OFR21: Precision Ag with Joe Luck

Joe Luck Author

10/26/2021 Added

18 Plays

Description

In this video, Joe Luck discusses how vital precision ag is, and how farmers can use it to conduct on-farm research on their operations.

Searchable Transcript

Search:

[00:00:00.812](upbeat music)
[00:00:09.480]I appreciate the opportunity to visit
[00:00:11.870]with everybody this morning.
[00:00:13.280]This has been a effort that really,
[00:00:16.160]we started back, I think almost seven years ago
[00:00:19.020]when Laura kind of started working heavily
[00:00:21.710]with on-farm research network,
[00:00:23.370]but trying to figure out
[00:00:25.070]how we can leverage precision Ag technologies more
[00:00:27.950]for a lot of our studies.
[00:00:29.310]And so I just want to recognize Laura and also Sarah Sivits,
[00:00:32.850]some of the data that you'll see in the presentation
[00:00:35.830]this morning kind of comes from some of the studies
[00:00:38.660]that they were instrumental in help us carry out.
[00:00:41.380]But really the question is, you know,
[00:00:43.230]how do we leverage these new systems
[00:00:46.220]to improve our, our studies that we're doing?
[00:00:49.610]And really the, the thing that I would like to point out
[00:00:52.290]is just a couple of barriers
[00:00:54.220]that I've seen over the past few years
[00:00:56.240]with on-farm research, just visiting with people
[00:00:59.340]that have private, maybe take part in or interpret studies.
[00:01:02.730]And, and one is obviously the time it takes
[00:01:05.430]to put the field trials out.
[00:01:07.140]So if you're doing all this manually and,
[00:01:11.490]and you're having to go out in the field and,
[00:01:13.100]and put flags out and measure
[00:01:14.800]and all these different things
[00:01:16.461]that takes a lot of time.
[00:01:17.860]So I think there's an opportunity there.
[00:01:19.650]The other thing comes down to how we interpret the results
[00:01:23.490]of these, these studies.
[00:01:25.120]So a lot of times, you know, it's confusing.
[00:01:28.110]So you can see at the bottom of this image here
[00:01:32.360]with the circles in red, you know,
[00:01:34.370]why is , why is five bushels an acre not significant?
[00:01:39.160]You know, why is, you know, $20, $30, $40 an acre,
[00:01:43.060]not significant in some of these studies?
[00:01:45.010]And what does that mean?
[00:01:46.080]Well, a lot of times it comes down to two things
[00:01:49.510]and these studies well, one could be, well,
[00:01:51.780]the treatment really had no effect,
[00:01:53.410]but with those wide differences, it could come down
[00:01:56.300]to the number of treatment replicates
[00:01:58.430]we had out in the field.
[00:01:59.980]So if we only had three full field land strips
[00:02:02.750]of one treatment compared to another,
[00:02:05.060]that's not a lot of data points.
[00:02:06.940]The second thing that can,
[00:02:07.773]that can really play a major role here
[00:02:10.050]is the field variability.
[00:02:11.640]So oftentimes we're kind of relegated
[00:02:14.180]to doing these field trials and field link strips.
[00:02:17.160]So if you look at this image here on the right, you know,
[00:02:22.480]we had these field link strips of different seeding rates.
[00:02:25.340]You know, this entire strip gets average into one point.
[00:02:29.510]And along with these other five,
[00:02:33.224]so six total replicates here, you know,
[00:02:34.160]that six data points that we have,
[00:02:37.010]they go into this statistical analysis.
[00:02:39.210]Well, there could be a lot of field variability
[00:02:41.790]within those strips, and we need to be able
[00:02:43.560]to try and tease that out.
[00:02:45.040]I think these are some of the opportunities
[00:02:46.810]that we have with precision Ag tech.
[00:02:50.370]We really can't,
[00:02:52.610]we can't skip the part about successful trial design.
