Episode 034: Granular Guidance for Nitrogen Management
Nebraska Extension Digital Agriculture
Author
05/27/2021
Added
1
Plays
Description
This episode features Bob Gunzenhauser, manager of Agronomy Science at Granular. Bob has spent a career in agriculture and was a product manager at with Encirca prior to Corteva's acquisition of Granular. In this episode, Bob describes how the Granular nitrogen management model tool within their Agronomy Insights platform works at a high level. He also details Granular's model execution processes, user flexibility within the platform, and discusses the future of nitrogen modeling in the context of new technologies emerging in the market. As Granular continues to gain traction as a solution for nitrogen management, this episode is an important listen for understanding how Granular continues to approach nitrogen management.
Granular Info:
Website: https://granular.ag/
Facebook: https://www.facebook.com/GranularAgriculture
Twitter: https://twitter.com/GranularAg
LinkedIn: https://www.linkedin.com/company/granular/
YouTube: https://www.youtube.com/channel/UCtBb2oaaeEGwfgYxcDJCC8Q
Bob's Info:
LinkedIn: https://www.linkedin.com/in/bob-gunzenhauser-6b282a65/
Twitter: https://twitter.com/BobGunzy
FarmBits Team Contact Info:
E-Mail: farmbits@unl.edu
Twitter: https://twitter.com/NEDigitalAg
Samantha's Twitter: https://twitter.com/SamanthaTeten
Samantha's LinkedIn: https://www.linkedin.com/in/samanthateten/
Jackson's Twitter: https://twitter.com/jstansell87
Jackson's LinkedIn: https://www.linkedin.com/in/jacksonstansell/
Opinions expressed by the hosts and guests on this podcast are solely their own, and do not reflect the views of Nebraska Extension or the University of Nebraska - Lincoln.
Searchable Transcript
Toggle between list and paragraph view.
- [00:00:09.727]Jackson: Welcome to the FarmBits podcast, a product of Nebraska Extension digital agriculture,
- [00:00:14.240]I'm Jackson Stanselll. Sam: And I'm Samantha Teten,
- [00:00:16.880]and we come to you each week to discuss the trends, the realities and the value of
- [00:00:21.120]digital agriculture. Jackson: Through interviews and panels with experts,
- [00:00:25.040]producers and innovators from all sectors of digital technology- we hope that you
- [00:00:29.200]step away from each episode with new practical knowledge of digital agriculture technology.
- [00:00:34.600]Jackson: Welcome back to the FarmBits podcast for this third
- [00:00:37.920]episode of our nitrogen management series. Sam: We are pleased to welcome Bob Gunzenhauser,
- [00:00:42.640]agronomy science manager at Granular for this episode of the FarmBits podcast.
- [00:00:47.200]Jackson: Bob will offer a little more about himself
- [00:00:49.600]and his role at the beginning of this episode. Sam: This episode focuses on nitrogen modeling
- [00:00:54.240]from an industry perspective and is intended to provide a look under the hood of a system that is
- [00:00:58.720]being used for nitrogen management commercially. Jackson: That's right, there's a lot of great
- [00:01:02.320]content in this episode from end to end, so let's get right into it with our interview with Bob.
- [00:01:10.320]So, what are some of the primary features of this Granular platform that you're offering to
- [00:01:14.800]growers and kind of what is your overall objective and you know helping up their
- [00:01:19.840]nitrogen recommendations this time of the year? Bob: Yeah, the Granular agronomy is kind of the
- [00:01:24.720]platform, the overall banner for what we're working on here. We do have Granular business
- [00:01:29.200]which is a farm management software that's really worth what the heart of Granular was as we
- [00:01:34.160]came to be part of Corteva in 2017. We also have Granular insights which is the kind of lighter
- [00:01:40.080]weight version, has elements of both in there. Then we also have acre value which is a farm land
- [00:01:46.880]evaluation service but Granular agronomy is kind of the heart of the focus on the agronomy side,
- [00:01:52.080]and we deliver that through certified service agents who are independent representatives of
- [00:01:57.120]our product. Granular agronomy right now is focused on nitrogen management, corn
- [00:02:02.000]fertility. We say fertility, it's phosphorus, potassium, lime, sulfur primarily in corn soybeans
- [00:02:07.920]and then also variable rate seeding also in corn/ soybeans, and that's been basically a carryover
- [00:02:13.200]from our prior in circuit years with Pioneer and so nitrogen management is our focus there.
- [00:02:20.880]What we do is again utilize that crop model we are running through thousands of simulations for
- [00:02:28.560]a given field. So, the inputs that we use there are our weather our soils management information
- [00:02:35.120]and crop genetics as well. This crop model can run through all those different scenarios of
- [00:02:42.000]looking at different zones within the field, they have different soil properties,
- [00:02:45.680]they have different yield targets and we have 20 years of weather history that bring in along
- [00:02:49.920]with the weather that you've already had this year. We take in your management, so if you
- [00:02:54.400]apply fall anhydrous with an inhibitor or you apply spring uan plus a top aggressive urea,
- [00:02:59.840]we handle those properly and the proper breakdown of the nitrogen into the system and with different
- [00:03:06.240]hybrids if you plant a long season or a short season hybrid. We bring that in as well so we try
- [00:03:11.040]to be as comprehensive as possible. We know where there's always things that a model doesn't catch,
- [00:03:15.600]but there's a lot of things we are trying to bring into that model to simulate and the end goal of
- [00:03:19.760]all this is to help growers to apply the right nitrogen rates across the field knowing that you
- [00:03:26.640]know soil textures vary, water flowing capacity varies but also for a given location weather
- [00:03:33.200]will change. And so we look at that 20 years of weather history to kind of give us a guide
- [00:03:37.760]as to what could happen in the future. We can't predict the future, no one's figured that one out
- [00:03:42.640]but we can definitely say you know okay here's what you could have and what's out of the 20 years
- [00:03:49.520]what's likely to that drought or what's likely to vary what years and we build that in. So,
- [00:03:53.840]then our recommendations come out to be based upon that you know kind of the middle range
- [00:03:58.960]between all those extremes you know it's going to be a recommendation of a rate that says
- [00:04:04.320]you know it's going to be sufficient in those years where you may have a lot of loss
- [00:04:08.640]but it's not going to be overly done. If we have a year where we have a drought and
- [00:04:12.000]the plant just doesn't take it up, I mean we can't solve for all of those yet those
- [00:04:16.160]scenarios but at least we can try to be in the middle of that to cover all the bases there.
