The agricultural department at SMSU has several major areas of study including ag business, agronomy, and ag education. SMSU also has ~50 acres of crop land donated by a local alumni and farmer which we routinely use for hands-on student learning. 2022 will be the eighth season this plot has been established and available for SMSU students. Agriculture focused majors are by far the most common type of student that uses the SMSU research farm for learning however I have been making inroads with Culinology, Business, and Art majors. My primary responsibility at SMSU is an educator, and depending on the make-up of the courses I teach, I estimate around 20-30% of my students each semester actively farm, with another 20-30% having some sort of close familial connection to farming. When I ask my farming students why they do certain practices on the farm during lectures, they often respond “I don’t know”. There is a lot of “Get in the tractor and drive” being told to my farming students and one thing I’ve realized as an educator, is that the farmer undergrads often can’t communicate the why of whatever practice they are conducting in their field. In order to remedy this, I demonstrate a variety of good, bad, conventional, and experimental farming practices at the SMSU research farm to provide visual aids for courses and hands on learning opportunities.
The purpose of this grant is to provide one of these experimental learning opportunities. It will allow students to see how additional N application in soybean can influence yields, how heavy weed pressure competes with the target crop, and finally, how drones can be used to investigate crop health/development. I choose varying N-rates as the primary variable of interest as the USDA-NASS estimates the average MN soybean grower fertilizes at 11 lbs of N per acre (USDA-NASS 2020). When this common N-rate is presented in class one of the most common questions I get is “What would happen to the soybeans if we fertilized at a higher rate?” The research project proposed by the grant will help me and the students arrive at the answer to this question in a hands on manner. While this proposal is primarily an education grant, my first objective is research aligned. I have structured the grant as such for clarity purposes, and the order of goals does not represent importance. It is easier to present the research project first, how it will be achieved, and then describe how the project will further student learning at SMSU, rather than presenting them in the opposite order. Said another way, GOAL 1 will provide the basis and meet of the learning opportunities in GOAL 2 and GOAL 3.

GOAL 1 : Describe how the weed canopy develops and how N rates change soybean canopy growth.

OBJECTIVE : Using drone imagery in conjunction with image analysis software, the canopy space of weeds (primarily lambsquarters) and soybeans grown at different N rates will be quantified throughout the season and correlated to season total yield. Weed canopy development will also be contextualized in the content of soybean growth and development.

GOAL 2: Demonstrate/Educate SMSU students on the advantages and limitations of drone technology and its potential place in the future of agriculture

OBJECTIVE : Using drone imagery collected 2022 summer, students will be able to determine how N rates influence canopy closure and yields in conventionally managed soybeans using image analysis in agronomy classes. The plots will also serve as a visual resource for local and up-and-coming (student) growers.

GOAL 3: Provide hands-on learning/training opportunities in agriculture to SMSU students.

OBJECTIVE : The vast majority of field work and image processing done to achieve this project’s goals will be accomplished by SMSU students.

In order to investigate and provide data on the role of N-rates on canopy space and development in soybeans, varying rates of N (10, 20, 30 and 40 lbs/ac) will be used. All fertilizer treatments will receive 23 lbs/ac of P and 30 lbs/ac of K. In addition to fertilizer treatments, an additional soybean (@ 10 lbs N/ac) + weed treatment will be included, as well as a weed only treatment for a total of 6 experimental treatments (10, 20, 30 and 40 lbs/ac, 10 lbs N/ac + weeds, and weeds). Each treatment will be analyzed in a randomized complete block design with a plot size of 8 rows (30” spacing), 60 feet long each, with a targeted planting population of 130,000 seeds/ac. Each treatment will be replicated a minimum of four times. Yield data and nodulation counts will be collected in autumn 2021 and analyzed with an ANOVA followed with a Tukey posthoc mean separation to determine significant differences between treatments. SMSU Agronomy owns all farm equipment needed to conduct this research.
Additional collected data will include weekly aerial photography performed by drone. SMSU owns a DJI-Mavic Pro drone with high resolution photography capabilities which is routinely used for courses and projects. Aerial imagery will be taken weekly at solar noon (to minimize the impact of shadows during image analysis) throughout the growing season until all treatments fill their canopy space. In order to analyze these images, a student worker will load them into the ImageJ program, isolate the plant parts that need measurement, and calculate the canopy space in feet2. Once the canopy space has been calculated, we can analyze each sampling date with ANOVA to determine any different in canopy development as a function of N-fertilization.
The best way to describe how ImageJ works is with a demonstration. As such, a representative image of how students use these drones to take aerial imagery and how they can manipulate the photos with the ImageJ software to address agronomic questions has been included in this proposal (Fig. 1). In the attached image, ImageJ was used to change the color of the corn plants within a research plot to be white, and the weeds to be red. The program that allows the color coding of plants will also “count” the number of red pixels in the photo, and with proper calibration, can estimate the area dominated by weeds, and those dominated by crops. In taking weekly aerial images, students funded by this grant will be able to experience a rudimentary form of precision scouting without having to purchase an expensive sensor, nor learn complicated programs needed to process NDVI data. ImageJ is relatively user friendly and I haven’t had any trouble teaching students how to use it in my “AGRO 390 Precision Ag” class.

Figure 1 ImageJ manipulated photo of weed presence (red) in a corn plot (bright white). ImageJ is also able to calculate the red area to estimate how much "canopy space" the weeds are taking up. Photomanipulation and analysis conducted by a Junior Agribusiness major in AGRO 390 Precision Ag.
Data gathered from these weekly aerial images will provide the SMSU agronomy program with access to a season long look at the canopy development of soybeans and weeds. Additionally we will be able to see how canopy development changes as a result of fertilization status, and if additionally what N fertilization rate will have an impact on yield. Once all of this data is analyzed, GOAL 1 will be fulfilled. Additionally as much of this analysis and field work will be performed by a SMSU student intern, GOAL 3 will also be addressed.
GOAL 2 and 3 will primarily be met through utilization of data generated by this project in the class room. The list in the “Project Deliverables” section of the proposal describes how data will be used in at least five of SMSU’s Agronomy courses and achieve GOAL 2 and 3.
Project Deliverables: (limit 14,000 char.)
While the focus of this grant is centered on student education, the preliminary results of this project will be shared with the local farming community during the annual SMSU agronomy field day. Last year this event was attended by ~100 local farmers and business representatives and SMSU agronomy was able to showcase the type of research it was conducting at the farm. This is a great opportunity to brag on the department and provide a venue to others on how up and coming drone tech can be used in a field setting. While the research won’t be done in summer of 2022, a poster of our findings will be presented in summer 2023.
I also believe a major strength of this student managed project is that it utilizes non-specialized equipment. One of the most cost effective NDVI sensors runs at ~$2000 and doesn’t include the drone upon which it gets attached, nor the software to analyze any collected data. Additionally much of the NDVI research is still being conducted and while researchers are starting to understand correlations between NDVI values and crop health/yield, implementation of this information on a farm scale has not yet been attained. Furthermore, drones are becoming increasingly popular and affordable, especially within the farming community. A quick search of drones and any number of agricultural terms on Youtube.com will result in a large number of high quality, farmer-filmed, which are using drones to capture their footage. This project will synergize the growing popularity of drones with hands-on student learning needs, and provide the images and data I need for future assignments in the agronomy courses I teach. Furthermore, since drones are already popular with my ag students, I will get easy buy in and participation from them.
As mentioned, SMSU agriculture students will be able to use the plot as an educational tool to learn various aspects of soybean growth and development. Data generated from the field plot will be used to provide at least 5 different learning opportunities:
AGRO 132 Crop Production + Lab: This class is the equivalent of an “Agriculture 101” class and is required by all ag majors at SMSU. As such it averages ~20 students each fall semester. For one lab, students will go to the field and make observations on how varying N-rates changed plant growth and canopy space. Once we return to school, students will be presented with yield data as a function of N-rate, and be asked to make a “Yield response to fertilization” graph using the real world data. In a series of questions associated with the assignment they will be able to determine what the most economic N-rate would be at varying fertilizer price points. An example of such a graph is presented below using fake data (Fig. 2).

Figure 2 Crop yield response to increase of N fertilizer rate
AGRO 212 Grain and Forage Crop Management: This class is required by Agronomy and Agriculture Solutions majors, and averages around 11 students every other fall. The material covered in this course is the most agronomist centric of the agronomy courses and focuses on the production of corn, soy, and alfalfa and topics such as optimal planting rates, best fertilization practices, and genetic traits that impact production. This course covers these crops to a greater detail than any of the other agronomy courses. The field plots of this trial will be used in conjunction with a corn fertility trial as well as part of a field trip to the research plots. Photos taken of the fertility plots and personal experiences will be used and communicated in future lectures as well.
AGRO 341 Principles of Pest Management + Lab: This class is required by Agronomy and Agriculture Solutions majors, and averages around 9 students every fall. This class goes to the SMSU field plots every week to create scouting journals, and collect pests for a curated collection as part of a semester long project. The soybean + weed treatment of this project will provide an interactive demonstration of how weeds actively impact yield and time series data from the weed only treatment will help demonstrate how quickly weeds can outcompete a soybean field if allowed to canopy.
AGRO 390 Precision Ag: This class is required by Agronomy, Agriculture Solutions, and Ag Ed majors. As such it averages ~20 students each fall semester. This course covers precision agriculture tools including variable rate technology, equipment auto-guidance, remote and on-the-go sensing, and gets into the mechanisms of how these tools are utilized. The major assignment in this class is when students get a chance to fly the SMSU drone, take photos, and develop a project/answer a question using the ImageJ program. Fig. 1 is an example of one such student project. This proposal will create a time series of soybean and weed canopy growth at different N rates and give students “hands-on-experience” in analyzing remotely sensed data to make conclusions. Furthermore, they would be free to use the soybean plots while they are still standing to develop their own semester project.
AGRO 454 Experimental Design in Agriculture + Lab: This class is an Agronomy major elective offered every other spring. As such I have only taught it once and it had 7 students. I’m actively proselytizing this course to the Biology and Environmental Science department and expect a greater and more diverse enrollment in the future offerings of this course. In this class, students use real world data to learn about experimental design in typical agronomic test plots including assessment of insecticide sprays, the need for replication and blocking, and eventually are able to run an ANOVA to interpret findings from the field. Each week student analyze a new dataset, and the time series this proposal will generate will provide ample opportunities to answer a variety of crop production questions, as well as demonstrate the many ways in which viable data can be collected (drones!).

Update:
At this point in the project, all data except yield has been collected. Soybean pods are filling up with some dry down occurring. Analysis of drone imagery is underway. The goal of this project and preliminary results have been communicated to ~120 individuals since funding was received.

Update:
Progress report in the attachment as tables are involved.

View uploaded report

Update:

View uploaded report

Update:

View uploaded report

Update:

View uploaded report

The primary purpose of this grant was to provide experimental learning opportunities to students both those in class, and those conducting the field work. This project has allowed students to see firsthand how additional N application in soybean can influence yields (or in our case, not at all), how heavy weed pressure competes with the target crop, and finally, how drones can be used to investigate crop health/development.

Our major finding was documenting how quickly weeds can outpace soybean growth when measured via drone. At 4 and 5 weeks post plant, our soybean plots with weed treatments outpaced the soybean only treatments in canopy space by around 10-20% points. It wasn’t until week 6 that the soybean only plots matched the canopy space that was reported in week 4 and 5 for the weed plots. By week 6, the soybean only plots had caught up in canopy space in comparison to the weed plots.

Another main finding, and good lesson for students is the data collected is only as good as the tools used to collect it with. Our drone sampling method found that soybean canopy coverage peaked at around 60-70%. This is simply not true as when we went to the field in late season, the canopy coverage was at 100%. This discrepancy is largely due to inadequacies in the pixel-color sorting algorithms that the computer program we used to analyze data uses. In essence, the leaves at the top of the plant canopy are a bright green color under sunlight, but those at a lower canopy level are a darker green due to shadows of the higher up leaves. The pixel sorting algorithms cannot recognize the darker green pixels and thus don’t “count” them in the canopy estimation. Taking photos at solar noon will minimize this impact but regardless it will still be there. Had we not ground truthed the soybeans canopy coverage, we very well may have believed we only achieved 60-70% canopy coverage. As such, all of our canopy closure values can be considered conservative estimations and additional means of data collection should occur.

The results (preliminary and final) of this project have been communicated to ~160 non-student individuals (as of Nov 15th 2022) since funding was received. Furthermore, the SMSU research plots have received 50-75 unique undergraduate students (~1/20th of the fulltime undergraduate SMSU student body) in the 2022-23 academic year. The soybean plots funded by this project gave SMSU agronomy the opportunity to augment/enhance student learning for ag focused students, and at the least, help introduce non-ag students to agriculture. This brief introduction for the non-ag students may be superficial at first, but it can help lead to a greater understanding of what types of jobs are involved in modern farming.

The primary manner in which this project will positively impact MN soybean farmers is via education and demonstration. Agriculture students at SMSU are the future crop scouts, consultants, and product sales service that our MN soybean farmers rely on, and as such, training investments made in today’s agriculture students, will pay major dividends once these students hit the job market. All agronomy students at SMSU need to complete an internship by the time they graduate. At the end of the semester they give a presentation on their learning experiences. One of the most common themes in these presentations is how technologically advanced farming is becoming be in precision planting, or using drones to scout 100s of acres at a time. It is clear farming will have a greater technological component going forward, but students still need to get their boots muddy and know how proper implementation of technological advances can benefit the grower. This grant helps achieve this goal.
SMSU is also the recipient of another MN soy growers grant in which we train students to pass the CCA exam. We have been largely successful with 5 out of 7 upperclassmen students passing. I also took the CCA exam in order to assess how well SMSU agronomy’s course material reflected the test material (and passed!) but was surprised at the number of precision agriculture questions the test asked. As such, I’m placing special emphasis on the AGRO 390 Precision Ag course going forward and will incorporate more tech in other agronomy courses. As detailed in my project deliverables section, students will use the plots and data associated with this project in multiple ways across 5 different classes. This breadth of learning options will make students better well rounded by the time they graduate and go on to take the CCA test.