[00:02:56.180]So just a couple of highlights here.
[00:02:59.130]When we, when we start off
[00:03:00.770]with these studies is really identifying the treatment
[00:03:03.650]that we want to go after.
[00:03:04.930]And so one of the, one of the examples I will give
[00:03:07.430]that's always been challenging is,
[00:03:09.550]you know, for using blended products like 10-34-O
[00:03:13.630]well, if you, if you're going to vary the rate of that,
[00:03:16.350]maybe to, you know, impact your phosphorous levels
[00:03:20.100]while you're also varying the nitrogen levels.
[00:03:22.210]So focusing on the element that we wanna,
[00:03:25.270]we wanna study, there's really important.
[00:03:27.410]And then doing rate studies with enough data
[00:03:29.760]to really generate response curves
[00:03:32.300]within our, our results is really key.
[00:03:35.100]And we'll talk a little bit more about
[00:03:36.230]that later, how we space those out.
[00:03:38.850]Checks are always really important,
[00:03:42.110]especially in a lot of these product tests.
[00:03:44.860]You know, what would have happened if,
[00:03:46.090]if I had done nothing, you know?
[00:03:47.620]Yeah. The crop might not have looked as, as good, but,
[00:03:50.750]but could, could that have been something I had avoided and,
[00:03:53.480]and really not taken a yield penalty
[00:03:56.120]instead of spraying the whole field for instance.
[00:03:58.350]But the other thing I would mention is,
[00:04:02.140]I'll use the term block or rep interchangeably here.
[00:04:04.990]But if you notice, like in this top study,
[00:04:07.640]we've organized this study into four blocks
[00:04:10.910]and each block contains our two treatments.
[00:04:13.050]And the bottom study, you can see
[00:04:15.180]that these reps or blocks are organized again
[00:04:18.780]with our three treatments.
[00:04:20.190]This time we're testing two different products
[00:04:23.173]and a check or do nothing, you know, strip it's important
[00:04:27.920]to have those across the field.
[00:04:31.320]And we have each one of the treatments in those
[00:04:34.600]you can see they're represented there.
[00:04:36.610]And then really the third thing
[00:04:38.230]is we try and randomize those,
[00:04:39.760]and that helps some take out that field variability impact.
[00:04:43.460]But you can imagine if I had
[00:04:44.293]high yielding areas on the east side here
[00:04:48.328]and really low yielding areas moving consistently
[00:04:50.880]to the west, if I keep the order of my treatments,
[00:04:55.040]the same across that field.
[00:04:56.760]So we basically, I'm biasing all those results in
[00:04:59.540]the same direction of that yield variability.
[00:05:02.100]So just a couple of basic things we really can't forget.
[00:05:07.130]And in terms of the data and how the number of reps
[00:05:09.470]and the importance there,
[00:05:11.370]let's look at it kind of a scenario here
[00:05:13.650]of what could happen.
[00:05:14.760]So here's this field study,
[00:05:16.410]we'll look at a couple of times,
[00:05:19.330]again, we've got six blocks or six reps across the field
[00:05:23.050]of each one of these treatments.
[00:05:23.907]And you can see this is the typical grower rate,
[00:05:26.830]I think 36,000 in this case.
[00:05:29.380]And we're going above that , and below that,
[00:05:31.040]to just look at the optimal seeding rate for this field.
[00:05:33.540]Well, if you look down in this chart,
[00:05:36.440]basically, we're looking at the partial profit
[00:05:39.090]for each one of those blocks,
[00:05:40.790]and you can see the rates here
[00:05:43.070]and what you notice is four out of six of those ended up
[00:05:46.410]having a lowest eating marae
[00:05:48.207]and the highest partial profit compared to the others.
[00:05:53.320]You look down here in the, in the results summary chart,
[00:05:56.380]you can notice, that we actually did see significance there
[00:05:59.030]in terms of partial profit with a $12/acre advantage
[00:06:03.070]of going to the lowest seeding rate in that case.
[00:06:06.689]So having the reps allows us to run
[00:06:08.760]the statistical analysis, increasing the number of reps,
[00:06:12.310]lets us detect these differences at a,
[00:06:14.530]at a more fine or higher resolution.
[00:06:17.930]Well, what would have happened in that study?
[00:06:19.550]So let's just have, for instance,
[00:06:22.130]two of those blocks and say, well,
[00:06:24.060]I just went out and threw these two blocks out,
[00:06:26.320]this is just how they happen to fall using
[00:06:28.290]that previous study.
[00:06:29.597]I'm only looking at these two sets of block data.
[00:06:33.357]And these two cases actually are our typical C536000
[00:06:39.000]in both cases would have indicated
[00:06:42.540]the highest a partial profit.
[00:06:45.220]In that case, we can't do statistical analysis
[00:06:48.630]because we don't have a minimum of free reps.
[00:06:51.390]And, and really, if you think that back,
[00:06:53.060]that first set of data,
[00:06:54.270]that basically , that decision could cost us $12/acre,
[00:06:58.350]if we make that choice moving forward.
[00:07:00.960]So again, the power of replications and,
[00:07:03.840]you know, for frankly where these,
[00:07:05.070]where these plots get placed is, is pretty critical.
[00:07:07.860]And here's kind of the worst example I would say
[00:07:09.830]from this is what if I go out
[00:07:12.190]and say, hey, you know,
[00:07:13.023]this is a new, new seed I've got for ya.
[00:07:17.280]You've been applying 36,000 seeds/acre.
[00:07:20.130]I promise you if you plant 40,000 seeds breaker,
[00:07:22.550]you're going to get a higher partial profit
[00:07:25.180]or marginal, never turn off that.
[00:07:27.500]And if we that just by happenstance, put that, that study,
[00:07:32.217]you know, one strip of that study here
[00:07:34.580]and block number six, this is the result,
[00:07:37.530]we would have gotten in that case.
[00:07:39.360]We would have actually seen by about $15 an acre
[00:07:43.030]or a higher partial profit with that higher seeding rate.
[00:07:46.440]So now, if you think about that,
[00:07:48.700]compared to what my typical practice was,
[00:07:51.350]that costs me $15/acre.
[00:07:54.000]If you think back to the original,
[00:07:55.520]the balanced well-designed study at tech
[00:07:58.620]nearly could cost me $27 per acre if I make that choice.
[00:08:01.940]So this, this issue of data quality and,
[00:08:06.570]and how we design these studies
[00:08:08.387]can really impact operations.
[00:08:10.250]And this is something we really had an opportunity again,
[00:08:13.680]for this type of,
[00:08:16.790]so how can we use precision Ag technology to help us out?
[00:08:20.160]Well, this is an example of a field trial.
[00:08:23.300]Basically we get AB lines equipment widths,
[00:08:26.120]from our co-operators.
[00:08:27.610]We create these prescription maps that have,
[00:08:30.600]you can see in this case are blocks
[00:08:33.510]that are randomly within those blocks.
[00:08:35.220]We have those different treatments that we're working with.
[00:08:37.880]And in this case,
[00:08:39.230]on this pivot irrigated field for soybeans,
[00:08:42.630]we're able to get 16 blocks out of that study.
[00:08:46.160]And in this, in this study,
[00:08:47.900]we were actually able to detect yield differences,
[00:08:50.550]statistically down to two bushels per acre,
[00:08:54.620]which is pretty impressive given that we weren't even able
[00:08:58.330]to use all the blocks that we designed.
[00:09:00.130]And I'll talk about that in a second, but,
[00:09:02.780]but the ability to design these in software upload
[00:09:06.690]those to the control monitor in the cab
[00:09:08.860]and do a product or a rate study with soybeans, corn, or,
[00:09:12.720]or maybe like the application rate of nitrogen,
[00:09:15.010]things like that.