- [00:04:21.680]Sam: So, you just described one of these challenges with the variability in weather
- [00:04:26.720]but can you elaborate on some other challenges with putting that right
- [00:04:29.920]nitrogen rate in the right place in the field? Bob: Oh yeah you know, with with nitrogen it's
- [00:04:34.720]you know the soil is a terrible place to store nitrogen you know but we've got to
- [00:04:38.160]put it there to grow our crop. You know we want to be able to put in the right time,
- [00:04:43.040]but there's also a lot of operational situations so that's why we have fall anhydrous in a lot of
- [00:04:47.040]places in the Midwest. But, we also when we put in the ground we're you know chancing it to loss,
- [00:04:52.960]and so that conversion of ammonium from anhydrous ammonia for example to to nitrate
- [00:04:57.920]we have opportunities for laws for leaching or denitrification can go either way up and down. We
- [00:05:03.680]can have a urea application that's applied in the summer when we have a volatilization loss as well,
- [00:05:08.880]and so our model underlying model simulates all those dynamics, so it looks at soil moisture,
- [00:05:14.400]it looks at soil temperature looks at soil organic matter even to figure out mineralization as well.
- [00:05:20.880]You know, soil temperature and moisture once we apply that urea
- [00:05:24.960]and if we do use a stabilizer whether it's an above ground or below ground stabilizer,
- [00:05:28.880]we similarly can simulate those as well. So, the challenges there are you know growers know that
- [00:05:35.520]you know nitrogen is expensive and this year is no exception. They also know that you know losing
- [00:05:41.440]nitrogen is not good for their pocketbook nor is it good for the environment right and so we want
- [00:05:46.080]to be able to show them, run through different scenarios and say hey this is what you could
- [00:05:50.400]you know with this operation with this practice you've been doing, here's what where you could
- [00:05:54.480]be losing nitrogen at in these different soil types and if you want to change to a different
- [00:05:59.920]approach maybe you want to use that stabilizer, maybe you want to go to a later timing maybe
- [00:06:03.520]you want to inject that instead of put it on the surface. We can show that yeah it can bring down
- [00:06:07.840]your nitrogen losses in those very situations. Jackson: So, when you think about these models,
- [00:06:13.520]I mean obviously models are driven on data and you've already talked about some of these
- [00:06:18.320]important data layers that you're bringing in on your side, but what kind of goes in I guess you
- [00:06:23.680]could go in a little bit more detail on those critical data layers but also talk about what's
- [00:06:27.680]necessary for a consultant or grower to bring to the table from the data side to get their
- [00:06:32.560]field set up to run this model properly? Bob: So, the primarily spatial data input
- [00:06:38.560]that we use is what we call environmental response units or ERUs. This is something that we developed
- [00:06:44.160]internally, and it's it's basically the best I can describe is kind of a rearrangement of the
- [00:06:49.520]USDA sergo soil proper layers so we take the sergo soils which everybody can get you know
- [00:06:55.040]look at the NRCS and review those and they were done over a number of years by different folks
- [00:07:01.200]and they have varying levels of accuracy right. I know their resolution in say Iowa is a little
- [00:07:08.240]bit better than I think in Nebraska or North and South Dakota, Kansas and so they you know there
- [00:07:15.360]are some locations some counties are very detailed even with Iowa and some are kind of coarse. So,
- [00:07:20.480]we take that basic information because we figured there were probably more right than wrong
- [00:07:25.680]but then we update them using high resolution elevation data. So, we use like Lidar for example
- [00:07:32.160]and from that Lidar high resolution elevation data, we come up with nine or ten different
- [00:07:37.360]topographic derivatives from that and you know slope curvature, elevation from the
- [00:07:44.240]stream bed things like that and so that through that we use a sampling process. So, let's take
- [00:07:50.640]you know a soil that you may have in Nebraska, Holdrege soil silt loams one to three percent,
- [00:07:57.600]so we would look in that area and look in that watershed where that soils at and say okay what
- [00:08:01.120]are the different topographic properties of that Holdrege soil loan one to three percent that's
- [00:08:06.160]mapped there, what are the different oh it's on this elevation, it's got this slope, it's got
- [00:08:09.760]this curvature okay. Then we use a and we repeat that for all the different soil properties or
- [00:08:14.880]soils found in that watershed and then we go back and remap them basically use the machine learning
- [00:08:20.320]process to say okay we find that you know they mapped the Holdrege silt-loam and these you know
- [00:08:26.080]kind of characteristics anywhere we find those characteristics from a topographic standpoint
- [00:08:30.560]let's place the Holdrege silt-loam there and let's where it's on a slope okay that's a different soil
- [00:08:35.280]or it's a depression it might be a fillmore you know something like that where we basically go
- [00:08:39.680]through and we can figure out where those soils should be and we sharpen the positioning of
- [00:08:44.240]the soils. We're still using the same USDA soil properties that the book values behind the scenes
- [00:08:51.600]now position them better, so that's one thing we first do. The next thing, we work with
- [00:08:55.920]growers is bringing in their yield data so multi-year yield analysis is working across
- [00:09:01.920]normalizing yield data for multiple years and we create zones of low, medium and high productivity
- [00:09:06.880]from that. So, these are kind of what we end up doing is kind of intersecting if you think
- [00:09:11.680]we have the soil in one layer and we have the ERUs one layer and the
- [00:09:17.200]multi-year analysis on another. They intersect with each other so we create some zones within
- [00:09:21.840]there. Another level is that we do also if you have irrigation you will section that out as well,
- [00:09:27.760]so we also have some options for our CSAs to brand other things that can intersect
- [00:09:32.800]and dissect if you will those ERUs. The ERUs will still have the same soil properties behind them,
- [00:09:39.760]but now we have different sections where we might assign different yield targets and that
- [00:09:44.080]yield targets are based upon that historical yield productivity. So, we know that even with ERUs we
- [00:09:49.440]try and get that sharpened but we know that yield doesn't always follow those nice you know lines
- [00:09:55.760]soil type maps, so then we intersect them again and say okay well you know hey this might be,
- [00:10:00.400]we may have three or four intersections with the soil type and they may have different yield
- [00:10:04.320]properties or yield targets through there and that might be due to things that we just don't capture
- [00:10:09.200]in the ERUs. They might be due to fertility issues or drainage or whatever it might be,
- [00:10:13.840]but we assign those different yield targets there and those are drivers for both those
- [00:10:19.440]yield targets and the sole properties are drivers into the model into our product there.
- [00:10:23.440]Sam: That's really interesting, so now that you have all this data how are you converting
- [00:10:30.480]that into a nitrogen rate is it from some of these trials that you described earlier
- [00:10:35.520]or can you go into that a little bit right? Bob: So, you know we do periodic updates of
- [00:10:40.800]the model in the product on the front end every so often, and those are updates that are based upon
- [00:10:47.440]field trials so we have a collaboration we have with the university of Nebraska. This year we have
- [00:10:52.320]had historical collaborations 2014, 15 and 16 what we call the pioneer regional nitrogen trials PRNT.
- [00:10:59.440]We have also a series of internal trials that we work with small plot work on pioneer farms
- [00:11:05.200]and also kicked off last year and again this year working with our CSAs to work with growers,
- [00:11:11.440]their customers to do nitrogen rate trials, so all that data is then used to fine-tune the model to
- [00:11:17.920]find gaps, places where we need to do improvements there and once we do a periodic update
- [00:11:23.600]those are brought in, but for this on the you know the every year a setup with the grower,
- [00:11:30.160]the CSA there's a time that's spent by the CSA and the grower to sit down say okay what's a
- [00:11:34.640]reasonable yield target for this field right is it 240 bushel yields is it 230, 220 whatever it is
- [00:11:41.520]and then these zones within these things we call them decision zones. We assign different yield
- [00:11:46.000]targets there and so we have a first process okay looking at multi-year yield analysis. We think it
- [00:11:50.960]could range in say 240 bushel corn for entire field. We can take that and kind of just spread
- [00:11:55.920]that across the field proportionally so okay this one over here might be 190 this is one here 260.
- [00:12:01.840]Is that really right Mr. farmer and we can adjust those to get them tuned in, so there's a process
- [00:12:06.720]that our CSAs have to sit down and kind of work through that with our growers to get those tuned
- [00:12:11.840]right there. So, that's the big input that we bring in and then also their management. Okay what
- [00:12:16.720]do you, when do you tend to apply what do you what are you gonna use are you gonna use anhydrous,
- [00:12:20.480]UAN, urea are you gonna put a variable rate map or dab application? Okay let's bring that in as well.
- [00:12:28.240]We can capture that what's your senior rate what's your hybrids and if we have variable rate seeding,
- [00:12:34.400]we can bring that data back in and use that as inputs as well. So, every zone may have different
- [00:12:39.440]as planted seedling rates. Those are inputs back into the model, as well.
- [00:12:44.240]So, we take a lot of that grower information as much as we can and put that back in the model.
- [00:12:49.200]So, it is a bit of a communication process with our CSAs to talk through that
- [00:12:53.440]with the growers, but we think that's a way to really dial it in there.
- [00:12:58.000]Jackson: So, if you have a grower right you talked about earlier you can kind of update
- [00:13:02.640]the model with weather as you go throughout the season, so if you've got a grower who's
- [00:13:06.240]making maybe multiple split applications over the course of the season whether that's fertigation or
- [00:13:10.480]maybe they're actually getting out there with a ground rig for two side dress applications,
- [00:13:14.560]how exactly does that work you know when you sit down with the CSA at the beginning of the
- [00:13:17.440]year and maybe you have this total nitrogen goal but you're not sure exactly when you want
- [00:13:20.720]to put those applications on is that something that Granular works well with or you know is
- [00:13:25.120]it kind of built for that sort of application? Bob: Absolutely, there's a couple of different
- [00:13:28.560]ways to go about that you know we the grower may say hey I typically I'm going to do this program
- [00:13:34.480]and the CSA will plug those in with some general dates. I mean this might be back in February,
- [00:13:39.600]March time period just like irrigation they don't know exactly when they're going to run the pivot,
- [00:13:44.880]but we think generally we're going to run it every you know a few days this amount per
- [00:13:49.040]inch for every time, and so they'll have those plugged in and then those get updated as
- [00:13:55.600]we you know communication with it, with a farmer there might be a planning situation where a farmer
- [00:14:00.720]says hey I what would this look like if I went to a split app would it instead put it on front,
- [00:14:05.360]what if I went to a two in season applications and they can run that scenario with the grower
- [00:14:10.640]and see if that does lower their application rates. It doesn't tie it in better also as we
- [00:14:16.240]go through the season you know here we're if we do our planning back in the winter, early spring
- [00:14:21.360]we don't know what the weather is going to happen or what's going to happen with that weather but
- [00:14:24.480]as we get into the season let's say we get into you know late May we've had some weather already,
- [00:14:30.000]we know it's been wet or it's been dry and so then we can update that recommendation
- [00:14:35.040]to say oh, well weather's occurred we had a lot of air we had a lot of you know margin
- [00:14:40.800]before because we were using lots of weather from previous years. Now we've had three more
- [00:14:45.280]four more months of weather we know it's a little bit tighter now we know kind of we
- [00:14:49.520]don't know what the weather the rest of the year's going to look like has looked like so
- [00:14:54.080]therefore we can fine-tune that recommendation even better there and then get that proper date
- [00:14:59.040]and timing and all that with from the grower. Jackson: Sure, and so when you have these multiple
- [00:15:03.760]applications you are I assume generating some prescriptions that folks can go out there and run
- [00:15:09.600]and transfer out there to the tractor when they are given these prescriptions,
- [00:15:14.800]how much flexibility is given to a producer to edit that prescription and say hey you know maybe
- [00:15:20.480]70 pounds in this area felt right you know when I was talking to my CSA two months ago,
- [00:15:25.040]but now I'm feeling more like 45 would be the kind of the economically optimal is that
- [00:15:28.960]flexibility given there within the platform? Bob: Oh yeah, absolutely so when we create
- [00:15:33.440]those recommendations it's kind of I would say it's the first cut I mean we give our best there
- [00:15:39.200]that's really the beauty of having a CSA involved with this. I cannot stress this enough having that
- [00:15:43.680]human intellect that trained local person involved there is super important there.
- [00:15:48.480]So, when we're going to go asking a grower to apply a certain amount you know they're going to
- [00:15:53.200]say the the the CSA is that person that interferes to say okay the model's saying say you know apply
- [00:15:59.920]average 90 pounds in an acre and the grower might say oh that's okay that's a little more than I
- [00:16:05.360]figured it would be or hey are you capturing the fact that you know my applicator can only go down
- [00:16:10.640]to 35 pounds of nitrogen per acre and get down to 10 gallons an acre like my UAN oh yeah we can plug
- [00:16:16.000]that in so they'll they can put our system allows you to plug in a minimum or a maximum allows you
- [00:16:21.280]to adjust everything up by 10 down by 10 whatever the adjustment might be, go into individual
- [00:16:27.760]decision zones and adjust those properly as well so that is all editable by the CSA because there
- [00:16:35.680]are times when that model will put out something that like wow that was not what I was expecting.
- [00:16:42.640]You know what I'll admit that that's what every model every model will have its faults every model
- [00:16:46.960]has areas where it does not know something and so that's where having that human loop that CSA is
- [00:16:52.800]making that call saying no that just doesn't smell right, let's work with the grower to adjust that
- [00:16:57.760]properly let's make sure it's going to be right there it doesn't happen not very often I don't
- [00:17:02.480]believe but it's just that one level of assurance that we can bring to the farmer to say we're going
- [00:17:07.680]to make sure it's going to work for your farm. Jackson: Yeah and I'll just throw this in there
- [00:17:12.720]this is one quote that i've always heard that I found fascinating,
- [00:17:15.280]all models are wrong to some extent, but there are some models that are extremely useful and I think
- [00:17:20.400]that really fits well with nitrogen modeling. Bob: That's exactly, I've used that same quote
- [00:17:24.480]when I talk about our product with others and say with our CSA saying yeah every model is
- [00:17:29.280]going to have errors, it's all going to be wrong in some place or another
- [00:17:32.400]but some are more useful than others there's still value in using them,
- [00:17:36.000]they still can provide us guidance and that's a really key thing I'd like to say is guidance
- [00:17:41.680]if I make a recommendation or prescription sounds like I'm and sometimes we do that,
- [00:17:48.400]but I like the word guides say here's generally where we want you to go here's the the concepts
- [00:17:52.640]there and so that's where our product really can work there is to say hey knowing all the soils
- [00:17:57.360]that you may have in a field knowing that whether you've had knowing the management, you've done
- [00:18:01.680]here's where we think is best but that CSA you and the farmer and CSA are that final arbiter making
- [00:18:08.000]that decision what they could put out there. Sam: That's an excellent point, so then how are
- [00:18:14.400]you evaluating how if how correct these guidance recommendations are there rate blocks out in the
- [00:18:21.440]field check strips how are you evaluating these are correct both for the producer
- [00:18:25.440]but also for you guys developing the model? Bob: Yeah you know, when we first introduced
- [00:18:30.240]this back in 2015, we did kick off some on-farm studies where we compared the grow rate versus
- [00:18:38.160]a you know a nitro that type of circa but nitrogen recommendation for our product in a lot
- [00:18:43.840]of those cases. As we went through those, we were winning quite a few of those you know up to 60-70
- [00:18:49.120]we felt that was a good call there more recently with these on-farm trials that
- [00:18:54.000]we're doing with growers, we're collecting that same kind of data to see okay what is the
- [00:18:59.440]nitrogen recommendations on the outside of the plot and within the plot, what was the
- [00:19:03.760]economic optimum nitrogen rate as well. So, we can really capture that new development that
- [00:19:09.200]we're looking at down the road here very soon is the ability to put in check strips, very easily so
- [00:19:15.040]it was something that our CSAs have been kind of doing part wise using our eo360 studio software.
- [00:19:22.720]They take the wreck and they go in there and cut it up and you know put some different you know a
- [00:19:26.640]grower rate versus you know the recommendation and we felt we need to really automate that,
- [00:19:31.760]so the new tool that's still forthcoming would allow you to say hey I'm going to do a script
- [00:19:36.960]that's taking the entire field and my application width is say 60 foot or 120 foot or 90 foot,
- [00:19:43.120]and it'll create two different blocks strips right next to each other right and so that way
- [00:19:47.360]we can intersect and say on one side is going to be a variable rate we're just going to cut that
- [00:19:51.360]and use that the other side might be a straight rate might be the growers or it might be the same
- [00:19:55.600]variable rate bumped up by 10 or down by 10, and we really want to be able to use that
- [00:20:01.600]to automate the data collection process and one thing we've heard back from our CSAs is
- [00:20:07.200]give me tools that really show how good I'm doing, so let me just let me show how good I
- [00:20:11.840]am how I'm delivering value to the farmers and duh yeah absolutely they should be doing that.
- [00:20:19.280]They're bringing, they want to show the value to their growers their customers every year
- [00:20:23.120]and I think those are proper tools to have there. Jackson: Absolutely and you know kind of on this
- [00:20:28.560]whole theme of looking forward, I mean I think that evaluation piece is really really critical,
- [00:20:33.200]but you know obviously the models have to continue to adapt to new products that are
- [00:20:36.880]coming on the market. We recently talked to Pivot Bio, you know who's producing this proven product
- [00:20:42.240]that is kind of a new way of supplying nitrogen to the crop throughout the growing season, how
- [00:20:46.320]are biological products and some of these other things like you know crop sensors or you know
- [00:20:51.520]nitrogen inhibitors that are coming on the market right now how are those able to integrate
- [00:20:55.280]into the model and how are you preparing for those as a new nitrogen management strategy?
- [00:20:59.120]Bob: Oh yeah so, in the case of stabilizers you know Corteva has a history of that within
- [00:21:05.040]servant instinct that's been a product has been around for many years and we do simulate
- [00:21:10.000]the effects of that within them all today. Same with the above ground inhibitors, so
- [00:21:15.520]a urea inhibitor we can simulate those as well and those are based off of public research data
- [00:21:21.760]that we've we've collected as we get into talking about with biologicals that gets a little more
- [00:21:27.040]complicated. You know, that's something we are studying, how we could bring those
- [00:21:32.000]in both from look at public literature as well as you know our own on trials as well.
- [00:21:38.000]So, those are definitely areas that we are very interested in. We know that that's going to be
- [00:21:42.720]a key part of a strategy towards nitrogen management, the future and we don't want to
- [00:21:47.040]leave that out when it comes to other sensors you know on-the-go machine sensors or remote sensing
- [00:21:54.240]we don't personally have anything in there yet. We definitely evaluate that you know
- [00:21:59.440]maybe I'll introduce a concept that I've been working out here is that there's a
- [00:22:02.800]when we talk about data inputs into a model there's the genotype typical data and there's
- [00:22:07.520]the phenotypic data, so the genotypic is maybe the soil data right we know some about the texture or
- [00:22:12.880]the waterline capacity, the organic matter, the topography it's very stable things. But,
- [00:22:18.560]it's also pretty hard to capture all that too you know, I think about again have the
- [00:22:22.880]coarseness of our servo maps and you know we have ideas of a proximal soil sensing using
- [00:22:28.960]that as a way to understand data but also then calibrate with measured soil properties and the
- [00:22:34.960]phenotypic data is what I think is like yield data or remote sensing data or other types of things
- [00:22:41.600]they tell you kind of the expression of the crop, what did what did it yield what does it, how green
- [00:22:47.040]is it, what is the chlorophyll you know these type of things. Those are great to know where they're
- [00:22:52.480]at in the field, but we don't know why we don't know why it yielded there just if you get a yield
- [00:22:56.640]map, it's like okay well hey look 240 bushels part of the field, we don't know what caused that and
- [00:23:01.840]so you need the genotypic data with a model to get you that point, and so the same thing goes
- [00:23:07.680]with these other these sensors, remote sensing has a lot of great potential because we can cover
- [00:23:13.120]millions of acres very easily, very cheaply, but what I am looking for from our scientific teams is
- [00:23:20.080]how do we take that phenotypic measurement and convert that into genotypic or give us guidance in
- [00:23:24.800]the genotypic are we seeing from a remote sensing that oh wow in August this part of the field is
- [00:23:31.120]lower you know biomass lower chlor has lower reflectance values is that due to nitrogen running
- [00:23:36.960]out or is it due to drought is it because it's on the hill, it's run out of water is it sandy
- [00:23:42.800]whatever it might be, so help me understand help us to understand why those things are going on
- [00:23:48.000]there from a genotype standpoint and then we can really unlock a lot of things with the phenotypic,
- [00:23:52.720]with the remote sensing that's the challenge inside there's a lot of solutions out there use
- [00:23:57.920]remote sensing by themselves, but they all have to kind of go through some sort of model to be useful
- [00:24:03.520]just otherwise they're just pretty pictures. Jackson: One thing we like to talk about a lot
- [00:24:07.360]on this show is aggregated data, and so I'd love to hear a little bit about how Granular you know
- [00:24:13.120]what your policy is towards aggregated data but then also how you're using that to help to inform
- [00:24:19.440]the model a little bit better and try to like you said get to that EONR for a specific field?
- [00:24:24.640]Bob: Yeah, so with aggregation of data we will never take you know a farmer's data and sell
- [00:24:32.880]it off with their information right the only time we're going to do something with that
- [00:24:37.120]is anonymously put together and study across you know large areas and so more specifically
- [00:24:44.960]on the ag business side you know where you have financial information involved there that's off
- [00:24:49.440]limits right. That's just yours only between you and your customer support manager, but that's it
- [00:24:57.200]so right we do collect millions of acres of yield data, millions of acres of vast plant data
- [00:25:02.480]and that's just a given there along with all this magic information that's applied through nitrogen
- [00:25:08.800]management tool. To be honest with you, those data sets while they're very broad and very large
- [00:25:18.160]without the individual context of what did you do with nitrogen, what did you do with fertility,
- [00:25:23.360]what did you do with your planting all that other contextual information. It doesn't make it as
- [00:25:28.560]useful as one would think right, so that's where the trials that we do on farm or at our research
- [00:25:35.280]locations are the most valuable to be honest with you because number one we've got a lot of
- [00:25:40.320]that context around what did they do, what's been applied, what are the different rates. Number two
- [00:25:45.600]we are varying the rates quite wide on those. I would rather have one you know,
- [00:25:53.040]one acres worth of small plot work with varying rates replicated four times. I'd rather have that
- [00:25:58.720]than a thousand acres of somebody applying the straight rate of nitrogen across all of them.
- [00:26:04.240]It'll be much more useful for me to have one acre of that information with the varying rates
- [00:26:08.720]because you know if you apply you know 200 pounds of nitrogen across all those acres
- [00:26:12.160]okay fine that's great. But, I don't know if that would have worked at 160. I don't know if that
- [00:26:16.720]would have worked at 240. So, that straight rate of information good for you, you can see where
- [00:26:21.280]your variability is at. That's great but for me from a wildlife standpoint it's not as useful.
- [00:26:25.520]Sam: Yeah, we've also seen a lot of data quality issues when you're dealing with a
- [00:26:30.960]lot of on-farm research files, so it's probably better to have good quality data than necessarily
- [00:26:36.560]some medium quality data from lots of sources, so that's really interesting.
- [00:26:40.160]Bob: Yeah, you know the axiom of you get what you pay for is so true you know
- [00:26:46.800]we would love to be able to create a better demand for or better reason not demand but
- [00:26:51.120]a better reason for growers to calibrate their monitors to set up their monitors properly their
- [00:26:55.360]applicators get the latest information properly because of all their things like well it's just
- [00:27:00.480]I'm making a map a big deal. I know what I did there's a certain amount of value there, but if
- [00:27:06.160]we show them say hey by doing this work now you can do more analysis on your farm. You get better
- [00:27:12.560]you know and maybe you know as you would maybe opt into a anonymous peer network of sorts perhaps you
- [00:27:18.800]want to be able to do a better job of collecting data then we create an incentive for that,
- [00:27:24.640]so that's you know we always kind of deal with it. As I've dealt with many
- [00:27:28.320]years of yield data you know one-third of it is really good stuff one-third it's okay and
- [00:27:32.800]one-third of it probably just throw away you know saying that's not true that's not making a first
- [00:27:38.000]half of a sales area on this though right. Jackson: So, you've kind of brought up the
- [00:27:44.640]financial side in one of your recent responses, and I don't know if I didn't catch this earlier
- [00:27:49.680]but is financial data being incorporated into the model outputs as well or is that something
- [00:27:54.400]that maybe the the new relationship with all this old en circa kind of you know model coming
- [00:28:00.480]into play with what granular brings on the financial side is that something that you're
- [00:28:03.680]looking to try to integrate here in the future? Bob: Yeah, and you know as we've been iterating
- [00:28:08.960]through different versions of our product that's of nitrogen that came over from the en circa days,
- [00:28:16.480]we did include there the cost of nitrogen so that we could you know compare that make this part of
- [00:28:22.320]the output there, and I think that's something we're going to incorporate this newer version
- [00:28:25.280]that we have built in with grip insights as well. At present, we don't bring in the cost and when
- [00:28:32.240]we make that decision for a variable rate we don't bring in that cost so we're not really
- [00:28:35.360]doing a true process today. That's an area where we're doing a lot of research on what maybe we
- [00:28:40.880]want to move towards in the future but right now it's not. I think you know when you deal with the
- [00:28:46.720]price of corn that could go from you know this year went from three dollars to five dollars
- [00:28:50.880]nitrogen from 30 cents to 50 cents a pound, definitely that's an area we need to be bringing
- [00:28:56.160]in so we know it's also a little bit hard to get growers to always provide that sometimes.
- [00:29:00.480]A little harder to cost to figure that out that cost there but you know in my studies
- [00:29:05.360]of doing that EONR analysis, if I do have three dollar corn versus five dollar coin
- [00:29:10.560]that's gonna shift my EMR up maybe 10-20 pounds of nitrogen. My cost goes up of fertilizer it
- [00:29:16.880]may move it down in our 10 or 20. So, it does it does figure in the equation yeah.
- [00:29:24.160]Jackson: So, I guess one of the other questions that I had as we kind of talk about the software
- [00:29:32.320]when you're thinking about these CSAs who are using it obviously these CSAs still want to have a
- [00:29:36.960]very usable interface to deal with, and they want to streamline the process as much as possible.
- [00:29:41.920]What have been some of the major focuses on the development side to make the program
- [00:29:46.160]easy to use and really appealing to these CSAs to get involved with y bulk editing?
- [00:29:51.600]Bob: That's the big one right there is the ability to say I want to apply these treatments
- [00:29:58.400]across these in these fields. I want to kick off variable rate scripts to run all those at once,
- [00:30:03.920]I want to export them all out at once as well. So, the bulk editing bulk setup is
- [00:30:08.960]really important there. When we first developed this a number of years ago it was like on a field
- [00:30:13.120]by field basis it was you know I almost want to say it's a bit of a perfect concept, but it was
- [00:30:18.320]you know the bulk editing was kind of an app you know. So, I'm like oh hey we proved
- [00:30:22.160]you can create variable rate risks well that's great, but I want to do a bunch of my time okay.
- [00:30:25.840]So, as we iterate on the product we've added that functionality, and I think that's the area that
- [00:30:31.440]we we heard a lot of feedback from CSAs. I want this bulk editing function, so I can you know work
- [00:30:36.960]with lots of growers and it makes sense because if we can give them the ability to do more work
- [00:30:42.640]quicker, they can now take on more customers. They get more profit it makes our service even
- [00:30:48.560]more valuable to them. So, let's do that let's make that very simple and so with
- [00:30:53.120]the the power of cloud computing, when we say hey we want to be able to kick off you know
- [00:30:58.000]recommendations for 20 different fields that can run there and as I mentioned this thing runs,
- [00:31:03.360]this model will run thousands of iterations. You take a you know a a field with maybe say
- [00:31:10.400]30 decision zones you have 20 years of weather at the minimum that's 600 simulations we have to run
- [00:31:16.960]there, may be more within that to find that right nitrogen rate. So, now you're talking maybe
- [00:31:21.680]five runs per recommendation or a decision zone by year, so now you're talking 3,000
- [00:31:27.440]take that across 20 fields that's you know 60 simulations we all have to do
- [00:31:33.200]with run through this now that all can run in in a tenth or a very small amount of a second.
- [00:31:40.160]But, you know if you stack them up one after time it'd be you know hours to get all done
- [00:31:45.360]beauty of cloud computing is we send that out spins up different examples of instances of
- [00:31:53.040]that runs all that work in parallel brings it all back simulates, it put it all together.
- [00:31:58.080]So, instead of taking hours to get all those recommendations we get back in a few minutes.
- [00:32:03.920]It's really important. Sam: Yeah, that's awesome.
- [00:32:07.360]Can you talk about where do you think the future of nitrogen modeling is going
- [00:32:12.800]maybe some of the ideas that you guys have in the pipeline or maybe even more futuristic
- [00:32:17.200]can you go into that a little bit? Bob: Yeah, I think
- [00:32:20.720]an area that we need to broach as an industry but we need to do it carefully is to show farmers a
- [00:32:27.200]little bit more how they're losing nitrogen how they may look through these different practices,
- [00:32:33.200]and we do that today to an extent. I think maybe bring up a little more and maybe figure that in to
- [00:32:38.960]allow them to figure that into their equation of recommendations may be useful. I would rather be
- [00:32:44.880]able to show them that information you know from an industry standpoint and help them to guide to
- [00:32:49.280]make better decisions then that be something that is dictated to them from a governmental
- [00:32:53.840]standpoint. Now, the government is there to everyone and you know all of our resources,
- [00:33:00.560]but I want farmers to do the right thing on their own, and if they have the right information I
- [00:33:04.960]think they can make the right decision, so that's an area where I really feel strongly having farmed
- [00:33:10.400]myself and coming from fifth generation farming. You know I would like to know that information
- [00:33:15.920]and so if you're telling me that hey the practice I've been doing for ever since grandpa told me
- [00:33:20.960]how to do it this caused me to lose a lot of nitrogen, and I could be doing better I could
- [00:33:25.360]be saving myself nitrogen I wouldn't have to be sent so much downstream or in the air
- [00:33:29.520]hey yeah great that's awesome so that's something I really believe is we need to surface more and be
- [00:33:35.440]and be more transparent as an industry around. I know it does bring in some questions of like well
- [00:33:40.000]how accurate is that estimation of loss is that something we should be you know looking at, and I
- [00:33:43.840]think you know we're going to use the best science we can there's always science is always evolving
- [00:33:48.720]it's never static, so do the best we can with what we we have and if we find improvements we'll have
- [00:33:53.600]that too. I think that further things down the road that are really exciting in the space as
- [00:34:00.400]we talked about was is biologicals, but also I'm going to say as we get into learning more about
- [00:34:05.040]the soil biology itself, so not just what would we put on there in terms of a product, but also what
- [00:34:12.560]kind of bacteria and fungi are in the soil and the technology that's coming out around that to allow
- [00:34:18.480]the analysis of that. It's really interesting it does cost a lot right now yes but anything
- [00:34:24.000]with DNA analysis certainly the price will come down over years and you know volume with that
- [00:34:28.560]and what I'm really excited about is learning are there certain soils that mineralize better
- [00:34:33.200]than others. So, take you know for a four percent organic matter soil even one that's on one side of
- [00:34:39.360]the fence versus the other side of the fence, do they both mineralize the same rates if not
- [00:34:44.480]why is it because this one has better bacteria does it have better whatever else there are
- [00:34:49.840]conditions there okay and so being able to test for that and understand that is great
- [00:34:54.560]you know the mineralization portion, the ability to mineralize nitrogen out of organic matter
- [00:34:59.680]really opened my eyes a couple years ago when we did some trials in Minnesota where we had
- [00:35:05.200]different nitrogen rates from zero to 280 pounds of nitrogen. So, we always you always
- [00:35:08.960]apply something that's very low and very high to really establish that UNR curve. At this location
- [00:35:14.960]near Mankato the zero nitrogen rate made over 200 bushel corn like wow that's crazy and the the high
- [00:35:23.680]nitrogen rate you know made 240-250 you know not a little bit more but not that much more that's like
- [00:35:30.320]there's something magical as well and you know we're pretty good it's like it was five percent
- [00:35:35.280]or so is really good, but there was something magical and I had you know we know more about
- [00:35:40.720]the biology there okay let's dig into it. What's going on there what is going to happen,
- [00:35:46.080]because if we can you know you can imagine down the road if we had a very lower cost way of doing
- [00:35:51.520]soil analysis and so doing your two and a half acre grids, your zones we came across that hey
- [00:35:57.280]these parts of your field mineralize better right and under the right conditions you won't need to
- [00:36:01.920]put on as much nitrogen wouldn't that be amazing? You could put on 30 or 40 lbs. less of nitrogen
- [00:36:09.360]and still have good yields here. Now it depends on the weather it may be more may be less whatever
- [00:36:13.840]it is, but we can still provide that guide. So, that's one thing I'm really interested too is can
- [00:36:17.760]we dig more in that soil biology and then perhaps down the road match up that biological product
- [00:36:24.000]with that soil biology. Are there things that we can do to match it to work together.
- [00:36:27.280]Jackson: Yeah, it does seem like that mineralization question is kind of the whether
- [00:36:31.760]you're dealing with sensors or models. That's really like the what we're trying to answer is
- [00:36:36.000]where is that other source coming from that we can manage to cut back our main
- [00:36:40.160]application with. So it's a tough question. Bob: You know, I think there's you know as
- [00:36:46.000]we talked about sensors and other technology I think those are still
- [00:36:48.880]important to consider too the question that in my mind is scalability. You know there's technology
- [00:36:55.760]that comes out that says oh yeah we can put this in your field and test it and we can measure
- [00:37:00.320]nitrogen and phosphorus potassium and information like that at that point, okay that's great but how
- [00:37:05.840]useful is that when I have two thousand, three thousand five thousand acres you know 30-40-50
- [00:37:11.200]fields and what happens when I drive over with a sprayer, what happens when I bury it with tracker.
- [00:37:19.920]So, I like that technology don't get me wrong I think those are great technology
- [00:37:24.560]to use for modeling purposes, so something like in our role here I would be all over that to study if
- [00:37:31.360]it can give me a good estimation of mineralization or water capacity or different things like that.
- [00:37:39.920]If that can tell me that that's great, but I'm going to probably use that data to build a model
- [00:37:43.920]because that allows me to expand over more acres, and I might have something that gives me a proxy
- [00:37:48.240]to help me understand what those values you know where I need to be up there. So,
- [00:37:53.680]again we're not technology adverse by any means we just want to make sure we're using the right
- [00:37:57.040]technology that's going to expand and scale across many acres and makes it easier for growers.
- [00:38:02.640]We don't, one thing we need to always consider when we look at all this technology is
- [00:38:06.560]what benefit do we think it brings to us but also what is the hassle factor.
- [00:38:10.800]Jackson: So, if our listeners are really you know peaked as far as models and they really
- [00:38:16.720]want to go learn more about what Granular is doing and what they're offering where would you
- [00:38:21.840]send them to learn more information? Bob: Well, definitely you know first thing will
- [00:38:25.120]be our website granular.ag. We have great location way there also hit up any of your local Pioneer
- [00:38:31.360]representatives, they could connect you with a local CSA of course if you know the CSA in your
- [00:38:37.280]region absolutely visit them as well. We have you know varying densities different numbers of CSAs,
- [00:38:42.960]so they may not be in your region per se but by connecting with a Pioneer representative
- [00:38:47.600]they can connect you with our digital business managers who cover the regions, and they can
- [00:38:52.320]connect you with the CSA or even assist you there as well. One thing that I'm really excited about
- [00:38:57.520]too is we offer what's called pro services so we may have a situation where a farmer
- [00:39:01.680]may be an area we do not have a CSA yet and but they're of large enough size to say you know we
- [00:39:07.760]want to try this out we have folks on our team that are agronomists have a background in farming
- [00:39:13.120]come from farms themselves usually and and for a fee we can set that up and basically be a
- [00:39:18.480]in a way a CSA for them right it's not to replace a CSA totally. In fact,
- [00:39:23.520]they also assist CSAs and some of the other day-to-day work but yeah there's a situation
- [00:39:28.720]where farmer says I really want this, but I am in a place where there's I don't know any CSAs.
- [00:39:33.120]We can still work with them through that pro services and so again talking with you
- [00:39:36.400]know checking out with our greater website talk with our local Pioneer reps. Connect with our
- [00:39:41.120]digital business managers we can get them set up. Sam: Awesome, so as we wrap things up can you
- [00:39:47.200]offer one piece of advice for listeners who are looking to
- [00:39:50.480]improve their nitrogen management practices? Bob: Documentation, just record what you're doing
- [00:39:56.320]if you've got any of the the popular controllers systems in your tractors your equipment today
- [00:40:01.440]document what you're doing. You know, capture as much you can around you know the product
- [00:40:06.720]you're using the rates that capture the dates of course that information may seem kind of boring
- [00:40:12.080]right now, but it could be useful down the road. Calibrate your yield monitors make sure
- [00:40:16.000]they're all set properly again maybe you're not doing a lot with them maybe you're making pretty
- [00:40:19.440]maps today but you don't know down the road that could be useful as well. I think one thing I would
- [00:40:23.920]always encourage farmers to do, and I know this is always a headache and back to that hassle factor
- [00:40:28.560]do trials set aside part of your farm do something crazy so and maybe not too crazy but do something
- [00:40:36.400]that's that's a little out of the box and don't do a whole lot of it just try it out
- [00:40:39.920]so do your own on the farm trials, different nation rates, record what you do where you do it
- [00:40:45.600]check it again with the yield monitor in the fall. Try out those biologicals, try out the stabilizers
- [00:40:51.360]try out those things see how they work on your farms and those are things that's going to get
- [00:40:55.520]you about down the road there better. Having variable rate equipment is useful we can work
- [00:41:03.600]with you if you have just flat rate I mean it's possible we could probably pick out a rate that
- [00:41:07.040]would probably suit you across the field, but you know moving to that new technology of variable
- [00:41:12.320]rate controls for your sprayers or applicators is going to get you a long ways down the road too.
- [00:41:21.120]so thank you so much to bob for joining us today on the podcast to talk about nitrogen management
- [00:41:26.080]yeah it was pretty cool having bob on it's been a while since we had a chance to see him this was
- [00:41:30.080]nue conference last year i thought it was a really good interview one of the best ones we've had
- [00:41:35.680]i really like that he focused on the slow process that they're taking to make
- [00:41:38.560]sure the model is correct in every new region that they're going into
- [00:41:41.920]um and then he's focused on getting the models right you know he wants to know where they're
- [00:41:45.360]wrong so they can improve them i think it's just a really good approach yeah and using quality data
- [00:41:49.840]with field trials but not just all the data and so i thought yeah that was a good point
- [00:41:54.240]i also like their local expertise using csas right and so it's not just let's put
- [00:41:59.680]this model out there everywhere and see what happens it's let's use someone that you're
- [00:42:04.640]comfortable with or familiar with to help you make it work for your farm exactly that was really
- [00:42:09.120]interesting yeah really interesting episode and next week we're looking forward to getting into
- [00:42:13.520]responsive nitrogen management which will be a new theme within this nitrogen management series
- [00:42:18.960]so i hope you'll tune in for that thank you again and we'll see you next week on farm beds
- [00:42:23.760]thank you for taking the time to join us today on the farmbits podcast if you enjoyed this episode
- [00:42:29.200]please subscribe to the podcast on spotify apple podcast youtube or wherever you listen to podcasts
- [00:42:35.760]to be informed about the latest content each week we welcome your feedback so if you have comments
- [00:42:41.120]or questions for us please reach out to us over email on twitter or in the reviews section of
- [00:42:46.400]your favorite podcast platform our contact information can be found in the show notes
- [00:42:51.440]we'd like to thank nebraska extension for their support of this podcast and their commitment to
- [00:42:56.000]providing high quality informational material to members of the agricultural community
- [00:43:00.720]in nebraska and beyond the opinions expressed by the hosts and guests on this podcast are solely
- [00:43:06.800]their own and do not reflect reviews of nebraska extension or the university of nebraska-lincoln
- [00:43:12.480]we look forward to joining us next week for another episode of farm bits
- [00:43:30.960]you
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.
Log in to post comments
Embed
Copy the following code into your page
HTML
<div style="height: 5.62em; max-width: 56.12rem; overflow: hidden; position:relative; -webkit-box-flex: 1; flex-grow: 1;">
<iframe
style="bottom: 0; left: 0; position: absolute; right: 0; top: 0; border: 0; height: 100%; width: 100%;"
src="https://mediahub.unl.edu/media/17042?format=iframe&autoplay=0"
title="Audio Player: Episode 034: Granular Guidance for Nitrogen Management "
allowfullscreen
></iframe>
</div>
Comments
0 Comments