OFR21: Sensor Based N Management with Samantha Teten
In this video, Graduate Student Samantha Teten discusses how farmers can use active crop canopy sensors to determine how much nitrogen their corn needs in real time. Sam shares the results of two years of research using this technology in non-irrigated systems.
icon search Searchable Transcript
Toggle between list and paragraph view.
[00:00:10.630]So, Jackson and I are actually, like, on the same team.
[00:00:13.300]We both, our advisor is both Dr. Luck.
[00:00:15.430]So, yeah, if you,
[00:00:17.160]we use the same type, similar technology.
[00:00:19.350]So if you're on the irrigated side,
[00:00:21.000]irrigated part in Nebraska,
[00:00:22.180]I encourage you to go to his.
[00:00:23.390]I'm focusing on, like Laura said, the non-irrigated part.
[00:00:34.840]So with that, I'll go ahead and get started.
[00:00:37.520]I believe there's a few collaborators
[00:00:39.940]that are joining us today that assisted in this study.
[00:00:42.940]So I want to thank them.
[00:00:44.950]Also I'm, like I said,
[00:00:46.310]focusing on the non-irrigated ones.
[00:00:48.010]There's also some irrigated sites.
[00:00:50.390]So even though I won't discuss them,
[00:00:52.050]I'd encourage you to go check those out.
[00:00:54.040]I'm starting on page 91 if you want to follow along
[00:00:59.920]to learn a little bit more about these
[00:01:01.350]active crop canopy sensors
[00:01:03.010]that we're using for nitrogen management.
[00:01:06.740]So, these active sensors
[00:01:08.740]are a part of this project called Project SENSE.
[00:01:11.540]There's a large team of us,
[00:01:12.950]I guess I should have introduced some of them
[00:01:15.950]here on this front page,
[00:01:17.550]but there's even more that are not listed there.
[00:01:20.250]So, this project has been going on for six years
[00:01:22.630]and it's a reactive approach.
[00:01:24.550]So maybe in your earlier discussions,
[00:01:26.620]you may have talked about some of the studies that were
[00:01:30.040]So, those are more predictive.
[00:01:32.670]The active crops canopy sensors are looking at
[00:01:35.020]what's actually going on in the season
[00:01:37.060]and they're making a recommendation based upon that.
[00:01:41.340]So, here's just an example to try to show
[00:01:45.220]how they work in the field.
[00:01:48.480]So these are the active crop canopy sensors.
[00:01:50.420]There's four of them mounted on the front.
[00:01:56.110]So, as they move through the field,
[00:01:59.780]you can see the NDRE are the crop indices,
[00:02:02.360]so what the sensors are reading,
[00:02:06.660]And as a response, the nitrogen target rate
[00:02:10.390]is also changing.
[00:02:12.280]And how the model is set up with these,
[00:02:14.710]the lower the crop indices,
[00:02:18.530]or the less healthy the crop looks,
[00:02:20.940]the more nitrogen is being applied,
[00:02:23.140]versus in areas where the crop looks healthier
[00:02:26.040]and the crop indices is higher,
[00:02:27.870]we're applying less nitrogen in the field.
[00:02:29.860]So there's some different philosophies there,
[00:02:32.170]but that is how the model that we are using works.
[00:02:36.210]They also need a couple of other parameters
[00:02:38.710]when you're setting 'em up in your field.
[00:02:40.620]The first is what do you think
[00:02:42.140]is the optimal nitrogen rate in the field?
[00:02:44.610]So if you're running on an irrigated site,
[00:02:46.810]your optimal nitrogen rate would be closer, would be higher,
[00:02:50.470]versus if you're in a non-irrigated site
[00:02:52.260]where you maybe don't have as much potential,
[00:02:53.890]your optimum nitrogen rate would be lower.
[00:02:55.760]So it's kind of just giving it a target range
[00:02:58.420]at which the sensor should operate in.
[00:03:01.400]We also need to know
[00:03:02.233]how much nitrogen was already applied in the field.
[00:03:04.910]If you put something on with your planter,
[00:03:06.960]if you put something on before planting,
[00:03:09.610]we need to know that.
[00:03:11.510]And then also nitrogen credits.
[00:03:13.130]So if there's anything that's going to be applied
[00:03:15.548]after this application.
[00:03:16.381]So things from irrigation credits
[00:03:19.780]or any other credits that you want to account for.
[00:03:23.530]And then we can also set a range
[00:03:25.530]of which the sensors can operate in,
[00:03:27.700]and that's a way to minimize your risk.
[00:03:29.580]So, if you know
[00:03:30.970]that you did not want the system to apply under,
[00:03:33.610]you know, 30 pounds or under 50 pounds,
[00:03:36.040]we can set that range and minimize, like I said, the risk.
[00:03:40.060]So we're running these sensors between stage V8 to V14,
[00:03:45.240]and the base nitrogen rate
[00:03:47.480]is recommended between 35 pounds and 75 pounds.
[00:03:50.980]So we want that applied
[00:03:52.530]at least two weeks prior to the sensors going out there.
[00:03:56.640]A couple of notes.
[00:03:58.120]So, we've been running these sensors on irrigated sites
[00:04:00.500]for six years.
[00:04:02.010]The dry land, or non-irrigated sites,
[00:04:03.760]has been going on for two years.
[00:04:05.280]And a couple of things we have found
[00:04:07.460]is that we actually need to apply earlier
[00:04:09.620]on non-irrigated sites in order to get
[00:04:11.640]an incorporating rainfall.
[00:04:13.410]And we also like to have a higher base rate
[00:04:16.400]that tends to do a little bit better
[00:04:17.770]in the heavier soils of Eastern Nebraska.
[00:04:22.670]So, just quickly on how these sensors work.
[00:04:25.810]They are collecting near infrared and red edge
[00:04:29.300]to calculate NDRE,
[00:04:31.140]but NDRE can change over time
[00:04:33.120]and it can also be different between field to field.
[00:04:35.930]So we try to calibrate that NDRE value
[00:04:38.250]by comparing it to a high nitrogen reference strip.
[00:04:41.480]So that as a high amount of nitrogen,
[00:04:43.650]that's showing what we are positive
[00:04:45.610]is a nitrogen sufficiency.
[00:04:48.553]that is what we are comparing the rest of the field to.
[00:04:51.090]So that's like what's shown here,
[00:04:53.050]we're comparing the part that we're applying to that value
[00:04:56.160]to, like I said, calibrate our sensors.
[00:05:02.500]Our treatment is the sensor-based treatment.
[00:05:04.400]We're comparing that to the grower's current management,
[00:05:07.760]and for most of these sites that was anhydrous,
[00:05:10.060]completely upfront for their full rate of nitrogen.
[00:05:13.070]But we did have one site,
[00:05:14.160]the first site I'll talk about,
[00:05:15.300]is a split application, as well.
[00:05:18.260]So we let the grower decide all of that,
[00:05:20.690]as well as the width of the treatments and the study area.
[00:05:25.590]So, as I said, this has been going on for six years.
[00:05:27.910]And one thing that we have found
[00:05:29.630]is that most of the sites are,
[00:05:32.300]with sensor-based management,
[00:05:34.740]are more profitable and more nitrogen use efficient,
[00:05:38.420]like I said, with the sensors.
[00:05:40.350]So, that's 64% of the sites.
[00:05:42.960]What we're trying to figure out is what is causing
[00:05:45.660]some of these sites to not be that way.
[00:05:47.230]We have some explanations, but we,
[00:05:50.710]on most of these sites, we're applying less nitrogen
[00:05:53.920]and still maintaining yield or even increasing yield
[00:05:57.170]because we're managing for that spatial variability.
[00:06:02.940]So on the site that I'm going to talk about,
[00:06:04.340]I'm going to cover four of them,
[00:06:06.870]the non-irrigated sites.
[00:06:08.770]And here's kind of an example
[00:06:09.910]of how those treatments are set up.
[00:06:15.360]So first, I'm on page 98 talking about
[00:06:19.190]the site that was up in Dodge County.
[00:06:22.020]This field, showing the rainfall amounts,
[00:06:26.110]so it did, it was really short on rain this year.
[00:06:29.780]We did get a rainfall right after the application,
[00:06:32.210]which is something we're really looking for
[00:06:33.890]to make sure that the nitrogen gets incorporated.
[00:06:35.810]So that part was good.
[00:06:37.510]But then later when the rain shut off,
[00:06:39.850]we did have some hurt in yield.
[00:06:43.150]So at this site,
[00:06:45.200]the grower applied some nitrogen with their planter,
[00:06:48.570]but then applied the rest of it at V6 with the coulter
[00:06:51.370]and for a total of 124 pounds of nitrogen
[00:06:55.040]versus the active crop canopy sensors,
[00:06:57.510]same 35 pound base rate,
[00:07:00.260]but then the sensors came back at V11.
[00:07:03.700]And we applied the nitrogen with a stabilizer
[00:07:07.140]and applied, actually, about the same average amount.
[00:07:11.130]So of course in some places, it's going to be lower,
[00:07:12.920]in some places it's going to be higher,
[00:07:14.510]but an average total amount of 123 pounds of nitrogen.
[00:07:20.330]So with those very slight differences
[00:07:22.600]in total nitrogen applied,
[00:07:24.060]the yield was non-statistically different, it was the same,
[00:07:27.130]and therefore marginal net return was also the same.
[00:07:29.680]But one thing I would like to point out
[00:07:31.800]with this collaborator, also worked with me last year,
[00:07:36.203]a site that was not very far away actually yielded
[00:07:39.160]closer to 250 to 260 bushels per acre.
[00:07:43.360]So as I mentioned before,
[00:07:45.230]rainfall was a limiting contributor at the site this year.
[00:07:50.330]Moving on to page 99,
[00:07:53.760]this is at Cass County.
[00:07:55.370]This collaborator did two sites with me,
[00:07:57.530]so I'm gonna compare those a little bit to each other.
[00:08:00.450]The site had a little bit more rainfall
[00:08:02.500]and did have a rain right after the application.
[00:08:06.840]So this was 161 pounds applied for anhydrous
[00:08:10.720]of the grower's current application
[00:08:12.852]and 30 pounds through map.
[00:08:16.100]So, the sensor application had
[00:08:19.030]a base rate of 70 pounds total,
[00:08:21.950]but that was split between anhydrous and map.
[00:08:27.580]and then the sensor application came in
[00:08:29.330]and only applied 125 pounds of nitrogen.
[00:08:33.260]One thing I would like to point out.
[00:08:34.600]So this site did not,
[00:08:36.670]the sensors did not perform as we had hoped.
[00:08:39.640]And so, I'll show that here in the results.
[00:08:41.300]And one of the theories is that it's because,
[00:08:44.870]you know, anhydrous applicators probably shouldn't be run
[00:08:47.100]at such a low rate, 40 pounds.
[00:08:48.900]It may be hard for them to keep that rate.
[00:08:51.590]The split based nitrogen maybe didn't perform as well,
[00:08:55.790]but, you know, looking at those results,
[00:08:58.880]so, a lot less of the nitrogen was applied
[00:09:01.750]and it was a 20 bushel loss in yield.
[00:09:05.050]So that's very unacceptable to adopt a technology
[00:09:10.120]if it's going to lose you that much
[00:09:11.610]in yield and profitability.
[00:09:13.520]So we're really looking into
[00:09:14.670]what could have happened at this site.
[00:09:16.410]So, as I mentioned,
[00:09:17.243]the base rate could have been a possibility.
[00:09:20.810]This site did not have cover crops
[00:09:22.410]versus the site that is nearby did have cover crops.
[00:09:26.120]And the other difference is this site
[00:09:28.110]had a little higher yield potential.
[00:09:30.340]And so, more nitrogen should have probably been applied
[00:09:33.620]by the sensors.
[00:09:35.820]But in comparison, the site, like I said,
[00:09:37.490]the next site was not very far away.
[00:09:40.030]So, same amount of rainfall, same management otherwise,
[00:09:44.360]other than the cover crops,
[00:09:46.600]did get rainfall afterwards.
[00:09:51.260]So, 175 pounds of anhydrous all up front
[00:09:55.110]versus 70 pound base rate of anhydrous
[00:09:58.620]and then our sensor application.
[00:10:01.240]But the sensors once again only applied
[00:10:03.750]121 pounds of nitrogen.
[00:10:05.500]So, very similar to the last site,
[00:10:08.730]however, it actually increased yield
[00:10:11.220]compared to the growers current management.
[00:10:13.440]So it must have been,
[00:10:16.680]you know, right at the right optimum nitrogen rate,
[00:10:18.960]but as well as accounting for
[00:10:20.760]that spatial variability in the field,
[00:10:22.920]which is then what was able to increase the yield.
[00:10:27.530]So the final site is at Saunders County.
[00:10:30.930]We were able to do a little bit more.
[00:10:32.760]So, there was also some rain limiting factors here,
[00:10:35.684]but we were able to do a little bit more testing here.
[00:10:37.690]So, some of those questions that we have
[00:10:38.930]is the timing of the application
[00:10:41.180]and the base rate of nitrogen that should be applied.
[00:10:45.260]As I've said,
[00:10:46.093]we think that that could be different for irrigated sites
[00:10:48.140]versus non-irrigated sites,
[00:10:49.890]or maybe what type of soil you have
[00:10:51.900]to make this technology work better.
[00:10:55.200]So we were testing that.
[00:10:56.560]Two different application timings on each half of the field
[00:11:00.180]and two different base rates for each of those timings.
[00:11:02.810]So a 35 pound base rate versus a 70 pound base rate,
[00:11:08.170]versus the grower, which was all anhydrous up front.
[00:11:13.720]So I'll first point out here that the grower treatment
[00:11:17.440]was the same for each half of the field.
[00:11:20.260]Yielded the exact same 220 pounds of bushels per acre
[00:11:24.380]for both half of the fields for the grower.
[00:11:27.356]So we can kind of compare all these treatments to each other
[00:11:30.520]with that in mind,
[00:11:32.290]but looking at the earlier application
[00:11:35.980]versus the later application,
[00:11:37.860]the more nitrogen was applied earlier by the sensors,
[00:11:42.120]which is a little bit confusing,
[00:11:44.170]because you would think that later on the season
[00:11:45.930]you would have more stress, and therefore the active sensors
[00:11:49.450]would recommend more nitrogen,
[00:11:51.060]but that wasn't the case, as shown here.
[00:11:54.230]So this later application applied less nitrogen
[00:11:58.310]and also had a lower yield.
[00:12:00.330]So we'd encourage a sooner, an earlier application
[00:12:04.040]and a higher base rate based upon the yield results,
[00:12:07.720]even though they're not statistically different.
[00:12:09.850]So this definitely needs to be tested more.
[00:12:14.500]So, briefly, just to summarize these,
[00:12:16.780]all of the sites,
[00:12:18.030]the active crop canopy sensors applied
[00:12:20.300]less than the grower's current management,
[00:12:24.170]however, on most of the sites, except for two of them,
[00:12:27.730]it also resulted in a little bit of loss in yield.
[00:12:31.110]Two of those, it was statistically different,
[00:12:34.270]very significant on that one site.
[00:12:37.890]And then that resulted in some loss in profitability,
[00:12:41.170]except for it can be really beneficial
[00:12:43.520]on some sites, as well.
[00:12:46.810]I'm going to skip through that
[00:12:47.750]so we have some time for discussion,
[00:12:49.310]but one thing I'd like to point out is these,
[00:12:51.150]we recognize that these active crop canopy sensors,
[00:12:53.790]we've been researching them for many years
[00:12:56.260]and they have their flaws
[00:12:57.410]and we're still learning more about them.
[00:12:59.210]So moving forward from here, though,
[00:13:01.990]we're looking at some other ways
[00:13:04.410]to use sensor-based technology, right?
[00:13:06.370]So there could be the use of drones and UABs.
[00:13:10.280]They are actually being tested on a lot of these sites
[00:13:14.890]and in some other studies that you can look through
[00:13:16.910]here in the book.
[00:13:18.030]And that's where you take an image
[00:13:20.350]and then you are turning that into a prescription map
[00:13:22.880]to apply within the next couple of days.
[00:13:25.550]That gives you more flexibility,
[00:13:27.820]although it may take more time.
[00:13:30.020]It can cause a little bit more work
[00:13:32.760]in that transition process
[00:13:34.120]versus the active crop canopy sensors,
[00:13:35.870]you go into the field,
[00:13:36.800]and it's applying that right, right away.
[00:13:39.420]I'll also put a plug in
[00:13:41.000]for the precision nitrogen management project
[00:13:43.980]that's coming in the next coming a few years.
[00:13:47.280]Laura will talk about that a little bit later,
[00:13:49.080]but that's looking at sensors.
[00:13:50.670]You can still continue to test these
[00:13:52.680]as we continue to learn more.
[00:13:54.370]You can look at nitrogen models or other nitrogen products.
[00:13:58.410]So, last year we didn't use inhibitors with our nitrogen.
[00:14:01.730]Our stabilizers, this year,
[00:14:03.010]we did and thought that that did help.
[00:14:05.160]So, you can test some of those different products.
[00:14:10.060]So with that, that's everything I have.
[00:14:12.360]If you guys have any questions.
[00:14:14.830]Hey Sam, what effect do you think,
[00:14:19.590]just in terms of nitrogen capture versus nitrogen loss
[00:14:23.640]between the two systems,
[00:14:24.920]like the grower base or kind of the traditional
[00:14:27.260]versus the sensor-based, I guess.
[00:14:29.270]Any thought on that?
[00:14:32.910]So that's something we've worked to capture.
[00:14:34.890]We take a lot of soil samples
[00:14:36.260]to try to capture, like I said, that loss.
[00:14:41.060]That's still kind of being analyzed and looking more into.
[00:14:44.040]One thing that we have found is that
[00:14:46.830]on these irrigated or sandier sites,
[00:14:49.730]the crop canopy sensors
[00:14:52.010]need to kind of operate a different way
[00:14:53.620]to manage, like you said, for that loss
[00:14:55.180]versus in heavier soils
[00:14:56.990]where we don't see as much nitrogen loss.
[00:14:59.170]We need to manage a little bit differently, like I said,
[00:15:01.760]with the more upfront potentially,
[00:15:04.440]and they need a little bit more rainfall to incorporate.
[00:15:06.860]So we're trying to capture that with some soil samples,
[00:15:09.280]but you can only take so many in a season
[00:15:12.220]to really see that.
[00:15:14.400]But we're improving the nitrogen use efficiency.
[00:15:21.110]I do see a question here.
[00:15:21.943]This is an excellent question in the chat.
[00:15:24.710]Did the less crop canopy cover affect how the sensors work?
[00:15:32.820]especially on the UAVs.
[00:15:34.930]So like I said, we're researching that.
[00:15:37.840]Soil pixels, when you pick those up,
[00:15:39.900]really influence the prescription.
[00:15:42.230]It's we try to minimize that a lot
[00:15:44.050]with the active crop canopy sensors,
[00:15:45.820]because those are positioned directly over the row.
[00:15:48.270]So as long as you're going pretty straight
[00:15:49.910]and you don't get off a little bit,
[00:15:51.930]because if you get off a little bit,
[00:15:52.790]you will get more soil,
[00:15:54.100]and that can definitely influence the rate.
[00:15:57.030]However, that's why we do recommend, you know,
[00:15:59.180]between V8 and V12.
[00:16:01.360]If you're within those ranges, it should be working well
[00:16:04.810]with the active crop canopy sensors.
[00:16:11.122]Those are both excellent questions.
[00:16:16.300]I'm going to go ahead and close the breakout room,
[00:16:18.250]but if anyone has any final questions for Sam,
[00:16:20.600]feel free to ask those now.
[00:16:23.720]We've got about one minute.
[00:16:30.903]I don't know if you want to also share a little, Sam,
[00:16:32.720]about how the timing with less canopy
[00:16:35.920]early in the season might impact?
[00:16:40.030]Like for, say, a V4 versus a V8 growth stage?
[00:16:45.610]Yeah, so NDRE really increases over time.
[00:16:50.420]And so, with the earlier crop stage,
[00:16:53.950]so there's a couple of things there.
[00:16:55.700]You know, your NDRE is not going to be as high,
[00:16:58.210]and so you would probably be applying too much.
[00:17:01.350]There's a lot of season left.
[00:17:03.590]But also you have to have that high nitrogen sufficiency,
[00:17:07.700]high nitrogen reference strip established,
[00:17:09.750]and it's that difference between them, right?
[00:17:11.240]So sometimes we've found that, actually,
[00:17:12.640]the more stressed fields do apply more
[00:17:16.840]and so they actually perform better.
[00:17:18.810]So there's, yeah, that timing component
[00:17:21.480]and capturing that NDRE at the right time
[00:17:23.740]is really important
[00:17:25.050]and something we're still trying to learn more about.
Log in to post comments