[00:09:16.970]This is something that can be done
[00:09:19.110]and I'll show you some resources
[00:09:20.810]for that here toward the end.
[00:09:22.870]Just a quick note, typically for the studies
[00:09:25.480]that we we've been planning when
[00:09:27.700]it comes to target seeding rates,
[00:09:30.260]usually we try and have three or 4,000 seeds per acre
[00:09:34.360]between those.
[00:09:35.200]So, you know, if you see that,
[00:09:37.000]that initial study of 36,000, and we want to test
[00:09:39.930]above that and below that by 4,000 seeds per acre,
[00:09:42.930]3000, I think would be the minimum.
[00:09:44.470]You'd want to go there on soybeans.
[00:09:46.710]Usually 30,000 seeds per acre is kind of the difference
[00:09:49.580]we like to see so that we know
[00:09:52.090]we've really putting a response out there from that.
[00:09:55.210]And then in terms of nitrogen,
[00:09:57.230]usually around 30 pounds per acre, that's the,
[00:09:59.850]the difference we like to see among those treatments
[00:10:02.800]to really try and generate that curve.
[00:10:04.470]And of course, with nitrogen,
[00:10:06.160]I know folks don't like to do that,
[00:10:07.590]but if you can go with a really low rate
[00:10:09.930]or zero in some cases out there,
[00:10:12.850]sometimes that can really help you define
[00:10:14.680]that nitrogen response curve.
[00:10:16.260]It's a great idea.
[00:10:17.093]I know it doesn't look pretty,
[00:10:18.040]but a couple other quick notes think really
[00:10:21.500]about your harvester widths just kind of the
[00:10:24.550]that's the data collection methods.
[00:10:26.260]So we really have to base our studies off that,
[00:10:30.320]usually for corn it's much easier
[00:10:32.270]because we always have that common denominator,
[00:10:34.240]a little tougher in soybeans.
[00:10:36.030]So we have to think about that when we design the plots
[00:10:38.260]and then really be careful with dry spreader.
[00:10:40.850]So if you're doing some type of a product rate study
[00:10:44.310]or something like that,
[00:10:45.440]dry spreaders' inconsistent patterns,
[00:10:47.990]you really have to be careful
[00:10:49.070]where you harvest in those studies.
[00:10:51.850]Just be really careful with that.
[00:10:54.490]The second thing is that once we designed the study,
[00:10:57.360]put them out of fuel.
[00:10:58.290]You really have to get the as applied data
[00:11:00.630]back from these studies.
[00:11:01.720]This is an example, that's that study we just saw,
[00:11:05.100]we've got our blocks out in the field.
[00:11:07.400]We didn't, didn't know that this would be planted
[00:11:09.780]at an angle, the versus what we designed the studies,
[00:11:12.641]but as applied data, it lets us know
[00:11:14.980]where those rates for hopefully successfully planted.
[00:11:19.360]And you even see here, I mentioned
[00:11:20.870]before the prescription wasn't successfully loaded
[00:11:24.570]until we got up to this past here, you know,
[00:11:27.120]about a quarter of the way up the field.
[00:11:29.410]So all these blocks and the bottom really were,
[00:11:31.850]had to be eliminated
[00:11:32.930]because there was no rate change in those blocks.
[00:11:35.080]So that's really set to, in my opinion,
[00:11:39.010]get the as applied data.
[00:11:40.240]So you can validate what the machine intended
[00:11:44.660]and tried to carry out in terms of that field operation.
[00:11:48.330]Just another example, here is a study
[00:11:50.180]that we designed with our prescriptions for nitrogen rates,
[00:11:54.510]got a text from the operator when they loaded that up
[00:11:56.930]into the machine, I said,
[00:11:58.833]Hey, look at these blocks, they don't look anything like
[00:12:00.930]what you sent your prescription map, is this okay?
[00:12:04.805]And this is actually an artifact
[00:12:05.700]of a, of the GreenStar 3 systems.
[00:12:07.780]They regrade these, these polygons.
[00:12:10.350]You don't notice this in large field areas,
[00:12:12.280]but when you come down to these small blocks,
[00:12:14.630]we're trying to do where we're only talking about,
[00:12:17.078]you know, 16 rows or, or 24 rows.
[00:12:21.090]Luckily we knew that and we designed this into the study.
[00:12:24.550]So we would still get a couple of harvester passes
[00:12:27.030]within each one of these.
[00:12:27.930]So, but again, you really have to validate
[00:12:31.110]what's going on out in the field.
[00:12:35.367]I unfortunately have I think two lectures
[00:12:38.010]on yield monitor data collection and quality check.
[00:12:40.900]I'm going to try and get away with it in one slide here.
[00:12:44.770]But you know, this is our baseline of data
[00:12:48.610]that we're trying to collect, assess what we do at a field.
[00:12:51.610]So we have some best management practices out there.
[00:12:55.060]And as you get towards the fall, you know,
[00:12:57.440]we always try and get those out to folks
[00:12:59.210]that are using these systems, but calibration is key.
[00:13:02.860]You know, calibration affects every single data point,
[00:13:06.860]really trying to hone that in,
[00:13:08.430]if you can't right at the time of the harvest of the plot
[00:13:11.930]is, is really a good idea.
[00:13:14.150]And of course, if you can do a multi-point calibration,
[00:13:17.080]so four more loads to kind of span that range of yields
[00:13:21.540]you might see in the field that's critical.
[00:13:24.000]And just recall that those,
[00:13:26.020]those individual calibration loads are very small
[00:13:29.080]because, you know, we're talking about a high-flow condition
[00:13:33.282]moving down to low-flow condition.
[00:13:35.560]You know, you only need
[00:13:36.393]60 or 90 bushels per calibration load
[00:13:38.860]to generate that curve. And actually, you know,
[00:13:40.970]if you're collecting way more than that,
[00:13:42.740]you're probably negatively impacting your results
[00:13:45.200]just because you want to maintain
[00:13:47.290]that flow condition throughout that load.
[00:13:49.970]And then when we're harvesting the plots
[00:13:52.650]and you're out in the field,
[00:13:53.720]speed changes are the worst enemy of a yield monitor,
[00:13:57.440]data quality.
[00:13:58.700]Anytime the operator speeds up or slows down,
[00:14:00.870]basically that's going to impact the flow
[00:14:02.660]through the machine, the calculation of area covered.
[00:14:05.810]So it's just really a double whammy in terms of,
[00:14:08.750]you know, collecting good data.
[00:14:10.770]So try and be consistent there.
[00:14:13.790]The, the yield data itself really has to be,
[00:14:18.040]it has to go through some hyper quality control check,
[00:14:20.930]and we, we typically use a fairly standard way of doing that
[00:14:24.910]through a software called Yield Editor.
[00:14:27.340]It's available from the USDA,
[00:14:30.328]and I've argued with a lot of statisticians, not this,
[00:14:32.840]but it's the bottom line is the machinery itself.
[00:14:35.910]We know can generate errors as we're collecting yield data.
[00:14:40.180]And again, if you think about the acceleration-deceleration,
[00:14:44.040]why do I have a yield data point out in that field
[00:14:47.610]that says 3000 bushels per acre?
[00:14:50.310]You know, that's just not reality.
[00:14:52.490]And if that data point gets left in here, and again,
[00:14:55.110]we know it's a physical issue that was generated
[00:14:59.560]that we have to eliminate that,
[00:15:00.650]or else our start studies are going to be,
[00:15:02.380]you know, biased horribly for that.
[00:15:05.400]So just a, a point we'll make
[00:15:08.210]that software is actually free.
[00:15:09.860]It's a little tough to get the data in and out of that,
[00:15:12.020]but we're working on hopefully an option to help with that.
[00:15:16.440]You can look and see that it makes a difference.
[00:15:18.560]So when we go in with a raw yield data,
[00:15:21.020]the same study we looked at before,
[00:15:23.360]I'm not seeing a significant difference in my,
[00:15:27.680]in my partial profit in this case
[00:15:30.100]when we actually post-process that.
[00:15:31.960]And then generally in this case,
[00:15:33.160]it was a lot of lead time going into harvesting the blocks,
[00:15:38.350]eliminated some low yield data points
[00:15:40.230]and basically shifted our yield estimates up
[00:15:43.420]within those blocks.
[00:15:44.930]You start to see statistically, we saw differences there,
[00:15:47.910]sweeping impact, certainly impact the study.
[00:15:52.220]A couple other quick points I would make,
[00:15:56.540]in terms of in-season observations.
[00:15:59.700]Really, I think critical here, you know,
[00:16:02.260]imagery's coming a long way and it's going
[00:16:04.030]to give us the opportunity for some of these plots
[00:16:06.920]to ever fly a field, get emergence counts,
[00:16:12.151]and counts at a high resolution,
[00:16:13.966]the cost on that, you know, continues to come down,
[00:16:16.310]but even just going out into these blocks and,
[00:16:19.560]and doing, you know, a few staying counts for population,
[00:16:22.840]making sure that the, you know,
[00:16:25.090]the, the rates and the ramp rate of that,
[00:16:28.830]you know, is pretty consistent. I think it's critical.
[00:16:31.610]We've, you know, we've had a lot of studies where,
[00:16:34.230]you know, plots have just been wiped out by some issue,
[00:16:37.200]you know, rain erosion, too much pressure on that.
[00:16:40.810]And so, anyway, being able to go in and document,
[00:16:45.310]you know, this was my seeding rate study,
[00:16:47.880]you know, here were kind of the emergence rates
[00:16:49.750]that we saw to go back and even put that into the study,
[00:16:53.610]I think is very important.
[00:16:55.350]And it's getting much easier to go out
[00:16:56.567]and do these geo reference stand counts
[00:16:59.000]with some of the apps and software
[00:17:01.290]and, and things like that that we have.
[00:17:03.540]But I mean, thinking about the future,
[00:17:05.780]you know, the imagery, again,
[00:17:09.460]the resolution I think is, is really key,
[00:17:12.920]but, you know, we, for instance, have studies like this,
[00:17:15.240]where we noted wind damage,
[00:17:16.730]you know, scouting out in the field,
[00:17:19.280]noticing wind damage out,
[00:17:20.910]and then getting that imagery to look at, you know,
[00:17:23.020]what was the scale and scope of all that wind damage?
[00:17:26.530]How did that affect our study that we have?
[00:17:30.689]That field in particular has a ridge or two,
[00:17:32.556]and you can see that wind damage really hit that area.
[00:17:35.190]You can even see some, what looked more consistent here.
[00:17:38.700]You know, this is some previous crop management issues,
[00:17:41.340]but on this issue on the right, you know,
[00:17:43.940]this is a much higher resolution image.
[00:17:46.590]And, and we're looking at the,
[00:17:48.820]the red edge and near infrared,
[00:17:50.770]but you're seeing a ground squirrel damage,
[00:17:52.810]you know, throughout some of the plots
[00:17:54.610]in this particular study.
[00:17:56.070]So being able to tie that back to
[00:17:59.110]what you're actually seeing out in the field,
[00:18:00.680]I think is pretty critical.
[00:18:02.920]So we talked, we talked a little bit about,
[00:18:05.630]you know, the rep issue,
[00:18:08.020]the Ag technology's going to let us increase
[00:18:10.600]the replications we have on our field.
[00:18:14.460]Being able to put the blocks in areas of the field
[00:18:17.250]can help us identify how the field variability
[00:18:20.070]impacted the results.
[00:18:22.150]This is another example of,
[00:18:23.610]of how we can do this.
[00:18:24.630]This is a little bit different because we don't have
[00:18:27.060]that nice block design.
[00:18:28.500]This is where you and I have to start looking
[00:18:30.080]at geospatial statistics.
[00:18:31.860]But you know, this was a split planner study
[00:18:35.980]for offense and defense of hybrids.
[00:18:39.462]What you see is the, you know, those,
[00:18:41.200]those varying across the field of those two hybrids.
[00:18:44.980]And, and if you averaged all that data out
[00:18:47.660]from, from, you know, each strip,
[00:18:49.740]basically you would have gotten a two bushel difference
[00:18:52.120]in the yield, you know, no significance,
[00:18:55.010]while it's the lever.
[00:18:55.843]What really the key here was, what was the soil moisture
[00:18:59.920]and water holding capacity.
[00:19:01.540]So this high sloping area
[00:19:02.940]move it down to the low sloping area.
[00:19:05.110]What you can see,
[00:19:05.943]we quantified this versus in our CS slope,
[00:19:09.600]basically you saw all of a sudden, and in these,
[00:19:13.040]these areas where there was high water-holding capacity,
[00:19:15.220]that offensive hybrid really out-yielded
[00:19:17.910]the, the defensive hybrid.
[00:19:20.510]And then as you move up the slope
[00:19:21.910]where we have less, less availability.
[00:19:25.050]So again, that's just helping us understand
[00:19:27.190]how that variability impacts the results of that,
[00:19:30.620]or trying to average all that data in there
[00:19:32.610]was clearly an underlying factor here
[00:19:35.120]that we had to account for in that.
[00:19:36.760]So that's most of what I have for the day,
[00:19:42.980]what I would leave you with is if you're interested in this,
[00:19:46.370]we're glad to support anyone that wants to try and do this,
[00:19:50.937]any type of study, where you want to try
[00:19:52.670]and leverage your technology.
[00:19:54.680]If you're interested in learning how to do this,
[00:19:56.330]for yourself, our team, Laura and others have created
[00:20:01.640]this learning modules.
[00:20:04.500]We've taken some of our data management workshops
[00:20:06.840]with these online. So you can,
[00:20:08.860]you can learn how to design these studies with this,
[00:20:13.100]going through these modules, and these are gonna be free.
[00:20:15.580]And we'll, we're
[00:20:16.413]just gonna probably a highlight this a little bit more
[00:20:17.990]at the end of the day. But again,
[00:20:20.690]I think there's a real opportunity out there.
[00:20:22.720]And if I were just to summarize these, you know,
[00:20:25.290]these five critical steps, you have to design it correctly.
[00:20:29.500]Once you deploy the field study,
[00:20:31.050]you really have to do that as applied data check ,
[00:20:34.640]in-season scouting is really critical
[00:20:36.880]to make sure that things, you know,
[00:20:38.620]were progressing appropriately through the field and,
[00:20:41.970]and there's not some external factor affecting the study.
[00:20:45.990]And then we get down to the yield data quality,
[00:20:48.280]the collection of that data ,extremely important.
[00:20:53.050]And then finally,
[00:20:53.883]we can go through that statistical analysis,
[00:20:55.780]maybe have to incorporate some spatial issues into that,
[00:20:59.800]but, but that's kind of the process by
[00:21:02.350]which I think we can really improve the information
[00:21:05.420]that we get out of these studies.
[00:21:06.820]And hopefully that's something that,
[00:21:09.290]that might be of interest to you in the future.
[00:21:11.010]And again, we're glad to support
[00:21:12.680]and you'll have contact information
[00:21:15.440]for several people here, but I'm always glad to visit.
[00:21:20.177]If, if there's anything I can help with, just let me know.
[00:21:22.780]So.
[00:21:24.142](upbeat music)

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

OFR21: Precision Ag with Joe Luck

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments