situations, like, that we can see a positive yield advantage to up to seven splits of nitrogen, which that's not even corresponding necessarily to using any kind of sensor technology that's just kinda related to the, you know, the system. And and so that's that's another component of this. You know, I've been told, and and I'm not a very deep plant physiologist or geneticist, but I've also been told that the long season hybrids grown down south, a hundred and twenty days plus, aren't quite as sensitive to early deficiencies that some of the shorter season, the hundred days and ninety five days we grow here are.

But the other aspect of that also people need to realize is that when you're down south, a very low organic matter compared to what we're dealing with in Minnesota. I mean, we we're here in in Southern Minnesota four. I know you guys got 6% in some of the places on the research station here. You get down to to places like like Arkansas or or Oklahoma, you know, you're down to 1% or less. And so really, if the crop is short, if you don't put it on, the soils aren't going to supply that.

In Minnesota, you know, when as long as it's warm and moist in that top layer, we're gonna continue to mineralize nitrogen out of that soil organic matter and continue to feed that crop. And so it's just a dynamic that that just really makes it difficult to use these things accurately in Minnesota. And so so, Jeff, let's let's talk a little bit about some of the the research results that you had in the past.

Yeah. Some of the earlier things that we did when we were looking at the green seekers, working with the green seeker and other sensors, is we had a lot of experience with the SPAD meter. So we had a pretty good idea of when the SPAD meter was pretty accurate at telling us, and basically, there was a lot of literature where, of course, published as well. The SPAD meter, once you get to about v 12, v 13, v 14, and up to r one, it's extraordinarily accurate at telling us what the nitrogen status of the crop is. So we suspected that these active sensors would be the same, but again, we go back to the fact that that some of the sensors, what they're measuring is the amount of biomass, and then that can reach saturation so that integrating that greenness of the crop and that those wavelengths in the scent active sensors is important.

So what we did in some of our earlier work, and this was probably in 02/1989, in that area, is we just set up some some NRAID trials where we kept some nitrogen, held it back, and we held it back to our normal split application time or side dress time around v six, v seven. But then in a couple other studies, we held have had some treatments in there where we held all the end back until v 12 or later, or we put on just a very little amount, like 15 or 20 pounds like a starter rate. Now in corn after beans, this usually was okay and didn't have a detrimental effect because often the plant isn't taking up a ton of N till it gets to v seven, v eight, and the soil can probably provide some, as as Brad said, especially in our high organic matter soils, as long as we've had decent climatic conditions. But on occasion, we would actually see we're holding back this n until up to v 12, and even sometimes at, like, v seven or v eight would actually result in a yield penalty. And it kinda makes sense if you have the wrong climatic conditions.

The plant's been stressed. Some hybrids are really sensitive to setting the number of kernels that they're gonna produce around v five, v six, and if they're showing any stresses, they're gonna produce fewer kernels. And, I mean, we didn't do kernel counts, so we don't know if that was the reason why we had reduced yields, but we did look at the amount of precip we had post sidedress in these later applications timings, because you it's critical if you're gonna pull back, you know, 80% or 75% of your N or more to these later stages to let the plant start to show some symptomology of N deficiency to help you sense it and get an idea of how much more N it's gonna need. It's critical that you get timely rainfall in a rain fed system right after that side dress application, especially if you're gonna delay a lot of the n till till v 10 or v 12 in some cases, which we did in some of these studies. And it didn't happen all the time that we got a yield penalty, but it did happen a couple of times, and it was enough that that it makes it a risky system to manage.

And personally, I think of and maybe we'll get into this a little bit later in this podcast, but I think of these sensor active sensor units similar to help maybe farmers identify the areas of their farms or fields that need a rescue treatment after years like the 2024 when we had 20 inches of rain in May and June, and we knew that a lot of these fields lost maybe as much as half of their nitrogen. But to try to manage thousands of acres, you know, in in the Upper Midwest with a management system like this would be a definite challenge. Well, I mean And adds risk.

Right. I mean, if you wanna define a rescue treatment as variable rate, I guess you probably can, but no one wants to walk into the season saying, yeah. We're gonna do a rescue treatment this year. I mean, that that's that's that's just salvaging what you can out of a crop. That's certainly not a game plan for making money crop farming.

And and so, you know, these these things maybe do have some utility if you get caught and you just have to do that, but certainly, you're not gonna walk into the season saying, yeah. We're gonna end up doing a rescue treatment this year.

No. You're right, Brad. And and I would also add to that that when you think of rescue treatments, especially the way the the economy for crop production is right now, the margins are really tight. If you're gonna go out there and do a rescue treatment on a field, knowing where to put a little extra in and maybe where you don't need it is gonna be maybe crucial to help making that give you actually give you a return on investment. And there's a lot of, you know, work that's been done across the Upper Midwest by some of our colleagues in other states, and they've shown that these rescue treatments can go out really late and sometimes still give a nice return on investment.

But, again, you're exactly right. You're not gonna plan for rescue treatment. The rescue treatment is what it is. It's a reactive I'm gonna try to get the best I can do out of this, and I had a a lot of discussions with farmers in the 2024 about what should we do. Do we need to should we go out there and put some more in on this field?

And then I had some farmers that just said, I got good crop insurance. I'm not gonna throw another dollar at this stuff. If I if I have to rely on crop insurance to cover my butt, it's okay.

Right. Definitely as a wrinkle. I mean, if if you're gonna rescue it and still bring it only within the the threshold where you're still collecting but not as much, what was the difference? You know?

And that that clearly farmers need to make that that determination. You know, I I found it interesting, Jeff, looking at the results of that trial. You had 25 sites where you only had one out of 25 where you had an yield increase, which based on statistics, that could have just been some other random factor, you know, when it's that slight. But but on the other hand, eight out of 25 had a decrease, and, you know, I don't think a farmer wants to be engaging in a technology that's gonna lose them yield one third of the time. I mean, we hear that consistently.

Yeah. You're right, and and it gets to the risk factor. It it it makes it a more risky system, and those eight to 25, some of those sites were ones where we delayed a significant amount of the n till beyond when we would recommend an application. The other factor that I would I would suggest is is that when we look at split application on medium and fine textured soils in Southern Minnesota in particular, which I've done a fair amount of, more recent data that some data that we've got collected in Southeast Minnesota says that about 20 to 25% of the time, we can expect the split application to have a yield advantage. Another factor of that is is about the same probability or maybe a little greater we've seen where split applications sometimes can get by with a little less n than when n is all apply is applied all at planting or preplant.

So there is the possibility for two potential advantages to that, but in this data that I'm talking about from Southeast Minnesota, these are split applications where we're putting on generally the majority of the n upfront at least two thirds, and then we're side dressing a smaller amount, and we're targeting the optimum time for side dress, which is generally on brain fed conditions around v four to v six, and maybe as late as v seven. But v five, v six is when most of those applications went on.

Right. And I've already mentioned the fact that we recommend multiple splits in irrigation, and the prime primary reason for that is those sites are prone to lose nitrogen. So you you just simply are are you're mitigating that by just not having the nitrogen out there until later. You know? So similar situation where there's other fields, non irrigated, little heavier textured soils, you you need to just simply identify what are the sites that are prone to lose nitrogen or have nitrogen availability issues, and those are the sites that are likely to respond in that way.

And so, you know, we've stressed that through our whole nitrogen smart program that this is a science. These factors are knowable. They may not be easily knowable. However, a lot of farmers, if they really get to know their fields, they can start picking that stuff out. So let's talk just very briefly about using drones because I think, you know, when when drones first came out and were being used for crop scouting, they were being used in almost an identical way that the crop sensors were.

And like I I've said about drones is they're basically your your your your color sensor, your light sensor minus a week because somebody has to go out to the field. They need to fly the drone, get it up there, get the pictures taken, bring it back down. You download it, then you need to because there's always gonna be some tilting, slanting. Somebody needs to justify the field edges to get it to conform with, you know, with the actual geometry of the field. Then you need to run some kind of a program to interpret the imagery that was captured by the drone.

You need to turn that into a application map. Then you need to send it off to whoever you're buying your fertilizer from, get in line, you know, and and then they're gonna come out and do it, and that could that could be a week. You know? And so, you know, there's just there's a lot of just disadvantages why using drones in this similar way. It it just never really took off.

I think I think a lot of people thought years years ago, and I say years ago, I mean, we're talking five years ago, you know, that there was gonna be great potential for this, but I think they've kinda realized that there's just some limitations to that. You know, maybe the one area, you know, and I know, like like, Dan Kaiser's done some work on this, and some of our folks on campus, you, Shannon, and so forth have also investigated wavelengths that aren't necessarily visible to the naked eye near infrared and stuff that maybe the crop can tell us something kinda before it gets in trouble, but there's still some sensitivity because that's only about three days, you know, before it starts losing yield. And so you still if you're gonna use that I mean, if a farmer's in a situation where they can do this all themselves, they probably can get away with it. But when it's done commercially, boy, that lag time can just be an issue.

Yeah. I think all the whether it's drones or just satellite imagery, it all has value. But but managing a system of it, again, it is gonna be challenging and maybe a little risky, and then you always have the weather to deal with too. Are you gonna get good days to take imagery? You gotta have pretty clear cloud free days to get good imagery, whether it's with a drone or whether with with satellite, And then like you said, you gotta take those photos.

You gotta stitch them together. You gotta determine what your prescription is. Now they do, and they have been working tremendously, I think, on AI software that does this and does it really well, and that gets better and better every day. And that may change how we see this stuff in five years or so.

Yeah. Exactly. And we've been talking about the potential of AI, like with crop models too, and we're not gonna talk crop models today. But AI definitely is going to change a lot of this stuff probably after you and I's careers, but but not too far out into the future. And so they'll just summarize here, you know, quickly that that for the most part, using sensors in in northern climates just hasn't really paid off for corn.

Really not part of our recommendation system. I know we get listeners from other areas, you know, kinda check with your local your land grant university and see what they're saying, you know, particularly if you're in Southern states. You know, maybe in the future, like we just got done saying, passive sensors using drones may actually have a little more utility.

One thing I would just add before we move on, Brad, is that we still have, you know, several scientists in our university and in the Upper Midwest that are researching this. So this is, you know, still kind of an active, very, very engaged science. It's not I wouldn't say that it's done and absolutely Right. You know, not not possible.

It's still it's still being developed and still being interpreted at this you know, as we speak.

And, you know, and and I guess to go back to the the time lag issue, you know, now we're seeing these huge drones that look like an aircraft, and they're they're we're doing pesticide applications, but they're also capable of dropping fertilizer too. And so it is possible that maybe drones are going to be able to do some sensing and drop fertilizer at the same time. That technology, don't think is there yet, but it it we may get there.

Yeah. That makes the most sense to me. Put the sensor on the drone and the and carry the fertilizer with it and do it all at one time.

Thank you to Brad Carlson, Extension Educator, and Jeff Vetch, Lead Researcher at the Southern Research and Outreach Center here in Waseca, Minnesota.

Do you have a question about something on your farm? Just send us an email here at nutmgmt@umn.edu. Thanks a lot for listening, and we look forward to seeing you next time.

Advancing Nitrogen Smart is proud to be supported by the farm families of Minnesota and their corn check off investment through Minnesota Corn.