Scaling On-Farm Research in Image-Based Fertigation with Customer-Driven Development

Jackson Stansell, Founder & CEO, Sentinel Fertigation Author

10/09/2023 Added

16 Plays

Description

Translating research to a commercial product is challenging. Doing so without involving customer-driven development, even during the research phase, is nearly impossible. This presentation will focus on the impact of customer-driven development of Sentinel Fertigation’s N-Time software for image-based fertigation and implications for research in agriculture and natural resources.

Searchable Transcript

Search:

[00:00:00.750]The following presentation
[00:00:02.220]is part of the Agronomy and Horticulture Seminar Series
[00:00:05.790]at the University of Nebraska Lincoln.
[00:00:08.940]Well, good afternoon everyone.
[00:00:10.140]Thank you for joining today, the Agronomy seminar.
[00:00:13.710]Welcome to everyone in the room,
[00:00:14.997]and for those online,
[00:00:16.800]I'm Guillermo (indistinct),
[00:00:18.480]have a pleasure to introduce yourself today
[00:00:21.360]to Jackson Stansell.
[00:00:22.500]He's the CEO and Founder of Sentinel Fertigation.
[00:00:27.030]Jackson is a UNL alumni.
[00:00:29.760]He's originally from Alabama.
[00:00:32.040]He did a bachelor's in engineering science
[00:00:38.668]at Harvard.
[00:00:40.110]And then in between, he did an internship here at UNL
[00:00:43.950]and then he came back after as a master's student with UNL.
[00:00:48.840]After finishing his master program,
[00:00:50.760]where he was working on using imagery
[00:00:54.499]to deliver better management options for farmers
[00:00:58.770]to manage water and nitrogen,
[00:01:01.080]and he founded Sentinel Fertigation
[00:01:05.010]that has been in the market for two years,
[00:01:07.200]and is we are covering more than seven states
[00:01:10.470]and 25 southern acres currently.
[00:01:15.540]So today we're hearing from him about
[00:01:18.930]how to scale on-farm research in image-based fertigation
[00:01:22.740]with customer-driven development.
[00:01:24.360]So the floor is yours, Jackson.
[00:01:26.100]Thank you for joining us today.
[00:01:28.200]Thanks Guillermo
[00:01:30.151]and pleasure to see all of you today.
[00:01:32.430]I know I see some familiar faces out there in the crowd
[00:01:35.010]and folks I've worked with before
[00:01:36.360]and folks have had the pleasure to meet.
[00:01:37.530]So I'm really happy to be with all of you here today.
[00:01:42.060]Again, talking about scaling on-farm research
[00:01:44.370]in image-based fertigation with customer-driven development.
[00:01:47.160]So just as a little bit of a disclaimer here,
[00:01:49.860]this is an agronomy and horticulture seminar,
[00:01:51.630]but I'm really gonna be talking a lot more about,
[00:01:54.360]kind of the iteration process that we went through
[00:01:56.820]engaging with farmers
[00:01:58.320]and how that really has impacted
[00:02:00.240]how we've gone about building Sentinel Fertigation
[00:02:02.367]and getting our image-based fertigation into the market.
[00:02:05.610]So there won't be a ton of agronomy information in here.
[00:02:07.470]I'm happy to answer those questions as well
[00:02:09.000]as we get to the end,
[00:02:10.290]but hopefully, this is still valuable.
[00:02:13.290]So I'll start off
[00:02:14.250]with talking a little bit about my journey.
[00:02:18.870]And my journey, as Guillermo mentioned,
[00:02:21.420]started in southeast Alabama, Dothan High School.
[00:02:24.930]Dothan, Alabama
[00:02:25.800]is actually the self-proclaimed peanut capital of the world.
[00:02:28.110]So roughly 50% of the nation's peanut crop
[00:02:31.110]comes from within a 100 miles of the town of Dothan.
[00:02:33.330]So grew up around peanut and cotton agriculture.
[00:02:35.670]I'm not from a farm myself though,
[00:02:37.381]just was exposed as a high school student,
[00:02:39.900]was recruited to play football at Harvard
[00:02:42.180]and that's what took me to the northeast,
[00:02:44.049]out of southeast Alabama.
[00:02:46.167]Spent the first few years
[00:02:48.540]in biochemical materials engineering,
[00:02:50.310]thinking I was going to get into biofuels.
[00:02:52.590]Spent a summer at Iowa State
[00:02:54.330]exploring my interest in biofuels,
[00:02:56.160]and decided I really did not enjoy
[00:02:58.140]being cooped up in a lab all day
[00:03:00.267]and went back to Harvard,
[00:03:02.670]started exploring opportunities
[00:03:04.200]to get involved with production agriculture,
[00:03:06.450]which I had continued to be around a little bit there
[00:03:08.490]in aims for that summer.
[00:03:09.870]And ultimately, ended up having the opportunity
[00:03:12.300]to come out here to the University of Nebraska
[00:03:14.850]on an internship in the Applied Plant Systems Program
[00:03:18.480]that allowed me to actually work here
[00:03:20.370]in the Department of Agronomy and Horticulture
[00:03:22.320]an on-farm research trial.
[00:03:24.420]And that ultimately convinced me that production agriculture
[00:03:27.660]and specifically, precision agriculture
[00:03:29.460]is where I wanted to spend my time
[00:03:31.500]and what I wanted to do with my life's work.
[00:03:33.210]So went back to campus,
[00:03:35.130]changed my major to a general engineering sciences degree
[00:03:38.640]with a focus in bio environmental engineering,
[00:03:41.100]and ultimately decided to come back out to Lincoln
[00:03:44.520]to pursue a master's in agricultural
[00:03:46.110]and biological systems engineering
[00:03:48.065]at the University of Nebraska Lincoln
[00:03:50.550]that turned into a PhD,
[00:03:51.870]and that PhD has since gone on pause
[00:03:53.760]as I've started working on Sentinel Fertigation
[00:03:56.340]and building that company.
[00:03:58.014]So I'd like to start off with just a big thank you actually
[00:04:00.900]to the agronomy and horticulture department
[00:04:02.730]for being what brought me out here
[00:04:05.700]to the University of Nebraska
[00:04:07.230]and has since been the springboard for this opportunity
[00:04:10.050]to build Sentinel Fertigation.
[00:04:13.290]There are a ton of mentors, advisors, and friends
[00:04:15.579]that I have here in the department.
[00:04:17.580]Dr. Martha Mamo was overseeing the program
[00:04:20.079]that brought me out here.
[00:04:21.150]Laura Thompson was my mentor,
[00:04:22.620]as well as Dr. Richard Ferguson.
[00:04:24.450]I've worked closely with Dr. Lila Puntel.
[00:04:26.610]Dr. Jim Schepers was a big part of welcoming me here
[00:04:29.915]in Nebraska when I first arrived that summer.
[00:04:32.850]And I've had an opportunity to work with Dr. Brian Krienke,
[00:04:35.010]Joel Crowther, John Parrish, Leo Bastos,
[00:04:37.080]Sal Ramirez, Jose Cesario,
[00:04:39.051]and so many others in the department.
[00:04:41.400]And I just wanna say thank you for helping to support
[00:04:45.090]and being such great advisors and friends in that journey.
[00:04:48.690]So today I'm focusing on Sentinel Fertigation.
[00:04:52.740]As Guillermo mentioned,
[00:04:53.640]I'm the founder and CEO.
[00:04:56.340]And our story kind of dates back
[00:04:59.220]to that internship experience
[00:05:00.900]where I was first exposed to the nitrogen problem
[00:05:04.080]here in Nebraska
[00:05:06.090]and also that we're facing nationally and globally.
[00:05:10.110]Nitrogen is incredibly difficult to manage
[00:05:13.110]because it is so complex.
[00:05:14.670]There are a ton of different ways
[00:05:15.930]by which nitrogen becomes available to plants.
[00:05:18.750]They're very hard to predict.
[00:05:20.490]They're very hard to measure.
[00:05:22.890]And doing nitrogen management well
[00:05:25.530]is really critical to profits
[00:05:26.910]as well as environmental stewardship.
[00:05:28.590]So I saw the nitrogen problem
[00:05:30.240]as a problem that was worth solving
[00:05:32.130]and really, it's the problem
[00:05:33.240]that continued to draw me out to Nebraska.
[00:05:37.290]When I got here,
[00:05:38.130]we started on a grant program that was looking at
[00:05:40.950]how we could use multi-spectral imagery
[00:05:43.290]at that time captured from drones
[00:05:44.910]to schedule nitrogen applications via fertigation.
[00:05:48.150]And at that point, we were doing that on a weekly basis.
[00:05:50.970]And so in 2019, we started working on
[00:05:54.240]on-farm research trials to prove out this technology
[00:05:57.660]and ultimately, developed a framework that we could use
[00:06:00.540]to schedule those nitrogen applications.
[00:06:03.000]I'll talk a little bit more about results later,
[00:06:05.280]but in general,
[00:06:06.690]those on-farm research trials showed very promising results,
[00:06:10.007]improvements in nitrogen use efficiency of 25%,
[00:06:13.410]savings to farmers
[00:06:14.610]of greater than 40 pounds of nitrogen to the acre.
[00:06:17.340]And with those results,
[00:06:18.360]we ultimately decided we couldn't let this technology die
[00:06:21.120]with a grant program.
[00:06:22.440]And we wanted to make sure
[00:06:23.580]that it was able to get out commercially
[00:06:25.800]and scale up and impact farmers here in Nebraska and beyond.
[00:06:29.760]And so now, that's what Sentinel is doing,
[00:06:32.610]and we're very excited about being on that mission
[00:06:34.890]to drive win-win solutions for both farm profitability
[00:06:38.970]and environmental stewardship.
[00:06:41.640]Some of the Sentinel team that's not here today
[00:06:45.197]is here on the board.
[00:06:46.860]We have a ton of folks here in Nebraska.
[00:06:48.690]We've got a great board with Stephan and Dan on it,
[00:06:51.600]both of whom have have great industry experience.
[00:06:53.492]And I'm really,
[00:06:55.170]I just want to mention all of these folks
[00:06:57.360]because they're really what makes Sentinel tick
[00:06:59.400]and has allowed us to be successful
[00:07:00.930]over the past couple of years.
[00:07:03.870]What Sentinel offers is N-Time,
[00:07:06.297]and this is what I'm basically gonna be talking about
[00:07:08.340]the rest of the talk today is,
[00:07:10.110]how we developed N-Time and what has gone into it since.
[00:07:13.230]What N-Time does is it ingests satellite imagery
[00:07:15.870]from two different sources.
[00:07:17.280]We use both Planet's PlanetScope dataset,
[00:07:19.560]which comes in at three meters per pixel
[00:07:21.600]and allows us to compute
[00:07:22.710]the normalized difference Red Edge Index,
[00:07:24.480]which is a vegetation index
[00:07:25.770]that is highly correlated to crop nitrogen status
[00:07:28.530]and changes in photosynthetic rate.
[00:07:30.690]We also ingest Airbus' Pléiades Neo constellation data,
[00:07:34.830]which is 30 centimeter per pixel,
[00:07:36.930]pan sharpened multispectral imagery
[00:07:39.607]that also allows us to compute
[00:07:41.760]the normalized difference Red Edge Index.
[00:07:44.070]And this is really, really impressive image quality
[00:07:47.250]because you're essentially getting
[00:07:48.660]close to UAV level resolution from space.
[00:07:52.020]And it's been a really great data set for us to use.
[00:07:55.410]From this imagery, we compute analytics.
[00:07:57.883]These analytics are based
[00:07:59.610]on our paired plot calibration technique,
[00:08:01.320]which I'll go into a little bit more here in a minute.
[00:08:04.530]But what these analytics allow us to do
[00:08:06.300]are isolate the impact of nitrogen from other stress factors
[00:08:09.210]that may limit yield potential in the field.
[00:08:11.323]They also allow us to quantify
[00:08:12.840]an nitrogen sufficiency index,
[00:08:14.340]which is a modified version of some of the work
[00:08:16.740]that's previously come out of UNL.
[00:08:19.211]We were able to classify crop nitrogen status
[00:08:21.810]and ultimately provide a seven day crop need outlook
[00:08:24.134]using the analytics.
[00:08:26.880]The value that farmer ultimately gets is a recommendation
[00:08:30.840]to either apply
[00:08:31.740]or not apply additional nitrogen fertilizer to the field.
[00:08:34.800]And that can be as simple as a text message
[00:08:36.510]or a push notification that just simply tells them
[00:08:38.520]you need to apply an additional 30 pounds to field A
[00:08:41.730]or it can be as complex as a prescription
[00:08:43.719]that provides different rates
[00:08:45.990]for different places in the field
[00:08:47.490]to respond to that crop's need site specifically.
[00:08:52.740]So N-Time originally came out of technology development
[00:08:56.820]in my master's program at UNL
[00:08:59.010]and that was centered on-farm research
[00:09:01.770]that resulted in customer-informed innovation for N-Time.
[00:09:06.210]So this technology originated honestly
[00:09:09.120]with a lot of the work that was done at UNL
[00:09:11.700]and other places around the country
[00:09:13.350]back in the late '80s, early '90s and into the mid '90s.
[00:09:16.770]And one of the papers that I read
[00:09:19.590]as I was getting started on this project
[00:09:21.210]was a paper by Dr. Schepers and Tracy Blackmer
[00:09:24.330]from back in 1995
[00:09:26.280]where they were using a chlorophyll meter
[00:09:28.800]to essentially schedule fertigation applications in corn.
[00:09:34.401]And so this was some of the early work that was done
[00:09:36.900]to use some of these sensor-based methods
[00:09:40.410]for scheduling fertigation.
[00:09:42.620]At that time, handheld chlorophyll meters
[00:09:44.790]were not exactly the most scalable solution
[00:09:46.620]to get data across the entire field,
[00:09:49.170]but there was very good efficacy shown
[00:09:51.630]and potential shown that this could be a great solution
[00:09:53.940]for scheduling fertigation long-term.
[00:09:57.540]After that, there were several decades that went by.
[00:10:01.100]UNL started engaging in plot scale research
[00:10:03.450]that was using essentially drone imagery
[00:10:05.483]and sensor readings to schedule fertigation events
[00:10:09.000]that showed that $30 per acre could be saved
[00:10:11.910]using these sensor-based fertigation scheduling techniques.
[00:10:15.390]Project SENSE was ongoing in the middle of the 2010s,
[00:10:18.907]from 2015 to 2017,
[00:10:21.030]that was showing that boom-mounted sensors
[00:10:22.800]for late vegetative sidedress could improve NUE by 18%
[00:10:26.410]and increased profits by $13 per acre.
[00:10:29.730]And there was also some work going on at Clemson
[00:10:31.800]in the late 2010s that was looking at using NDVI
[00:10:35.130]to schedule cotton fertigation and improving NUE by 74%.
[00:10:39.860]So that was a pretty significant savings,
[00:10:42.000]but that was some of the first work
[00:10:43.320]that was actually published
[00:10:44.640]that showed that you could use images
[00:10:47.070]that were captured from UAVs, from other aerial vehicles,
[00:10:49.800]or even from satellites
[00:10:51.540]to schedule fertigation applications.
[00:10:56.100]So with that, we had this, you know,
[00:10:59.370]literature background, this technology background,
[00:11:01.830]and we were also facing a pretty significant issue
[00:11:05.370]here in Nebraska and throughout the rest of the country,
[00:11:07.650]which is that in irrigated crop production systems,
[00:11:10.830]there's a really significant probability of nitrate leaching
[00:11:13.950]and nitrate contamination in the groundwater.
[00:11:16.650]And so, facing this particular concern,
[00:11:20.040]we sought out to develop
[00:11:21.210]our image-based fertigation scheduling framework.
[00:11:24.510]The idea here is that we would collect imagery
[00:11:27.450]at a high frequency throughout the growing season.
[00:11:30.270]We would calibrate each image
[00:11:31.980]with what we have now called Sentinel plots.
[00:11:34.790]At that time, they were just indicator plots
[00:11:37.380]that we were using in the field,
[00:11:39.000]which consisted of multiple different nitrogen rates.
[00:11:41.010]They were essentially an embedded nitrogen rate trial.
[00:11:42.990]We would calibrate each image with those
[00:11:45.180]and then perform an analysis
[00:11:46.740]and apply a recommendation algorithm
[00:11:48.270]that would say whether or not
[00:11:49.380]additional nitrogen fertilizer needed to be applied
[00:11:51.810]and then generate a site-specific prescription
[00:11:54.480]with the correct rate to apply
[00:11:55.620]in different locations in the field,
[00:11:57.120]and then ultimately,
[00:11:57.953]deliver that recommendation to the farmer.
[00:12:01.050]The whole idea here, which I'll kind of,
[00:12:02.940]if I can figure out the laser pointer, good.
[00:12:05.520]The whole idea here was to essentially maintain
[00:12:09.210]an ideal set point.
[00:12:12.120]As the crop nitrogen status fluctuates,
[00:12:14.580]we wanna maintain crop nitrogen status
[00:12:16.290]in an ideal set point,
[00:12:17.730]and then we're gonna use fertigation,
[00:12:19.620]turn fertigation on and off
[00:12:21.570]to trigger that crop's positive nitrogen status response,
[00:12:24.930]right?
[00:12:25.763]So this is a constant monitoring and feedback control system
[00:12:28.410]where we're constantly gauging
[00:12:29.460]that nitrogen status of the crop
[00:12:30.960]and then recommending an intervention,
[00:12:32.400]bringing that nitrogen status up to the set point,
[00:12:34.890]allowing the plant to use that nitrogen
[00:12:37.200]and any other sort of natural processes.
[00:12:39.360]And as that sufficiency starts to dip again,
[00:12:41.400]then we turn on fertigation and allow it to intervene
[00:12:44.310]and keep the crop at a solid nitrogen status
[00:12:46.933]for yield potential.
[00:12:51.150]Core to this technology
[00:12:52.400]is our paired-plot calibration technique.
[00:12:55.110]For those of you in the room
[00:12:55.950]that are familiar with sensor-based nitrogen management,
[00:12:58.687]these concepts are not exactly new.
[00:13:00.991]But what we did is we actually combined a concept
[00:13:05.400]that came out of UNL,
[00:13:06.360]which was the high nitrogen reference strip,
[00:13:08.640]and combined that with a little bit of a response index
[00:13:11.190]that came out of Oklahoma State,
[00:13:12.840]where there was essentially
[00:13:13.830]a low nitrogen strip that was used
[00:13:15.930]and compared to the rest of the crop
[00:13:17.280]to see what the crop's response to nitrogen was.
[00:13:19.800]And here at UNL,
[00:13:20.910]that reference strip was being used to say,
[00:13:24.210]if the crop had additional nitrogen,
[00:13:26.070]would there be a positive response
[00:13:27.720]to that additional nitrogen?
[00:13:29.370]What we decided to do was to pair
[00:13:31.800]these high and low nitrogen calibration strips,
[00:13:34.980]the high nitrogen strip being 60 pounds
[00:13:37.230]ahead of the rest of the crop,
[00:13:38.940]and the low nitrogen strip being 30 pounds
[00:13:40.650]behind the rest of the crop,
[00:13:42.960]and close proximity
[00:13:44.490]to capture the benefits of both different techniques.
[00:13:47.700]The low nitrogen strip would allow us to see
[00:13:50.460]if the crop had 30 pounds less nitrogen,
[00:13:52.740]would it truly be approaching a deficiency?
[00:13:54.900]So essentially,
[00:13:55.733]are we getting enough nitrogen from natural sources
[00:13:57.480]to overcome the limitation of synthetic nitrogen,
[00:14:01.410]right next to a strip that indicated whether or not
[00:14:04.830]the crop would respond to additional synthetic nitrogen
[00:14:07.410]and allow us to calibrate.
[00:14:08.910]The low nitrogen plot
[00:14:09.780]also gave us about a week look ahead time
[00:14:12.398]on crop nitrogen need, right?
[00:14:15.359]Because if you average out nitrogen flux
[00:14:17.340]over the course of the growing season,
[00:14:19.050]you're looking at about three to four pounds
[00:14:20.910]of nitrogen per acre per day.
[00:14:22.740]And if you multiply that by seven days,
[00:14:24.330]you're looking at roughly 30 pounds of nitrogen to the acre
[00:14:26.880]that a crop is gonna take up in a given week.
[00:14:28.950]So with a 30 pound negative offset,
[00:14:31.920]we're able to see, you know,
[00:14:33.510]how much nitrogen is that crop going to need
[00:14:35.540]in the next week
[00:14:36.720]and stay ahead of any imminent deficiencies.
[00:14:40.500]So we took this framework
[00:14:42.510]that was reliant on this paired-plot calibration technique
[00:14:45.540]out into on-farm research.
[00:14:47.670]While I was a graduate student,
[00:14:48.750]we ran 14 different on-farm research trials
[00:14:51.357]that were in the northeast part of the state,
[00:14:53.260]northeast, north central part of the state,
[00:14:55.830]the central part of the state,
[00:14:57.300]and we also had one trial
[00:14:59.010]in kind of the west central part of the state
[00:15:01.290]out in Sutherland.
[00:15:02.730]And the way that we laid out our on-farm research trials
[00:15:05.550]was using essentially a randomized complete plot design
[00:15:09.990]that consisted of 12 15 degree pie-shaped sectors,
[00:15:13.593]four different reps of three different treatments.
[00:15:17.921]And the three different treatments that we were looking at
[00:15:20.040]were a grower management treatment
[00:15:22.020]where the grower managed,
[00:15:23.700]followed their own management in their own sector,
[00:15:25.710]a risk tolerant treatment.
[00:15:27.570]So where we were essentially prioritizing nitrogen savings
[00:15:32.370]over yield potential
[00:15:33.540]by waiting until the absolute last minute
[00:15:35.610]to apply nitrogen via fertigation,
[00:15:37.620]and a risk-averse treatment
[00:15:39.450]where we were responding more quickly
[00:15:41.610]to indicated approaching deficiencies in the crop.
[00:15:45.000]And so in every trial, we had two sensor-based treatments,
[00:15:47.670]the risk tolerant, risk averse,
[00:15:49.230]and the one grower treatment.
[00:15:51.900]What we found across these research results
[00:15:55.560]is that our risk averse treatments
[00:15:57.780]when applied over the course of the entire growing season,
[00:16:00.060]so basically from V6 through R3
[00:16:02.760]is what we would consider
[00:16:03.630]the entire course of the growing season
[00:16:05.160]for fertigation management.
[00:16:06.861]In this case,
[00:16:08.304]we were able to improve efficiency by 25.5%
[00:16:13.211]versus what growers were doing already on their farms.
[00:16:16.763]And at a corn price of $5.58 per bushel
[00:16:20.280]and a nitrogen cost,
[00:16:21.360]which is high at a $1.6 per pound
[00:16:24.780]because the last time we ran this analysis
[00:16:26.190]was when we were at all time high fertilizers,
[00:16:28.980]we're able to increase profits by $27.91 per acre.
[00:16:33.180]So those results were pretty outstanding
[00:16:35.760]because it showed that
[00:16:36.720]we could improve environmental stewardship
[00:16:38.310]by improving the efficiency with which nitrogen was applied
[00:16:41.550]while also improving farm profitability.
[00:16:45.750]So ultimately, the outcomes of this development work
[00:16:48.510]is that we developed a patent-pending technology
[00:16:50.880]that consisted of both the calibration technique
[00:16:52.830]that used these embedded paired-plots,
[00:16:54.808]as well as this real-time nitrogen monitoring recommendation
[00:16:57.780]and intervention framework
[00:16:59.368]for automating fertigation recommendations.
[00:17:03.333]I also developed the N-Time desktop software in MATLAB,
[00:17:06.981]which is a prototype application that we ultimately used
[00:17:10.470]to develop the web application that Sentinel now offers.
[00:17:14.430]We compiled all of this on-farm research data
[00:17:17.130]that validated the technology
[00:17:19.020]and then also built an initial customer network.
[00:17:21.540]We worked with about 10 different farmers
[00:17:24.270]over the course of these on-farm research trials.
[00:17:26.758]They involved their agronomists
[00:17:28.680]in much of the management that we did,
[00:17:30.990]and we also worked with different NRDs around the state.
[00:17:33.540]Central Platte NRD was one that was exceptionally involved
[00:17:37.146]in this process of investigating image-based fertigation.
[00:17:41.070]And this network was absolutely critical
[00:17:43.530]to getting the technology
[00:17:45.660]out of the academic research domain
[00:17:47.760]and into the commercial domain.
[00:17:51.540]So now we'll talk a little bit about commercializing N-Time,
[00:17:54.257]and what I wanna focus on here is
[00:17:56.088]how customers are the essence of Sentinel's growth
[00:17:58.830]and success,
[00:17:59.993]and ultimately, pushing N-Time forward.
[00:18:04.680]N-Time had several customer-identified problems
[00:18:08.610]as a research solution.
[00:18:12.120]The first one was that UAV-based imagery
[00:18:14.310]does not work for monitoring at scale.
[00:18:17.220]The reason for this is that it's operationally difficult.
[00:18:20.580]So when I say operationally difficult,
[00:18:22.830]what I mean here is,
[00:18:24.390]wind often impacts drones
[00:18:28.020]and how well they're able to fly
[00:18:29.310]and how consistently they're able to capture imagery.
[00:18:32.108]There's an immense amount of data
[00:18:34.170]that's generated in a single UAV flight
[00:18:36.300]for even a typical quarter section.
[00:18:38.700]Dealing with wireless connectivity in rural parts of America
[00:18:42.270]makes that really hard to get that data to the cloud
[00:18:44.970]where it needs to be processed
[00:18:46.140]or you have to have a really robust computing station
[00:18:49.290]that you're moving with mobily
[00:18:51.390]and processing in between fields.
[00:18:53.730]On top of that, it's very labor-intensive.
[00:18:55.590]So right now under FAA regulations,
[00:18:57.990]there's really no way to do nighttime flight.
[00:19:00.900]There's no way to do non-line of site flight
[00:19:03.120]without many waivers from the FAA.
[00:19:06.060]And so you have to have an individual
[00:19:08.520]who's carrying this drone around to different fields
[00:19:10.650]who's able to maintain line of sight.
[00:19:12.330]And if you're wanting to cover hundreds of fields in a day,
[00:19:15.300]you're gonna have to employ quite a fleet of people
[00:19:17.340]to make sure that that gets done.
[00:19:18.870]And ultimately, if you're employing that many people,
[00:19:20.880]you're paying for the insurance in all of these drones,
[00:19:23.160]you're paying for the data processing capabilities,
[00:19:24.960]it gets very cost-prohibitive.
[00:19:27.150]We were flying UAVs at research scale
[00:19:29.610]and it was really hard to even get six fields done
[00:19:31.650]in a single day,
[00:19:32.550]get them processed
[00:19:33.810]and get recommendations turned around in 24 hours.
[00:19:36.570]So this was a huge problem that had to be solved.
[00:19:39.750]During research, we were also using very small rate blocks,
[00:19:42.750]and this was because we had to fit these rate blocks,
[00:19:45.045]so these Sentinel plots that we use for calibration
[00:19:47.880]into those treatment areas, right?
[00:19:49.976]And we were trying to include four different rates in those.
[00:19:53.430]And what we found is that they were,
[00:19:55.680]they were effective for research,
[00:19:57.240]but they're horribly ineffective for scaling this technology
[00:20:00.270]with the amount of processing that you have to do
[00:20:01.950]to make sure
[00:20:02.850]that you're looking at the right geospatial locations
[00:20:05.220]that actually achieved the appropriate rates.
[00:20:08.040]What we found
[00:20:08.970]was that we often had inaccurate establishment,
[00:20:11.280]and I don't mean that we'd never achieved the right rate,
[00:20:13.500]but I meant that the right rate
[00:20:15.180]was not achieved at the right place.
[00:20:16.830]And so we actually had to manually look at as applied data,
[00:20:19.710]move the geospatial locations to the areas of the field
[00:20:22.080]that actually got the right rate received
[00:20:24.270]so that we could execute the research.
[00:20:26.010]And ultimately, that made it very operationally demanding,
[00:20:28.440]both for the operator of the machine itself
[00:20:30.780]as well as for the people that were processing the data.
[00:20:33.090]There was a ton of time
[00:20:33.960]that went into establishing these plots.
[00:20:36.270]And then there were communication challenges.
[00:20:38.760]If you had a farmer that was applying these
[00:20:40.170]with their own equipment,
[00:20:41.130]did they know the right swath width of the equipment?
[00:20:43.860]Did they know the heading and the AB line
[00:20:45.153]that that equipment was gonna be traveling in the field
[00:20:47.340]so that we could properly position these plots?
[00:20:50.220]So we realized that we needed to make bigger rate blocks
[00:20:53.610]and make these in a way
[00:20:54.443]that could be more operationally integrated
[00:20:56.737]into farms at scale.
[00:20:59.220]Finally, we were using variable rate pumps
[00:21:01.410]to do all of this research
[00:21:04.627]and farmers repetitively said,
[00:21:06.960]these variable rate pumps are too expensive
[00:21:09.090]for me to pencil out in my operation right now
[00:21:11.730]without demonstrating a significant amount of ROI
[00:21:13.980]on these pumps.
[00:21:14.813]So we had to make sure that the technology was positioned
[00:21:18.270]in such a way that it could work
[00:21:19.890]without a variable rate pump being in the field.
[00:21:24.270]So some of the pivots that we made to the technology
[00:21:26.940]in order to launch the technology at scale
[00:21:29.910]is that we moved to satellite imagery only.
[00:21:32.520]Now the interesting thing about this is,
[00:21:34.830]so I founded the company two years ago yesterday.
[00:21:37.650]So we just passed our two year birthday.
[00:21:40.380]From September 21st, 2021
[00:21:42.900]until we launched on June 6th, 2022,
[00:21:45.780]we went about converting this desktop application
[00:21:48.270]into a web application
[00:21:49.680]that could be used by multiple users,
[00:21:51.840]different authentication,
[00:21:53.940]folks from all different geographies
[00:21:55.560]could access the system.
[00:21:57.510]Our plan was to make this transition
[00:22:00.960]to satellite imagery slowly, right?
[00:22:02.700]We were going to incorporate drone imagery
[00:22:04.440]in the first year.
[00:22:05.478]We were gonna be operating at limited scale.
[00:22:07.140]We had the money to pay for a drone provider,
[00:22:09.930]so we were gonna combine Planet's satellite imagery
[00:22:12.540]with a commercial drone provider.
[00:22:15.660]And we got halfway through the month of June,
[00:22:17.730]that commercial drone provider had not delivered any imagery
[00:22:21.000]seven days after they had taken all the initial flights
[00:22:23.760]across the 8,000 acres we were operating on.
[00:22:26.460]And I had to make a decision
[00:22:28.590]in the best interest of our customers
[00:22:30.570]to terminate that contract
[00:22:32.663]for that monitoring for the rest of the summer,
[00:22:35.250]pay out whatever we owed on that,
[00:22:36.900]and pick up Airbus' Pléiades Neo.
[00:22:39.330]So by force, we went completely satellite in the first year.
[00:22:43.950]The amazing thing about this is that Pléiades Neo
[00:22:46.320]was a brand new constellation.
[00:22:47.610]It was in its first growing season.
[00:22:49.308]And I think if this constellation hadn't been active
[00:22:53.250]because it was essentially the only tasking
[00:22:55.292]Red Edge satellite that provided the resolution
[00:22:57.510]that we needed to replicate drone imagery,
[00:23:01.380]we might've really, really struggled in that first year
[00:23:04.320]to make a successful transition to satellite imagery.
[00:23:08.280]But what Pléiades Neo did was,
[00:23:09.113]they were able to come in
[00:23:10.380]with 30 centimeter per pixel imagery, provide the NDRE,
[00:23:14.130]the Red Edge Band that they're using on these satellites,
[00:23:16.530]exactly mimicked what we were using in drone imagery.
[00:23:19.620]So the NDRE response was exactly the same
[00:23:22.200]and it plugged right into our algorithms in the first year.
[00:23:25.500]So that was a really key part to making sure
[00:23:28.620]that we could succeed at scale in 2022.
[00:23:32.040]We also introduced the concept of indicator slices.
[00:23:34.590]So in research, we had only done these indicator blocks
[00:23:38.977]that were established with non-fertigation, right?
[00:23:41.610]They were field machinery.
[00:23:42.630]They were in a rectangular shape or a square shape.
[00:23:45.420]And what we kept hearing from farmers was that,
[00:23:47.550]well, couldn't you just use the pivot
[00:23:49.320]to fertigate early in the season
[00:23:51.180]to establish these nitrogen rate blocks
[00:23:53.850]and could we just use pivot speed
[00:23:55.410]instead of a variable rate pump to do that?
[00:23:58.230]Thought, sure, we can do that.
[00:23:59.700]And so we actually started implementing
[00:24:01.350]these indicators slices,
[00:24:02.520]which were great
[00:24:03.720]because they took an operator outta the picture.
[00:24:05.610]We could essentially drop a prescription,
[00:24:07.470]a speed prescription into the pivot control software,
[00:24:10.380]tell it where to speed up and slow down.
[00:24:12.090]There was no human that was intervening in that process.
[00:24:14.250]And we had good enough GPS
[00:24:15.840]with the size of slices that we were making in the field
[00:24:19.470]to be confident in where those were placed.
[00:24:21.390]And as you can see in this image right here,
[00:24:23.850]we actually ran a speed prescription.
[00:24:25.080]This is from this year.
[00:24:26.190]You can see just ever so slightly where that canary,
[00:24:29.220]which is the red,
[00:24:30.360]and the reference, which is the green,
[00:24:31.830]are located in each part of the field.
[00:24:34.020]And so we know
[00:24:35.070]that we're getting these accurately established
[00:24:36.810]and we can do this geospatial geo-reference calibration
[00:24:40.770]very easily using these sorts of speed scripts, right?
[00:24:43.650]So this was another big pivot
[00:24:45.720]that we made to make this process scalable.
[00:24:48.840]We transitioned to full-field management.
[00:24:50.670]So not just making recommendations
[00:24:52.440]on a sector-by-sector basis,
[00:24:53.820]but saying for this field,
[00:24:55.740]we're gonna take into account
[00:24:56.910]the management zones that you have,
[00:24:58.560]the soil spatial variability,
[00:25:00.270]and essentially using a minimum allowed depletion concept
[00:25:03.240]from irrigation management.
[00:25:05.310]In your worst zone,
[00:25:07.140]it's showing that you need to apply nitrogen fertilizer.
[00:25:09.150]So we need to go ahead and recommend to the entire field
[00:25:11.975]that you need to apply nitrogen
[00:25:13.500]to make sure that that worst zone stays in good shape.
[00:25:18.810]We moved from having application logging
[00:25:21.180]being a mandatory part of the platform
[00:25:22.770]to being something that was optional
[00:25:24.690]because we realized that there really wasn't a huge impact
[00:25:27.120]to having that in our algorithm,
[00:25:29.640]'cause ultimately, it's almost entirely data-driven,
[00:25:33.360]and research grade data just is not necessary at scale.
[00:25:36.630]You know, it's something that you, kind of,
[00:25:37.800]get very accustomed to
[00:25:39.720]when you're working in a research domain,
[00:25:41.400]but having every single data point
[00:25:43.860]that you think you need for research, right,
[00:25:46.890]just simply is not practical
[00:25:48.570]once you start operating over 8,000, 20,000 acres.
[00:25:51.960]So that was another thing that we had to make sure
[00:25:54.570]that number one, I understood,
[00:25:56.190]but also our team understood,
[00:25:57.390]is that some of these situations are not gonna be perfect
[00:26:00.060]and we need to be operating in a robust way.
[00:26:04.350]So our N-Time desktop app,
[00:26:06.690]this is what it looked like two years ago.
[00:26:08.887]So I wrote this,
[00:26:10.110]it was a desktop app that was coded in MATLAB.
[00:26:13.580]It was really ugly (chuckling),
[00:26:15.180]really didn't look very good at all.
[00:26:16.800]And if I were a farmer,
[00:26:17.633]I never would've picked this up, right?
[00:26:19.530]We were generating these PDF reports
[00:26:21.240]that didn't have great ratios, great scale,
[00:26:24.840]but it had the information that you wanted to see
[00:26:27.810]as a grower.
[00:26:28.643]What was your recommendation?
[00:26:29.670]How much total fertilizer are you gonna need
[00:26:31.680]in this fertigation pass?
[00:26:33.180]How much time is it gonna take
[00:26:34.230]to complete this fertigation pass?
[00:26:36.030]And it adjusted the injection rate based on the depth
[00:26:38.880]that you were going to be using for the irrigation pass.
[00:26:41.430]And it ultimately process the imagery automatically.
[00:26:45.060]Over the course of eight months,
[00:26:47.070]we went from that to this web app
[00:26:50.460]that we launched on June 6th, 2022.
[00:26:53.310]And so this is what farmers picked up
[00:26:55.890]and used over 8,000 acres in the first year.
[00:26:58.530]We had our field dashboard,
[00:26:59.730]which showed the status of the field,
[00:27:02.070]so was a recommendation to apply nitrogen,
[00:27:04.852]what is your percentage of sufficiency in the field,
[00:27:08.670]provided insights
[00:27:09.870]that for every single paired Sentinel plot in the field,
[00:27:13.463]we were able to see what the NDRE values
[00:27:15.900]and the SI values were.
[00:27:17.528]It had prescriptive capabilities
[00:27:19.830]and it also had the ability to log nitrogen applications
[00:27:23.640]for farmers.
[00:27:26.070]With that software in 2022,
[00:27:28.980]we went from these research statistics,
[00:27:31.680]which you'll notice this per acre savings here
[00:27:34.140]is a little bit different than what I showed earlier
[00:27:35.550]because that changes no matter,
[00:27:37.590]basically alongside the corn
[00:27:39.420]and nitrogen price that you use.
[00:27:41.430]But what is consistent here
[00:27:42.540]is that 25% improvement in nitrogen use efficiency
[00:27:45.690]and the savings of 43 pounds of nitrogen to the acre
[00:27:47.820]from our research trials.
[00:27:49.920]If you look at our 2022 results,
[00:27:51.930]across the 42 sites that we got a complete data set for,
[00:27:55.530]we were able to replicate what we saw in on-farm research
[00:27:58.440]in these commercial trials versus a farmer baseline.
[00:28:01.200]We were able to improve nitrogen use efficiency by 23%
[00:28:04.350]and save farmers 42 pounds of nitrogen to the acre.
[00:28:08.310]To me, this was a very, very, very validating dataset
[00:28:11.460]to say that we're going to see consistent results
[00:28:13.500]that we had seen already on farm research
[00:28:15.990]in commercial scale applications.
[00:28:18.870]Last year, this made a really big impact
[00:28:20.880]because of the price of nitrogen, right?
[00:28:22.770]We were at about a dollar per pound of nitrogen.
[00:28:25.230]Our farmers on average were paying about 90 cents per pound
[00:28:28.650]of nitrogen that they used last year.
[00:28:30.567]And so that ultimately resulted
[00:28:31.920]in about $40 per acre in savings,
[00:28:34.200]which made a massive dent
[00:28:35.479]in improving their farm profitability.
[00:28:41.160]Then between 2022 and 2023,
[00:28:44.310]we actually had to rebuild the entire N-Time system
[00:28:48.390]from scratch.
[00:28:49.389]The firm that we had worked with here in Lincoln
[00:28:52.890]to develop the web application that we used in 2022
[00:28:56.310]chose a technology stack
[00:28:57.810]that was not well fit for geospatial data.
[00:29:00.930]It required a lot of custom coding
[00:29:03.859]of geospatial processing algorithms.
[00:29:06.000]It was all written in C#.
[00:29:07.710]And so what we decided to do
[00:29:09.120]as an engineering team last year
[00:29:11.070]was to totally gut the back end of the system
[00:29:13.801]and replace all of the code with Python,
[00:29:17.670]which I'm sure some of you in this room are using right now
[00:29:20.220]to do some of your data analysis.
[00:29:21.960]And you probably are aware that GeoPandas, Rasterio,
[00:29:24.720]Fiona Shapely,
[00:29:25.800]some of these libraries exist
[00:29:27.660]that already have built in open source functionality
[00:29:30.390]for geospatial processing, right?
[00:29:32.490]So we converted completely to Python on the back end
[00:29:35.280]and we actually realized
[00:29:36.180]we had to totally gut the front end code as well
[00:29:38.520]because the firm we had worked with had injected C#
[00:29:42.450]into all the JavaScript on the front end.
[00:29:44.430]So they had actually built some backend algorithms
[00:29:46.590]into the front end
[00:29:47.670]that had to be totally ripped out and removed and replaced.
[00:29:50.610]So our two software engineers worked together on this
[00:29:54.330]and over the course of the entire last off season,
[00:29:56.940]so from August of '22 through May of this year,
[00:30:00.840]they rebuilt N-Time and this is what it looks like today.
[00:30:03.720]And as you can see,
[00:30:04.770]we have some of our high resolution imagery
[00:30:06.930]from Airbus being shown here
[00:30:08.100]where you can see pivot tracks quite clearly.
[00:30:10.980]You can see where there were some,
[00:30:13.050]there was a little bit of disparity
[00:30:14.190]and where the planter shut off
[00:30:16.170]at the west side of the field.
[00:30:18.210]You can see kind of this gravity irrigated corner up here.
[00:30:22.290]You can see some of our moderate resolution imagery
[00:30:25.050]from Planet.
[00:30:26.250]You can see where we've introduced kind of a,
[00:30:27.943]a nitrogen tracking model,
[00:30:29.640]as well as a nitrogen applied model
[00:30:31.650]here on the insights tab and the status tab,
[00:30:34.491]where we're starting to do more NDRE histogram usage,
[00:30:38.876]then where we've improved our prescriptive capabilities
[00:30:42.840]for variable rate fertigation.
[00:30:44.640]So this is where things stand today.
[00:30:49.080]Adoption-wise,
[00:30:50.340]in 2022, we operated on 8,000 acres.
[00:30:52.710]We had 23 different farms that use the technology
[00:30:55.410]and we had three certified service provider partners.
[00:30:57.840]So these are dealers that are working with us to extend
[00:31:00.870]and offer this solution to their farmer customers.
[00:31:03.840]In 2023, we've operated on 18,030 acres.
[00:31:07.800]This has recently gone up 'cause we have an orchard grass,
[00:31:11.130]four orchard grass fields in Nevada
[00:31:12.750]that we've just picked up
[00:31:14.100]and also some small grains out in western Nebraska
[00:31:17.670]that we're starting to monitor.
[00:31:18.780]Yet this fall, we have 74 farms
[00:31:21.960]that are currently using the technology
[00:31:23.460]in 22 certified service provider or dealer partners
[00:31:27.333]that are located in four states.
[00:31:29.670]We're now operating in eight total states here in the US,
[00:31:32.637]including Illinois, Wisconsin, Kansas, Texas,
[00:31:35.850]Alabama, Tennessee, Nevada,
[00:31:37.740]and of course, Nebraska.
[00:31:40.470]What's been really interesting from 2022 to 2023
[00:31:43.860]is that one out of every five farmers that we returned
[00:31:46.770]have decided to use this technology
[00:31:48.360]across all of their irrigated acres,
[00:31:49.688]which is a really good sign
[00:31:51.930]for what they believe the technology is capable of.
[00:31:56.820]And we continue to build with farmer feedback.
[00:31:58.972]So this has been a really critical part
[00:32:01.020]of how we've developed things through on-farm research,
[00:32:03.540]working with collaborative farmers
[00:32:04.860]to determine what these problems are with the system,
[00:32:07.260]what's gonna limit their use of it.
[00:32:09.150]And this year, we've had trials throughout the US.
[00:32:13.710]We had 16 Precision Nitrogen Management trials
[00:32:16.542]with UNL Extension this year
[00:32:18.540]where they were evaluating this technology
[00:32:20.310]against grower practice
[00:32:21.780]or against both adapt and grower practice.
[00:32:24.780]I'm really excited to see what comes out of those trials
[00:32:27.294]this year.
[00:32:28.494]We have worked with the AGLaunch Accelerator
[00:32:31.470]out of Memphis, Tennessee
[00:32:32.820]to execute a corn and cotton trial in Tennessee.
[00:32:36.092]We did a rainfed corn trial with ag partners co-op
[00:32:40.440]down in northeastern Kansas this year
[00:32:43.470]where we were monitoring a rainfed field
[00:32:45.570]and determining whether or not
[00:32:47.070]they should make an aerial application of nitrogen
[00:32:49.320]around tassel.
[00:32:50.820]And we ultimately recommended that they hold off
[00:32:53.220]on that application,
[00:32:54.390]and so, we've saved them 30 pounds.
[00:32:56.940]We'll see what their yield response looks like there.
[00:32:59.790]We've had a cotton trial this year with the USDA
[00:33:02.460]in the Texas Panhandle at the crop product
[00:33:05.430]or Conservation Production Research Lab.
[00:33:07.710]And then we've also operated at N Rec
[00:33:09.750]with some of the USDA LATR sites up there
[00:33:12.303]that are comparing aspirational management
[00:33:14.640]versus current farmer management.
[00:33:18.150]Some of the barriers that we're currently facing
[00:33:20.250]to scalable success are addressable markets,
[00:33:23.010]technology delivery,
[00:33:24.420]and sales channel strategy.
[00:33:26.790]When we went into this,
[00:33:28.740]I was assessing our fertigation market that we could address
[00:33:32.100]as essentially the 1.2 million fertigated acres of corn
[00:33:35.460]that we have here in Nebraska.
[00:33:37.320]And then the additional,
[00:33:38.607]basically 2.2 million acres of fertigated corn
[00:33:43.500]that exists throughout the rest of the United States.
[00:33:46.020]That really leaves us with only three and a half million
[00:33:48.570]fertigated acres that we have to make a dent in
[00:33:51.330]and use as our addressable market.
[00:33:54.750]For our investors
[00:33:56.040]or people that were considering investing in us initially,
[00:33:59.100]this was much too small of a market for them to say
[00:34:02.100]it was a worthwhile investment to make.
[00:34:07.659]And to this point, we're starting to,
[00:34:10.050]I think already see
[00:34:11.670]that making a a big dent into that small market
[00:34:14.070]is gonna be a bit of a challenge, right?
[00:34:15.330]Because you're running into hesitancy
[00:34:17.730]and adoption from farmers.
[00:34:19.320]Not every farmer's gonna want to pick up
[00:34:20.820]and use a new technology to adjust,
[00:34:24.420]adjust their practices
[00:34:25.320]that they've been using for the past 20, 40, 50 years.
[00:34:29.970]And we're starting to identify some new opportunities
[00:34:32.820]to expand this marketability of our technology
[00:34:37.129]and hopefully, grow and scale the technology.
[00:34:40.470]The number one opportunity that we're seeing is that
[00:34:42.120]not just corn can benefit from this technology.
[00:34:44.790]And I'll talk about some of the crops
[00:34:46.470]that we operated on in 2023.
[00:34:49.110]Trial monitoring is starting to be something
[00:34:50.760]that some of our partners are seeing a value proposition in,
[00:34:54.030]which has been really, really interesting.
[00:34:55.350]So especially
[00:34:56.466]some of the biological solutions providers out there,
[00:34:59.820]think your pivot bios, your sound agriculture's,
[00:35:02.760]some of those folks are seeing this technology
[00:35:05.940]as a way to monitor
[00:35:07.320]how those biologicals are performing in real time
[00:35:09.840]and provide guidance on whether or not
[00:35:11.760]you need to intervene with synthetic nitrogen fertilizer
[00:35:14.310]before there actually is a yield reduction
[00:35:16.800]as a result of cutting back on nitrogen fertilizer.
[00:35:19.830]Foliar feeding is something that folks are looking at
[00:35:21.780]with this technology.
[00:35:23.280]We're starting to make some preliminary moves
[00:35:27.180]towards recommending nitrogen applications for sidedress
[00:35:30.600]and rainfed and irrigated operations.
[00:35:33.256]We've started to perform retroactive analysis.
[00:35:35.760]So for farmers that have conducted a nitrogen trial
[00:35:38.532]over the course of the growing season,
[00:35:40.110]they can actually enroll that field in the platform
[00:35:42.167]and we can go back and see crop response
[00:35:44.730]to in-season applications
[00:35:46.020]to see if they made an impact on that crop
[00:35:48.180]over the course of the growing season.
[00:35:50.220]The Department of Defense has actually indicated
[00:35:52.230]that there might be some applications
[00:35:53.490]for international monitoring of crops
[00:35:56.490]to identify where there might be some stress there.
[00:35:59.130]And then even marketing,
[00:36:00.600]our customers have come and said,
[00:36:01.680]well, you should rename those indicator blocks
[00:36:03.510]to Sentinel plots
[00:36:04.343]because then we'll actually know what you're talking about
[00:36:05.880]because nobody could pick up on indicator blocks.
[00:36:09.000]So our customers are the ones that are driving
[00:36:11.580]all of these potential value propositions.
[00:36:13.800]These are none of the things that we went into,
[00:36:16.235]into the company thinking this technology could do,
[00:36:18.990]and these are the things that customers
[00:36:20.700]are either forcing us to do or asking us to do.
[00:36:24.120]We've seen farmers use N-Time so far
[00:36:26.070]to optimize fertigation applications,
[00:36:27.930]to monitor regenerative practices,
[00:36:29.760]whether that be cover crops, biologicals,
[00:36:31.770]manure applications,
[00:36:33.208]switching to no-till, so on and so forth,
[00:36:36.030]reducing their risk exposure.
[00:36:38.980]So thinking about fertigation as a way,
[00:36:40.920]and especially data-driven fertigation
[00:36:42.630]as a way to hold back nitrogen
[00:36:45.180]in case of an adverse weather event
[00:36:46.800]such as a hailstorm, right?
[00:36:48.900]So pushing that risk potential back
[00:36:51.030]and their risk exposure back,
[00:36:52.560]and then also improving their operational efficiency.
[00:36:54.570]So especially for big operators who have 60 to 80 pivots,
[00:36:58.872]it's very helpful for them to know
[00:37:00.930]which pivots actually need their attention in a given week
[00:37:03.030]so that they can better allocate their time
[00:37:04.530]and their labor force.
[00:37:06.780]So I mentioned that currently
[00:37:08.370]N-Time is being used in other crops.
[00:37:10.620]And these are all crops that customers have basically said,
[00:37:14.220]we are going to enroll this field with you
[00:37:17.220]whether you like it or not,
[00:37:18.210]or whether you're ready for it or not,
[00:37:19.590]we want to monitor these crops.
[00:37:21.771]And so we've ultimately, you know,
[00:37:24.120]done this for free this year
[00:37:25.140]just because we don't know how the technology responds
[00:37:27.570]or we didn't in these other crops until this point,
[00:37:30.330]but what we've seen is that
[00:37:31.200]N-Time has had success this year in potatoes.
[00:37:33.480]So we have a field here
[00:37:35.160]that's split with potatoes on the west side
[00:37:37.110]and corn on the east side.
[00:37:40.050]I'm still waiting to hear exactly how harvest went,
[00:37:42.300]but all indications
[00:37:43.230]are that Wisconsin's had a bumper potato crop this year
[00:37:46.260]and we were able to cut back one fertilizer application
[00:37:49.200]on these fields.
[00:37:50.670]We've heard from other agronomists in the area
[00:37:52.110]that they were recommending a cutback of 20 pounds.
[00:37:54.960]So assuming that yields held up on this particular field,
[00:37:58.170]we should show a pretty similar improvement
[00:38:00.030]in nitrogen use efficiency in potatoes.
[00:38:02.280]We've had popcorn in Illinois, oats in western Nebraska,
[00:38:05.070]cotton in Tennessee, cotton in Texas,
[00:38:07.500]soybeans in central Nebraska,
[00:38:09.510]rye and orchard grass in the platform so far.
[00:38:12.930]And what's been most interesting about this is because,
[00:38:16.290]essentially, the N-Time algorithms rely
[00:38:19.650]basically on just soil spatial variability data
[00:38:24.803]and essentially, crop reflectance response,
[00:38:27.570]they translate really quickly to other crops
[00:38:29.964]versus modeling solutions
[00:38:31.770]that require knowledge of essentially
[00:38:34.290]how those crops are taking up nitrogen
[00:38:36.690]in different environments,
[00:38:37.950]how the soil is interacting in those different environments
[00:38:40.470]because these are site-specific calibration plots.
[00:38:43.806]And we're simply looking at,
[00:38:46.260]kind of, the geospatial variability,
[00:38:49.440]the crop response, at least in the imagery,
[00:38:51.990]again pending yield response
[00:38:54.360]seems to be pretty consistent across all these crops,
[00:38:57.360]which is really exciting
[00:38:58.470]because it gives us the opportunity to scale faster
[00:39:01.500]than a model-based solution.
[00:39:04.350]So with this information,
[00:39:06.300]we've recently revisited our market opportunity
[00:39:09.330]and see that there's a much greater opportunity
[00:39:11.550]for this technology to be successful
[00:39:13.230]to expand and to impact farmers
[00:39:15.990]across a broad range of crop production systems,
[00:39:18.384]crops and geographies.
[00:39:20.370]So domestically,
[00:39:21.750]we now have identified 12 million fertigated acres
[00:39:24.390]that are addressable with this technology,
[00:39:27.240]either in the near or long term.
[00:39:29.310]About 9 million of those are currently addressable.
[00:39:32.850]And the additional 3 million or more in orchards, vineyards,
[00:39:35.640]and nut trees,
[00:39:37.080]14 out of the 15 top fertigated crop categories
[00:39:39.960]are immediately addressable.
[00:39:41.250]And there are 276 million immediately addressable acres
[00:39:44.160]in the US
[00:39:44.993]if you take fertigation out of that requirement
[00:39:47.460]and you consider pasture and other things
[00:39:50.090]that could stand to benefit from this
[00:39:51.997]sort of nitrogen monitoring.
[00:39:54.330]The global chemigation market
[00:39:56.040]is expected to reach $42 billion this year,
[00:39:59.400]and we've seen that farmers
[00:40:00.540]are currently wasting $200 billion per year in nitrogen
[00:40:03.960]just with the inefficiency
[00:40:05.310]of how they're currently operating internationally.
[00:40:08.070]So we're now seeing
[00:40:09.480]that there's a pretty tremendous opportunity for this market
[00:40:12.451]to capture a large market segment.
[00:40:17.040]Some of the problems that we continue to identify,
[00:40:18.810]our customers identify through sales and through scaling
[00:40:22.400]are that our plots are still not quite big enough.
[00:40:26.190]So we've continued to run into challenges
[00:40:28.650]with AB line communication,
[00:40:30.810]with getting our prescriptions right,
[00:40:32.100]with getting plots established,
[00:40:33.330]and still having to process that as applied data
[00:40:35.192]even as we've made those plots bigger.
[00:40:38.136]So what we're now doing is,
[00:40:39.870]is moving to a big block approach
[00:40:43.740]where we have much larger blocks
[00:40:45.660]that no matter how you intersect those blocks,
[00:40:48.480]you will get to the right rate somewhere in the center
[00:40:51.090]because they're multiple times,
[00:40:52.710]the width of the equipment
[00:40:54.276]and are kind of direction agnostic.
[00:40:57.720]We've talked about the issues with UAV imagery,
[00:40:59.970]the issue with satellite imagery is clouds.
[00:41:03.180]And so we've started to run into immense cloud cover issues
[00:41:05.850]as we've moved into the eastern part of the United States.
[00:41:08.130]And so now we're having to address the issue
[00:41:09.930]with cloud cover,
[00:41:11.040]and how that impacts the amount of time between images.
[00:41:14.520]And communication's been another thing
[00:41:15.960]that we've had to address.
[00:41:17.220]We never really had to address this at research scale
[00:41:19.740]or when we could talk to every farmer ourselves,
[00:41:22.219]but now we don't really have the bandwidth
[00:41:24.570]to talk to every single farmer on a weekly basis
[00:41:27.120]to make sure that they're engaging with the platform.
[00:41:29.400]And so we've had to automate
[00:41:30.840]much more of our communication processes.
[00:41:32.820]So N-Time now sends automated text messages
[00:41:35.130]and push notifications to farmers
[00:41:37.470]to encourage them to engage with
[00:41:39.720]and interact with the imagery
[00:41:41.250]and the insights that are coming through the platform.
[00:41:44.070]As we look at our innovation roadmap
[00:41:45.990]to continue to address some of these challenges,
[00:41:48.870]some of the things that I'm really excited about
[00:41:50.430]are artificial intelligence-driven calibration.
[00:41:53.210]So I've referenced Dr. Schepers several times
[00:41:56.852]in the course of this presentation,
[00:41:59.307]but he and Kyle Holland worked on,
[00:42:01.709]kind of the virtual reference concept
[00:42:04.860]that's been used in nitrogen management.
[00:42:07.170]What we're working on right now
[00:42:08.430]is taking that concept to the next level
[00:42:10.860]using artificial intelligence
[00:42:12.390]and the dataset that we're building
[00:42:13.680]that essentially has a nitrogen trial in every field,
[00:42:17.070]applying artificial intelligence with these daily image sets
[00:42:20.970]to build out an algorithm that can virtually calibrate
[00:42:24.811]and simulate calibration in a paired-plot way
[00:42:28.320]of every single field that we have
[00:42:30.000]on a management zone specific basis.
[00:42:32.130]And so that's some of the new technology we have
[00:42:34.140]that will continue to allow us to use this big block concept
[00:42:37.650]and still calibrate throughout the entire field.
[00:42:41.100]We're also working on model-sensor fusion right now,
[00:42:43.680]so we're working on incorporating the DSAT crop model
[00:42:46.560]that came outta the University of Georgia,
[00:42:47.970]University of Florida,
[00:42:48.803]and the International Fertilizer Development Center
[00:42:51.240]in Muscle Shoals, Alabama.
[00:42:53.040]We're forming an API with that solution
[00:42:55.200]and we're going to use remote sensing data
[00:42:57.420]fused with the DSAT crop model
[00:42:59.970]to ultimately get a full picture of crop nitrogen,
[00:43:03.210]and hopefully come to a mass balance
[00:43:05.250]on crop nitrogen monitoring.
[00:43:07.595]And we're also working on operational efficiency
[00:43:10.050]and environmental,
[00:43:11.108]social and governance value delivery as well.
[00:43:15.360]We're distributing N-Time via our Sentry network,
[00:43:17.520]so we're working with agronomists, co-ops,
[00:43:20.837]precision ag services agencies
[00:43:23.160]to get this technology into the hands of growers.
[00:43:25.920]And what we believe this allows us to do
[00:43:27.690]is make sure that there are boots on the ground
[00:43:29.430]that can verify this technology
[00:43:30.990]and work directly with these farmer relationships
[00:43:33.750]to ensure success of the technology.
[00:43:36.360]So I'll kind of wrap up.
[00:43:37.620]I have a feeling that I'm way over time
[00:43:40.649]that I was allotted here,
[00:43:43.170]but some things that I just would like to call out
[00:43:45.570]is some customer-focused concepts
[00:43:47.970]that we are working to try
[00:43:49.200]to implement at Sentinel Fertigation right now
[00:43:51.840]to ensure that we continue to meet customer needs.
[00:43:54.960]The first one is ruthless prioritization.
[00:43:59.010]Personally, I have a tendency to look at new shiny things
[00:44:01.950]and want to attack every new shiny thing
[00:44:04.440]and we only have so much bandwidth, right, as a team,
[00:44:06.810]as people to truly address all these different challenges.
[00:44:10.110]And so what we're working on right now
[00:44:12.450]is only working on projects or features
[00:44:14.730]that solve customer problems,
[00:44:16.470]that we know are customer problems
[00:44:18.720]and that are also within our scope of capabilities.
[00:44:21.360]We've gotten a ton of questions about
[00:44:23.190]can we help with irrigation scheduling?
[00:44:25.560]Can we monitor for pests, for disease, for weeds?
[00:44:31.500]None of those things
[00:44:32.400]are within our core scope of capabilities.
[00:44:35.100]And so we are pushing all of those projects off of our plate
[00:44:39.390]and not addressing those right now.
[00:44:41.190]And instead we will look to partner with companies
[00:44:43.470]that are solving those challenges
[00:44:44.970]because we believe that our mandate right now
[00:44:47.580]is to solve nutrient management challenges.
[00:44:49.320]And so that's what we're gonna continue to focus on.
[00:44:52.470]We also,
[00:44:53.580]while being a highly prioritized organization,
[00:44:56.910]we're also trying to be opportunistic, right?
[00:44:59.007]And so if we're making these observations
[00:45:01.800]or we're in 10 or 20 customer conversations in a week
[00:45:05.190]and the same trend comes up in 15 of those 20,
[00:45:08.190]we need to probably pay attention to that
[00:45:09.930]and identify if there's an opportunity
[00:45:11.940]to pursue something new
[00:45:13.020]that maybe should be prioritized
[00:45:14.490]over anything else that's currently on our board.
[00:45:17.580]We're also kind of taking the step
[00:45:19.680]of incremental improvement.
[00:45:21.150]So what small enhancements,
[00:45:22.590]what small changes can we make
[00:45:24.464]that better return value to our customers, right?
[00:45:29.130]Is there something small that we can do,
[00:45:31.200]such as shifting how we display our information
[00:45:34.920]in the insights tab
[00:45:35.940]or how we communicate with our customers
[00:45:38.670]that can ultimately make this a much more usable product
[00:45:43.230]and allow them to continue to adopt it,
[00:45:45.270]continue to use it,
[00:45:46.230]and ultimately advance their stewardship
[00:45:47.930]in nitrogen management practices.
[00:45:49.890]And then targeted innovation
[00:45:51.330]is kinda the last customer-focused concept
[00:45:53.820]that we're working on adopting right now.
[00:45:57.180]So it's not just innovating simply because we can,
[00:46:00.150]there's a lot of data that we have
[00:46:02.011]and a lot of opportunities that we have
[00:46:04.680]to innovate using that data,
[00:46:06.570]but unless it's solving a critical problem for our customers
[00:46:10.410]or for our business,
[00:46:11.520]it's not really a great use of our time, right?
[00:46:14.220]And so, it's something that we're trying to do
[00:46:16.380]to really target our innovation at those critical problems
[00:46:19.290]that we're particularly focused on.
[00:46:23.130]And then we're also adopting some product
[00:46:24.510]and innovation processes.
[00:46:25.590]I won't dive into this
[00:46:26.423]because I wanna save some time for questions,
[00:46:29.700]but the big thing here
[00:46:30.930]is we are adopting a very iterative process right now
[00:46:34.500]where we're trying to release new capabilities regularly,
[00:46:38.400]show them to customers,
[00:46:39.480]gather that feedback very quickly
[00:46:41.550]and incorporate them into a more comprehensive technology
[00:46:44.520]that addresses those customer needs.
[00:46:47.250]And I'll finish up here just talking about our impact
[00:46:49.470]and a couple takeaways.
[00:46:50.580]So between 2022 and 2023,
[00:46:54.180]we've operated on roughly 26,000 acres
[00:46:56.760]primarily in Nebraska.
[00:46:59.340]If you take our 2022 data
[00:47:01.800]and extrapolate it across this year's acres,
[00:47:03.720]which largely the data indicates
[00:47:05.610]that we should see similar results to what we saw last year,
[00:47:09.290]we will have seen a over a $2 million increase
[00:47:12.660]in farm profits,
[00:47:14.310]we will have saved nearly $700,000 on nitrogen costs
[00:47:19.290]for the farms that we've worked with,
[00:47:21.450]and we will have mitigated
[00:47:23.220]almost 1.1 million pounds of nitrogen
[00:47:25.320]from ever entering the environment.
[00:47:27.840]To put that in context,
[00:47:29.580]that's a little bit less than 1%
[00:47:31.770]of the annual increase in nitrogen applied to corn in the US
[00:47:35.370]over the past 23 years.
[00:47:38.700]By 2026, if we continue on our current growth trajectory,
[00:47:42.120]what we have planned
[00:47:43.920]and continue to produce the same results
[00:47:45.960]that we're seeing right now,
[00:47:47.640]we should see an increase of over $93 million
[00:47:50.190]in farm profits using this technology,
[00:47:52.920]there should be a savings of nearly $32 million
[00:47:55.620]on nitrogen costs accumulated over the next four years
[00:47:59.130]or between 2022,
[00:48:00.210]the four years, including 2022 to 2026.
[00:48:03.630]And then we should have mitigated over 50 million pounds
[00:48:06.570]of nitrogen from ever entering the environment,
[00:48:08.400]which is a really, really exciting thing.
[00:48:10.077]And that starts to be nearly 50% of that annual increase
[00:48:13.928]in nitrogen applied to corn in the United States
[00:48:16.980]year over year since 2000.
[00:48:19.200]So that starts to be a really noticeable impact
[00:48:21.750]and something that gets me excited,
[00:48:23.820]gets our team excited,
[00:48:25.350]and we're looking forward to achieving.
[00:48:28.260]So last few takeaways.
[00:48:30.990]We are seeking to use these customer-driven capabilities
[00:48:33.870]to continue to enhance farm profits
[00:48:36.030]and environmental stewardship.
[00:48:38.130]And it's my belief that customers
[00:48:39.690]are at the core of successful R&D.
[00:48:41.400]That doesn't matter whether that R&D
[00:48:43.170]is commercial or academic.
[00:48:44.901]I saw customers be a huge part of the R&D
[00:48:47.730]that we did at UNL.
[00:48:50.400]They should be involved before and during R&D,
[00:48:53.010]as well as after.
[00:48:54.420]So I think, you know,
[00:48:55.710]really when we're going to,
[00:48:57.330]it's Sentinel, at least when we're proposing grants,
[00:48:59.190]we're talking to customers on the very front end of those
[00:49:01.800]to see is this a problem
[00:49:02.730]that's even really worth us trying to solve for you.
[00:49:06.420]From what I've seen,
[00:49:07.290]customers will point you to the right problems to solve,
[00:49:09.630]but generally, they don't point you to the right solutions,
[00:49:12.210]right?
[00:49:13.043]And so they're able to identify the problems
[00:49:14.910]that they have really well,
[00:49:15.780]but the best way to solve them
[00:49:16.890]is typically left to the people that are an expert
[00:49:20.010]in that particular domain to propose the right solution
[00:49:23.610]and ultimately gather feedback on that solution.
[00:49:26.190]And the last thing that we're learning is that
[00:49:27.960]customers should be chosen carefully, right?
[00:49:30.330]So we've had a few bad customers
[00:49:32.236]over the course of the last couple years
[00:49:35.339]and one of the best pieces of advice
[00:49:37.770]that I've gotten from our investors,
[00:49:40.380]from board members, from advisors
[00:49:42.540]is it's okay to fire customers
[00:49:44.970]because sometimes,
[00:49:45.803]there are folks that want to use the technology,
[00:49:48.490]and maybe they don't truly want to use the technology,
[00:49:50.574]but they have some other motivation or some, you know,
[00:49:54.390]preexisting negative notion about the technology
[00:49:57.330]that ultimately leads to bad outcomes.
[00:50:00.000]And it's better to have fewer good customers
[00:50:01.980]that achieve good outcomes
[00:50:03.120]and continue to push you in the right direction
[00:50:05.123]than it is to have a ton of customers
[00:50:07.020]with a lot of bad customers that may,
[00:50:08.910]may tarnish what the technology is truly capable of
[00:50:11.130]by not embracing it fully.
[00:50:13.200]So hopefully, that was good information and valuable.
[00:50:18.000]I'd be happy to take any questions that y'all might have.
[00:50:20.340]We open the floor for questions
[00:50:22.077]and for those online,
[00:50:23.130]you can feel free to put them in the chat
[00:50:24.990]and we can start addressing them.
[00:50:28.915][Audience Member 1] I noticed
[00:50:29.748]in one of your last slides there
[00:50:31.140]that the cost savings on nitrogen
[00:50:35.250]was about one-third of the total change in profits.
[00:50:40.170]Where'd the other two-thirds come from?
[00:50:42.300]Yield, actually.
[00:50:44.400]So last year,
[00:50:45.360]and then this is against the farmer baseline.
[00:50:47.670]So we had a high-yielding corn year last year,
[00:50:50.490]and I fully recognize that.
[00:50:52.680]Against a farmer baseline,
[00:50:54.120]last year, 71% of our fields actually had increased yield.
[00:50:58.320]And if you look at the price of corn
[00:51:00.240]and nitrogen fertilizer last year,
[00:51:02.730]increased profit on farms
[00:51:04.020]was actually somewhere in the neighborhood of $120 per acre,
[00:51:07.320]the fields that we operate,
[00:51:08.580]the farms that we operated with.
[00:51:10.410]Looking at 10-year nitrogen prices,
[00:51:12.660]you're looking at more like $77 per acre
[00:51:15.140]in increased profitability last year.
[00:51:16.560]But that's kind of where that,
[00:51:18.450]that three times higher farm profit
[00:51:21.870]versus nitrogen cost savings came from.
[00:51:24.450]We'll see how consistent that continues to be,
[00:51:26.677]but that's the reason behind it.
[00:51:31.515][Audience Member 2] Is there always
[00:51:32.348]a useful detectable difference between the canary
[00:51:35.430]and the reference plots?
[00:51:37.557]There's not always a detectable difference
[00:51:40.440]and that's ultimately what we're trying to achieve.
[00:51:43.380]So ideally, what we should see happen is that
[00:51:46.920]the canary will start to show a difference
[00:51:50.070]from that reference plot and with an intervention.
[00:51:52.380]So with an nitrogen application,
[00:51:53.760]it comes up to show a non-detectable difference.
[00:51:56.280]And over the course of time,
[00:51:57.210]we should see that happen repeatedly
[00:51:59.130]over the course of the growing season
[00:52:00.390]if we're intervening in a real time fashion
[00:52:02.610]in getting that crop back up to sufficiency.
[00:52:05.730]And then towards the end of the growing season,
[00:52:07.713]as the crop seneses, as it reaches physiological maturity,
[00:52:12.240]what we should start to see is that
[00:52:14.220]all of our fields are approaching a deficient state, right?
[00:52:17.310]So the canary should start to show more difference
[00:52:19.680]as it dries down faster than the rest of the crop,
[00:52:22.650]which fortunately,
[00:52:23.760]we're seeing in most of our fields right now,
[00:52:25.740]and it's an indicator that we're basically within 30 pounds
[00:52:29.970]of where that optimal nitrogen rate really should be.
[00:52:32.250]So that's what we expect outta that response.
[00:52:37.152][Audience Member 3] I was intrigued
[00:52:38.190]by your 70% retention that you mentioned early on.
[00:52:44.130]Does that relate to choosing your customers carefully?
[00:52:48.390]I would say that two of the customers that we lost
[00:52:50.848]were, you know,
[00:52:51.799]a situation of choosing our customers carefully
[00:52:55.170]and there were three
[00:52:56.776]that decided to go in a different direction
[00:52:59.250]that were due to communication actually.
[00:53:00.924]We did not serve them very well
[00:53:03.000]and they didn't understand the value
[00:53:05.190]of what we were providing.
[00:53:06.600]I had an intern actually that I kind of made responsible for
[00:53:11.190]about half of our grower segment
[00:53:12.990]and I think it was a little bit too early for him
[00:53:15.060]to really embrace the customer service side.
[00:53:17.790]And so they were customers
[00:53:18.960]that hadn't heard from somebody in a month from our team
[00:53:21.960]and that did not go over very well with them
[00:53:24.180]and they decided to move on on because of that.
[00:53:26.961][Audience Member 3] Would you project that
[00:53:27.990]that'd be a number you'd have in the future, 70%?
[00:53:32.580]I project us to have 80%,
[00:53:34.740]if we can score 80% retention or higher in the future
[00:53:37.800]across all of our acres, we will be successful.
[00:53:40.188]So that's what we're shooting for is 80%.
[00:53:42.702][Audience Member 3] I can see where that'd be important.
[00:53:44.700]Absolutely.
[00:53:48.238][Audience Member 4] Interesting presentation,
[00:53:49.410]thank you very much.
[00:53:51.540]I read this as that you have checks
[00:53:54.750]in every field you're looking at.
[00:53:57.150]I don't know where all the data that you get comes from,
[00:54:00.185]satellites, I presume correctly.
[00:54:02.550]Correct.
[00:54:03.383][Audience Member 4] And so instead of maybe
[00:54:06.330]having the grower adjust their nitrogen program,
[00:54:11.190]you put the checks in there
[00:54:12.630]and let that data drive recommendations,
[00:54:16.230]don't add or add or whatever,
[00:54:18.360]and timing wise. Exactly.
[00:54:20.197][Audience Member 4] Okay. Yep.
[00:54:21.993][Audience Member 4] And I used to work
[00:54:24.900]for a commercial company
[00:54:26.940]and we had an agronomist in Texas
[00:54:29.730]that worked with about a 1000 pivot operators.
[00:54:34.950]And his mantra was,
[00:54:36.900]I don't want to change your fertilizer rate necessarily,
[00:54:40.860]I just wanna change when you apply it.
[00:54:43.217]Sure.
[00:54:44.050][Audience Member 4] And you're providing the data
[00:54:44.970]to drive that application, yay or nay,
[00:54:47.580]is what I'm reading into what you're talking about, correct?
[00:54:50.984]Exactly, yes. [Audience Member 4] Okay.
[00:54:53.619]And to put that in a little bit of of context,
[00:54:56.160]we are a fertigation scheduling tool, so it's exactly that.
[00:54:58.920]We're trying to time those applications better.
[00:55:01.080]So there have been some growers,
[00:55:02.190]we've not changed that rate at all,
[00:55:03.630]and it's actually just been better timing,
[00:55:05.040]whether that's earlier or later,
[00:55:06.630]and then there are other growers.
[00:55:07.929][Audience Member 4] Where they used...
[00:55:08.762]You ended up with whatever they would use normally,
[00:55:12.570]but you just had a better timing concept or schedule?
[00:55:17.040]Correct. [Audience Member 4] Okay.
[00:55:17.873]And scaling up this year,
[00:55:18.870]we've actually seen in Illinois and in western Nebraska,
[00:55:21.517]especially in western Nebraska,
[00:55:23.400]where they've gotten more rain this year
[00:55:24.720]than what they're used to,
[00:55:26.164]they had more flushing events early in the season
[00:55:28.920]and we've actually recommended more nitrogen
[00:55:31.110]than what those farmers have typically applied
[00:55:33.330]to make sure that we're keeping up with yield potential
[00:55:36.155]in those areas.
[00:55:37.299][Audience Member 4] And those checks
[00:55:38.310]will demonstrate that?
[00:55:40.080]Correct. [Audience Member 4] Okay.
[00:55:42.716][Audience Member 5] Nice job. Jackson.
[00:55:45.480]In your slides up here,
[00:55:47.070]you showed spatial variability in some of those slides,
[00:55:51.060]that layer is really the crop chlorophyll or biomass layer,
[00:55:58.020]but that's driven by another layer, the soil layer.
[00:56:03.240]And so, do you at this point incorporate a second layer?
[00:56:09.240]You talked about artificial intelligence.
[00:56:12.570]What are the chances of putting those two together?
[00:56:16.770]Is that in your vision?
[00:56:18.960]Yeah, absolutely.
[00:56:19.980]So we currently use management zones
[00:56:22.800]that are based on preferably soil electrical conductivity,
[00:56:28.020]soil organic matter,
[00:56:30.390]slope elevation,
[00:56:31.770]and multi-year yield history.
[00:56:33.540]And so those management zones we actually use to drive
[00:56:36.330]the placement of those calibration plots in the field.
[00:56:39.360]So we represent each individual management zone
[00:56:41.280]in a field with those.
[00:56:43.170]And our goal now with the nitrogen model
[00:56:46.020]that we're incorporating with DSAT
[00:56:48.300]is to spatially reference that model
[00:56:51.420]and calibrate the initial conditions with that soil data,
[00:56:54.540]layer on the remote sensing data
[00:56:57.300]and fuse that in to calibrate the model in real time
[00:56:59.970]and then use it to out project
[00:57:01.980]over the course of the next two weeks
[00:57:03.150]what those needs are gonna be, so.
[00:57:06.604][Audience Member 6] Jackson, nice presentation.
[00:57:08.843]My question relates to,
[00:57:10.603]you are projected to grow, right,
[00:57:13.020]as a company with the technology you are utilizing.
[00:57:18.030]What is your sense in terms of
[00:57:22.500]capacity building of the crop advisors,
[00:57:26.190]the education we need to do at the university?
[00:57:29.310]If you could share something along that line
[00:57:31.500]on what we need to be thinking about
[00:57:34.482]if you're going to need the workforce of technology
[00:57:38.820]verse agronomists.
[00:57:42.270]That is a really, really hard question.
[00:57:46.380]Thank you (chuckling)
[00:57:50.820]Right now, I think,
[00:57:53.520]what we've seen is that we don't have enough agronomists
[00:57:56.160]that are digitally inclined.
[00:57:59.190]And so, I think
[00:58:00.240]that's the biggest gap in education right now
[00:58:02.640]is improving that digital,
[00:58:04.800]you know, inclination of agronomists.
[00:58:07.290]We're seeing some young agronomists
[00:58:08.940]coming out into the market
[00:58:10.650]that are starting their own
[00:58:11.490]independent crop consulting businesses.
[00:58:13.380]And we've actually seen some success
[00:58:14.550]with getting our technology in their hands
[00:58:16.590]and they're working with young producers
[00:58:17.970]that are excited about this.
[00:58:19.080]But we also have a lot of older agronomists
[00:58:21.300]that maybe have never picked up on these tools
[00:58:24.960]and have been very, honestly, dismissive
[00:58:27.210]when we first talked to them.
[00:58:28.260]You know, they've got their models
[00:58:29.580]that they've built for the past 30 years
[00:58:30.930]and they're not gonna change them
[00:58:31.860]or incorporate new data to influence those.
[00:58:34.950]And so I think just continuing to build a workforce
[00:58:37.530]that is, you know,
[00:58:38.363]equipped with knowledge of digital tools
[00:58:42.270]and how to use those
[00:58:43.320]is probably the biggest thing that the university can do.
[00:58:45.750]And I think ideally,
[00:58:47.730]using the usefulness of these digital tools,
[00:58:52.180]which are really good for in-season adjustment
[00:58:55.080]and making decisions in real time
[00:58:58.890]with the recommendations that we need to have
[00:59:01.710]as kind of a standard operating procedure
[00:59:04.290]to just give people, you know,
[00:59:05.520]bounds in which to think, right?
[00:59:06.750]So if we think about the UNL nitrogen algorithm
[00:59:10.530]and the publications that come out about that algorithm,
[00:59:13.710]it would be very helpful to have right alongside it,
[00:59:15.810]this is how it best pairs with a digital tool, right?
[00:59:19.200]To show where the strengths and weaknesses are
[00:59:21.150]of each one of these tools,
[00:59:22.500]'cause our tool right now is horrible for planning nitrogen
[00:59:25.470]and estimating the total amount that you need
[00:59:27.300]and giving you the ability
[00:59:28.290]to make a budget for the next year.
[00:59:29.640]The UNL nitrogen algorithm does a great job
[00:59:32.340]of doing that right now
[00:59:33.240]to at least give you an idea of where that's gonna be.
[00:59:36.030]But the nitrogen algorithm
[00:59:36.960]does a horrible job of accounting for
[00:59:38.820]what the weather conditions are going to be
[00:59:40.380]and what that crop yield potential actually is.
[00:59:42.540]And that's where our tool shines, right?
[00:59:43.887]And so I think having those pieces of information
[00:59:46.410]well compared for producers and agronomists
[00:59:48.540]is another big education component
[00:59:50.490]that could come outta the university.
[00:59:53.700]Well, thank you Jackson.
[00:59:54.870]Unfortunately, we run out of time,
[00:59:56.580]but he will stay around for some questions for the audience,
[01:00:00.030]but we need to keep this on time.
[01:00:02.040]So thanks again for being here.
[01:00:04.385](audience clapping)

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

Scaling On-Farm Research in Image-Based Fertigation with Customer-Driven Development

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments