Applying Precision Ag Technologies to enable Variable Rate Soybean Population Studies in Nebraska

University of Nebraska – Lincoln Author

08/31/2020 Added

27 Plays

Description

Dr. Joe Luck discussed using planter prescriptions and yield monitors to evaluate the optimal soybean seeding rate.
Feedback form

Searchable Transcript

Search:

[00:00:12.240]Good morning.
[00:00:13.073]This is Joe Luck with Nebraska Extension.
[00:00:15.230]And today I'm gonna share with you just a little bit
[00:00:16.990]about how we can apply precision ag technology
[00:00:19.630]to conduct an on-farm research.
[00:00:21.280]In particular, we're gonna be looking at
[00:00:23.540]seeding rate studies and in the case study,
[00:00:25.820]we'll look at today, we're gonna be talking
[00:00:27.410]about soybean seeding rates.
[00:00:29.430]I would also like to recognize
[00:00:30.890]Sarah Sivits and Laura Thompson,
[00:00:33.150]two of our extension educators that have helped out
[00:00:35.590]with this project in particular.
[00:00:38.850]We have several opportunities to do seeding rate studies.
[00:00:42.410]This is an example of a typical field link strips,
[00:00:45.850]where we could use maybe a manual approach
[00:00:48.140]to set up and conduct that study even if we had to go out
[00:00:51.120]and manually change rates on a planter.
[00:00:54.210]It's been very successful in a way to go out
[00:00:56.720]and just help producers look at
[00:00:58.970]how their typical seeding rates are performing compared
[00:01:02.150]to higher and lower rates.
[00:01:04.250]What we're gonna really be talking about today
[00:01:05.950]is how we can use precision ag, data sets,
[00:01:09.430]and also the technology in the field to help us maybe create
[00:01:13.170]a little bit more complex studies.
[00:01:14.760]So the example you see here is a prescription map we created
[00:01:18.410]for a on-farm research study.
[00:01:20.660]You can see the treatment blocks across the field,
[00:01:24.490]so we have four different seeding rates.
[00:01:25.970]We actually have 16 blocks throughout this field,
[00:01:29.660]each block containing one of those four seeding rates.
[00:01:32.290]So this was actually developed in a software package
[00:01:36.180]and the prescription map was created.
[00:01:38.340]We could then take that prescription map and upload that to
[00:01:41.190]the variable-rate monitor inside the cab or the tractor,
[00:01:44.530]if that planter had that capacity.
[00:01:47.040]And it's important to note that there are commercially
[00:01:49.350]available farm management software packages that will now
[00:01:51.860]allow us to design these kinds of studies.
[00:01:54.530]And that makes it really handy because then we can basically
[00:01:57.110]translate that directly to the monitor in the cab
[00:02:01.260]to upload this prescription.
[00:02:04.130]I also wanna point out that each one of those systems,
[00:02:06.810]whether we're doing variable-rate seeding,
[00:02:09.470]variable-rate dry or liquid product application,
[00:02:12.550]those systems typically record the as-applied data.
[00:02:16.640]So where and at what rate the products
[00:02:19.050]were applied every second.
[00:02:20.380]So that's what we're seeing here.
[00:02:22.240]Every second, that planter is trying to estimate
[00:02:24.950]how many seeds per acre were planted
[00:02:26.860]as we moved through our plots.
[00:02:28.840]And this is just a really nice quality control check.
[00:02:31.720]We'll look at this a little bit here in a second.
[00:02:34.950]If we know that the systems didn't respond appropriately,
[00:02:37.740]it's probably a good opportunity to take that data out.
[00:02:40.760]So you can actually see in this case,
[00:02:44.230]the southeast portion of the field,
[00:02:45.670]which was planted first and move into the northwest,
[00:02:49.360]the planter actually, wasn't responding
[00:02:51.210]to our different blocks.
[00:02:52.320]And that was because the prescription wasn't loaded properly
[00:02:55.490]into the monitor.
[00:02:56.850]The operator was able to stop and make that change,
[00:02:59.210]so then you can start to see where the planter
[00:03:01.640]actually started to respond appropriately.
[00:03:04.480]And this is just good to know because someone might think
[00:03:07.830]that this study wouldn't be useful, or in fact,
[00:03:11.710]we didn't expect the operator to plant this
[00:03:14.020]at an angle across the field,
[00:03:15.420]we were just kind of expecting an east-west planter pass.
[00:03:20.870]So one might think that's gonna be useless,
[00:03:22.840]but actually with the as-applied data,
[00:03:24.500]we can go back and look at areas where the planter
[00:03:27.010]successfully planted according to our rate blocks,
[00:03:30.240]and then look at the yield monitor data and verify that.
[00:03:33.820]So this study was actually not lost at all.
[00:03:37.610]You're still able to get some good information.
[00:03:40.120]And you can see that here in the summary table,
[00:03:42.100]in particular for this, we'll look here at the yield data,
[00:03:46.450]we had yields ranging from 62 average bushels per acre
[00:03:51.550]to 65.
[00:03:53.090]It's important to note that we designed this for 16 blocks.
[00:03:57.610]If we looked at complete blocks, so all four target rates
[00:04:01.560]were represented successfully
[00:04:03.130]from the planter and the harvest data,
[00:04:05.680]we ended up with only eight for this summary.
[00:04:09.570]But even with eight, we were able to detect
[00:04:11.850]significant yield differences down
[00:04:13.420]to three bushels per acre, or just slightly over three.
[00:04:16.850]You can see here the 62 to 65,
[00:04:19.350]those were statistically different.
[00:04:22.300]Another thing that also might affect the ability
[00:04:25.070]to detect that high of a resolution is the field itself.
[00:04:28.390]So if we have a very uniform field, not a lot of soil
[00:04:33.520]textural differences or terrain, for instance,
[00:04:36.870]that actually helps us in this type of study
[00:04:39.340]know that the field is probably not gonna benefit
[00:04:42.020]from a variable-rate approach across the field.
[00:04:45.260]So if this yield difference, if say that had been
[00:04:49.530]much higher, say it it was 10 bushels per acre
[00:04:52.830]was our least significant difference,
[00:04:54.380]that actually might be a clue that
[00:04:55.930]there is a lot of variability in that field.
[00:04:58.680]And we might need to go back and analyze the data
[00:05:01.200]and determine if maybe a true variable-rate approach
[00:05:04.440]would be of benefit,
[00:05:05.530]so maybe raising rates in certain parts of the field
[00:05:08.470]and lowering rates in another part of the field.
[00:05:10.440]So there's kind of two different clues we can look at there.
[00:05:14.350]As we always do on-farm research network,
[00:05:16.940]we really need to look at profitability here in some way.
[00:05:20.260]And what you can see here is the grower typical rate
[00:05:23.410]of 150,000 seeds per acre.
[00:05:26.010]It's statistically less than what would have been
[00:05:28.900]if we'd had to drop the seeding rate down to say 125
[00:05:32.180]and even 100,000 seeds per acre.
[00:05:34.230]So the true optimum rate may be somewhere
[00:05:38.420]in between those two rates,
[00:05:40.250]but at least with this information, we can show the producer
[00:05:43.970]that they could actually get some money back.
[00:05:49.040]We really have to think about quality design of the studies.
[00:05:53.620]So always replicated randomized treatments
[00:05:56.690]organized into blocks.
[00:05:58.460]If we're using maybe a product, we really have to think
[00:06:01.620]about check strips, zero rates of that product,
[00:06:05.100]just to say what would have happened if we didn't use it.
[00:06:08.260]Typically, in these types of seeding rate studies
[00:06:10.910]or so forth, that might be the grower.
[00:06:14.890]In this case, again, 150,000, you can see that the rest
[00:06:18.290]of the field here was planted at that rate.
[00:06:21.950]We need to think about two other things.
[00:06:23.870]One is with technology, we need to span a pretty wide range
[00:06:28.180]in terms of our rates.
[00:06:30.010]So if I were to say let's plant a corn seeding rate study
[00:06:34.280]at 34,000, the next treatment might be 33,000.
[00:06:38.960]I'm not really confident that the planting technology today
[00:06:41.930]can hold us to that, that fine of a resolution.
[00:06:44.680]So typically in our corns rate studies, we might suggest
[00:06:49.340]3,000 or 4,000 seeds per acre between treatments.
[00:06:53.300]On our soybean seeding rate studies,
[00:06:54.850]we might think about something closer
[00:06:57.530]to 30,000 seeds per acre between our treatments.
[00:07:01.740]For instance, in nitrogen rate studies,
[00:07:03.510]we've typically said not less than 30 pounds per acre
[00:07:06.910]of nitrogen between rates.
[00:07:09.220]So just something to think about there.
[00:07:11.140]Also really need to consider the applicator itself,
[00:07:14.810]what's the width of that equipment
[00:07:16.270]versus the harvester width,
[00:07:17.890]because we're using yield monitor data
[00:07:20.230]to assess these studies, and oftentimes,
[00:07:24.070]especially on soybeans, those two do not match up very well.
[00:07:27.410]A lot of times in corn, it's much easier because a harvester
[00:07:30.610]might be half the planter width,
[00:07:32.010]but in soybeans that could be different.
[00:07:33.810]So something to think about as we set up these blocks.
[00:07:40.530]Another aspect that we've started to take advantage
[00:07:42.990]of lately is our in-season scouting
[00:07:45.560]and trying to digitize that data.
[00:07:47.450]So what you see here is a picture of our prescription map
[00:07:51.110]on the left.
[00:07:52.130]We can actually use software to go in
[00:07:53.960]and identify exact blocks where we wanna go out
[00:07:57.200]and do some in-season scouting.
[00:07:59.380]In this case, we've been using app called SMS Mobile,
[00:08:04.470]which SMS is also a software package
[00:08:06.730]you can use to design these studies.
[00:08:09.280]This allows us to prescribe where we're gonna go out
[00:08:12.440]and then merge that data with the geo reference location.
[00:08:17.470]So the neat thing about SMS Mobile is it will actually
[00:08:21.020]direct you sequentially through your plots.
[00:08:24.260]So it'll tell you a heading for you to take
[00:08:26.550]and then the distance to take and walk you through
[00:08:29.180]where you're supposed to be in,
[00:08:30.030]and then you can enter in that data.
[00:08:32.140]In this case, we're typically doing stand counts
[00:08:35.320]early in the season, and then it'll actually tie that data
[00:08:38.630]back to the location.
[00:08:40.740]We could also do this with a handheld GPS device.
[00:08:44.400]So our cooperating educator, Sarah Sivits, in this year
[00:08:48.560]has gone out and taken stand counts across our blocks
[00:08:52.870]and recorded those data here along with the intended rate.
[00:08:56.800]So we can actually go back now and look at
[00:08:58.540]average seeding rates and just to be confident and ensure
[00:09:01.530]that as we went up in our target rates,
[00:09:04.110]we also went up in our emerged rates of plants.
[00:09:09.090]And that gives us another clue if there's some issues
[00:09:11.780]we need to go back and look at maybe in the data.
[00:09:16.980]We've already talked about some data sets to collect.
[00:09:20.100]We really are starting to get some interesting information
[00:09:23.500]from the georeferenced in-season crop scouting data.
[00:09:27.290]Also aerial imagery, if you have the ability to acquire that
[00:09:30.590]at some point during the season.
[00:09:32.500]If we do have an issue that we could detect
[00:09:35.690]in our crop scouting using that georeferenced method,
[00:09:38.670]sometimes we can go back and tie that into different areas
[00:09:41.930]of the field using imagery that we might not be able to do
[00:09:45.730]or have the time to do while we're actually out walking
[00:09:48.820]through the field.
[00:09:51.340]In the end like again, we're using yield monitor data
[00:09:54.230]with these studies to conduct the analysis.
[00:09:58.870]So we are trying to get as many blocks or as many reps out
[00:10:03.330]in that field as we possibly can.
[00:10:05.660]A couple of interesting things.
[00:10:07.010]So if you look at the chart here in the bottom right,
[00:10:11.150]the areas in red might be locations where we didn't really
[00:10:16.480]get the same analysis in terms of marginal net return
[00:10:20.660]as the average.
[00:10:21.493]So again, our highest marginal net return
[00:10:24.410]was at the lowest seeding rate.
[00:10:26.550]Notice that in block two and in block six,
[00:10:29.890]that might not have been the case.
[00:10:31.280]So if we would have just gone out and put one strip
[00:10:35.080]of this out, even the four rates,
[00:10:37.600]we might not have gotten quite the same answer.
[00:10:40.680]One thing that I think is really interesting is
[00:10:43.570]if you look again at block two and the small blue circle
[00:10:47.360]and block three, I'm sorry, block four,
[00:10:50.540]if you compare to a higher rate
[00:10:52.860]to the actual growers intended in both cases,
[00:10:57.100]the higher seeding rate would have generated
[00:10:59.300]a higher marginal net return.
[00:11:01.620]And that's one of the things that really,
[00:11:03.570]it's a little bit disturbing when we have consultants out
[00:11:08.440]in the field that just wanna drop one little block
[00:11:11.500]of a different rate, or maybe the growers rate
[00:11:14.080]compared to their block.
[00:11:15.790]The inability to do these replicated studies really
[00:11:19.090]takes out the opportunity to get to a true answer
[00:11:22.270]with this averaging.
[00:11:23.410]So just really interesting that two out of four,
[00:11:28.670]two and four, rather would've given us a different answer
[00:11:31.120]if we were just looking at the growers
[00:11:33.190]and maybe even a higher rate.
[00:11:34.910]So something to think about as we look at and analyze data.
[00:11:40.487]I wanna thank you for tuning in a little bit.
[00:11:42.000]This was just a bit of information about
[00:11:43.790]some of the variable-rate studies where we're trying
[00:11:46.240]to use precision ag tech to speed things up.
[00:11:50.185]We look forward to some questions, and if you ever
[00:11:53.500]wanna reach out there's some contact information.
[00:11:56.390]Thanks.

The screen size you are trying to search captions on is too small!

You can always jump over to MediaHub and check it out there.

Comments

0 Comments

Applying Precision Ag Technologies to enable Variable Rate Soybean Population Studies in Nebraska

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments