Source: UNIVERSITY OF ILLINOIS submitted to
USING PRECISION TECHNOLOGY IN ON-FARM FIELD TRIALS TO ENABLE DATA-INTENSIVE FERTILIZER MANAGEMENT
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
EXTENDED
Funding Source
Reporting Frequency
Annual
Accession No.
1008818
Grant No.
2016-68004-24769
Project No.
ILLU-470-612
Proposal No.
2015-08733
Multistate No.
(N/A)
Program Code
A5160
Project Start Date
Mar 1, 2016
Project End Date
Feb 28, 2022
Grant Year
2018
Project Director
Bullock, D. S.
Recipient Organization
UNIVERSITY OF ILLINOIS
2001 S. Lincoln Ave.
URBANA,IL 61801
Performing Department
Agricultural and Consumer Econ
Non Technical Summary
Food and agricultural systems manipulate the nitrogen cycle to great benefit, but chronic inefficient use of nitrogen fertilizer has led to the hypoxic "dead zone" in the Gulf of Mexico and the leaching of nitrates into groundwater. Mismanagement of phosphorous fertilizer has led to algae-choked bodies of fresh water. There is a critical need to generate reliable, site-specific estimates of optimal fertilizer application rates and timing that will be readily adopted by farmers. To generate high quality data, we will use precision agriculture technology and develop CyberGIS-based software to run low-cost, large-scale, on-farm agronomic field trials conducted by farmers during normal operations. Our research will address the Agricultural Production Systems Program Area (A5160) Priority, in that we will increase agricultural productivity and food security by generating data used to provide management advice that will allow farmers to achieve crop yields with reduced fertilizer input use while limiting nitrogen contamination of the nations waters. Additionally, project results will aid in the development of socially sustainable agro-environmental policy.
Animal Health Component
0%
Research Effort Categories
Basic
70%
Applied
30%
Developmental
0%
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
1025210202030%
6011510209030%
1120320202030%
6105210301010%
Goals / Objectives
Project objectives are to (1) develop a data-driven on-farm precision-agriculture-based research-extension-teaching infrastructure that raises farm income by facilitating movement of data and valuable management information between farmers and researchers; (2) provide to the public and policy makers accurate information about the costs and benefits of agricultural fertilization policies, especially in regards but not limited to U.S. corn and soybean production; and (3) train and place selected under-represented students in positions to improve agricultural research and agribusiness collaborations between the U.S. and Latin America.
Project Methods
Activities and Efforts:Use precision agriculture technology to design and conduct thirty (as part of the proposed project), later hundreds of long-term, large-scale, on-farm field trials in U.S. and abroad with international partnersRecord site-specific yields in field trialsMonitor yield and nutrient loss on six of the fields under trialDevelop "Data-Intensive Farm Management" softwareDevelop "On-Farm Research Design"(OFRD) softwareContribute to design and construction of state-of-the-art 160-acre field laboratoriesRun experiments on field laboratories to study relationships between fertilizer management strategies, yield, and water qualityGenerate, store, maintain, provide access to massive amounts of data on yield and water quality responseAnalyze the entire data set to estimate "meta response function," used to give advice to non-participating farmers and determine efficient yield-water quality output combinationsBuild extension and consulting infrastructure to bring farm data to research and research results back to farmersDetermine how research results are best communicated to farmers for relevance and accessibilityProvide advice for efficient policyTrain a cohort of undergraduate agriculture students in computer science, statistics, and "big data"Evaluation/Measures of Success:Number of field trials run 2016-2019 (We have pledged to run at least 100 trials in total in the four years.)Number of on-farm field trials conducted after 2019 (We would consider running 100 experiments per year by 2021 to be an excellent startWhether OFRD software is beginning to be used by researchers worldwide in the conduct of large-scale field trialsWhether DIFM software is being adopted commercially, and whether consultants are attending our training workshops.(We hope to be training at least 25 consultants per workshop, with two workshops per year by 2021)Number of FAB scholars working in international business or international research (Our goal is five.)

Progress 03/01/20 to 02/28/21

Outputs
Target Audience:The target audience reached by our efforts during this reporting period included farmers, crop consultants, Extensioneducators, environmentalists, researchers, educational professionals, software developers, and companies and consumersinterested in optimal fertilizer application. Changes/Problems:The principal change in the DIFM project was its duration. Project personnel received a $4 million USDA-NRCS Conservation Innovation Grant, as part of the On-Farm Trial program, to continue the development of the project's cyber-infrastructure. The DIFM project began collaborating with research personnel at the Oracle corporation to create the concrete structure of a trial design/data processing/crop management recommendation system. Working with Oracle is helping the DIFM project deal efficiently with a challenge it has been struggling with all along, which is to develop that cyber-infrastructure in order to scale up DIFM activities to tens of thousands of trials per year. In previous reporting years, the DIFM project experienced a challenge finding Spanish-speaking American students to train for research in South America and in developing software for the DIFM infrastructure and database. The team was able to recruit Francisco Gamino, who has greatly contributed to DIFM in the last twelvemonths.Francisco built and rebuilt a Virtual Machine (VM), Virtual Cloud Network (VCN), and a file system on a new Oracle tenancy. The purpose of the file system was to archive data and store it for manipulation. We uploaded all of the data and ran scripts on it from the VM. These scripts perform tasks such as parsing data depending on its type and renaming data to follow new conventions. The scripts can also flag harvest and planting files if they do not contain a record of the crop they represent.Franciscoalso worked on online oracle web applications, and on installing Apache on a VM for running other services. What opportunities for training and professional development has the project provided?DIFM continued supporting students involved in the project: Paul Hegedus, Technician and PhD student, Agroecology. Paul has placed 100% of his time into developing his own database and learning how to migrate it into the DIFM Database. He has written computer applications to automate the process of data cleaning, scraping from Google Earth Engine and conducting analysis of on-farm precision experiment data. Sasha Loewen, PhD student, Agroecology. Sasha's research is focused on precision ag technology based trials on crop seeding rates on organic farms, including his own farm in Manitoba, California. Hannah Duff, PhD Student Agroecology. Hannah's research is focused on using precision agriculture trials to parameterize economic models that will allow determining the economic tradeoffs of allowing ecological (non-cropped) areas in field to reap the ecosystem service benefits/problems in Montana dryland wheat production Georgio Morales, PhD Student, Computer Science, Machine Learning. Studies and experiments began in developing an AI-based approach to yield prediction. The first part of the crop yield prediction experiments consisted of the proposal of an end-to-end yield prediction framework that takes advantage of the representational power of Convolutional Neural Networks (CNNs) and the ability of a probabilistic graphical model, such as a Conditional Random Field (CRF), to explicitly modeling dependencies between elements of the output. The proposed Convolutional Neural Network - Conditional Random Field (CNN-CRF) model receives a patch input that corresponds to a small region of the field containing radar images and on-ground data and the output is a patch of the same dimensions containing the estimated yield values of said region. Our CNN-CRF was compared to a CNN designed by us called Hyper3DNet and a simple CRF. The resulting comparison demonstrated that the CNN-CRF performs significantly better than the CRF; however, the results obtained by the CNN-CRF and the CNN models are not significantly different. The second set of experiments consisted of using information from previous years to predict a complete yield map for the next year. We compared the prediction results using some variations of our Hyper3DNet network and different loss functions to those obtained by other models proposed in previous work, specifically other CNN architectures, a stacked autoencoder, an ensemble of neural networks, and a multilinear regression. So far, the results show a consistent superiority of our proposed CNN model over the compared models. Amy Peerlinck, PhD Student, Computer Science, Machine Learning. The second area of study was in advancing our previous work in generating optimized prescription maps for on-farm experiments. In order to make the experimental prescription generator publicly available to researchers and farmers that are part of the CIGOFT process, we have created a website in cooperation with the Oracle Cloud Infrastructure. This preliminary website is available athttp://trialdesign.difm-cig.org/home. In order to create a working, accessible platform on the oracle cloud, several components had to be put in place and made to cooperate with each other: a server to run the application itself on the internet, a separate database server to store and access data, and the application itself, which includes the back-end logic as well as the front-end user interface. This application has been fully deployed and is functioning correctly within the initial scope of the experimental prescription guidelines. Some key functionalities are the creation of a grid fitted to a field boundary, and the options to create several different types of experimental trials: a fully randomized trial, a randomly stratified trial across yield and/or protein bins, a strip trial, and a randomly stratified trial with rate jump minimization between consecutive plots to minimize strain on farming equipment. The latter is accomplished by using a machine learning approach called a Genetic Algorithm, which performs a metaheuristic search across the different options, which therefore results in a different outcome each time this algorithm is run, maintaining a certain level of randomization necessary for appropriate statistical trials. The farmer or researcher using the tool can choose several different options to create these trials, such as plot size, the number of bins to split yield and or protein, how the bins are to be determined, fertilizer rates to apply, etc. Aolin Gong,Ph.D. Optimal length of plot trials for experiment design. University of Illinois, Urbana-Champaign.Aolin has been working on her dissertation chapter over the last twelvemonths, focusing on the nitrogen management in the U.S., as well as in China. Additionally, she worked for the World Bank as a short time consultant for fivemonths, exploring the relationship between the agricultural transformation and digital development. She is now interning for IFAD on supporting projects in rural areas of the world. Brittani Edge,Ph.D. Management zone delineation. University of Illinois, Urbana-Champaign. Jaeswok Hwang,MS. Agricultural Economics and Precision Agriculture. University of Illinois, Urbana-Champaign. Junyuan Li.,MS. Agricultural and Consumer Economics. University of Illinois, Urbana-Champaign. Leyton Brown, MS. Technical Systems Management. University of Illinois, Urbana-Champaign. How have the results been disseminated to communities of interest?Results have been disseminated to communities of interest using various communication platforms, including peer-reviewed publications, conferences, and symposium presentations, Extension publications and events, interviews, news and magazinearticles, quarterly newsletters, a frequently updated website, field days, international and national meetings, public interviewsvia agricultural press, university events, and several agricultural and university workshops. As Principal Investigator, David Bullock has participated in several presentations to expand awareness over the past year, including the following: (*invited speaker) ASA/CSSA/SSSA Annual Meetings, "Symposium--Coupling the Power of Digital Agriculture, Experimental Design, and Modeling" Bullock, D.S. Phoenix, AZ 11/8/20 - 11/11/20 "Overview of the Data-Intensive Farm Management Program" (*Invited Speaker) NAMPO Ag Expo 2020 (Virtual) Bullock, D.S. Pretoria, South Africa 9/9/20 "On-Farm Precision Experimentation in the Data-Intensive Farm Management Program" (*Invited speaker) AAEA Track Session on Precision Agriculture Bullock, D.S. Virtual 9/10/20 "An Economic Evaluation of Site-specific Input Application Rx Maps" 3rd INFER Online Symposium on Agri-Tech Economics for Sustainable Futures Mieno, T., and D.S. Bullock Newbury, UK (virtual) 9/21/20 "Using On-farm Precision Experimentation to Optimise Seed and Nitrogen Fertilizer Rate Management in the Free State, South Africa" 3rd INFER Online Symposium on Agri-Tech Economics for Sustainable Futures Delport, M., D.S. Bullock, et al. Newbury, UK (virtual) 9/21/20 Overview of the Data-Intensive Farm Management Project and Possible Paths for Helena-DIFM Collaboration Helena company Bullock, D.S. Virtual 11/20/20 Empirical Assessment of Interactions among Genetics, Environmental Factors and Managed Inputs in Corn Yield Response AIFARMS Moose, S.P. and D.S. Bullock Virtual 12/17/20 Economic and Ecological Sustainability of Crop Production through On-farm Precision Experimentation Digital AgriTech 2021 Virual Summit Bullock, D.S. Virtual 2/16/21 What do you plan to do during the next reporting period to accomplish the goals?Continue expanding on-farm field trials using the improved DIFM automated infrastructure to scale up the number of trials we can do with more efficiency. We will fully launch the Oracle/DIFM database and create a front-end tool to work with crop consultants and farmers. DIFM researchers and students will continue publications.

Impacts
What was accomplished under these goals? 1. The DIFM team made significant progress in DIFM's IT software and database (led by Keith Curran). In the last twelve months the following was accomplished: Created Oracle Autonomous Data Warehouse Database, Tables, Indexes, and Schema Created MS SQL Enterprise Database Tables, Indexes, Schema. Provisioned and Configured Oracle Cloud Interface (OCI) Tenancy for Collaborating University Partners and Stakeholders Created and Configured Virtual Cloud Network; private and public subdomains Configured User Groups, Roles and Policies for Oracle Cloud Identity Services Established Parameters for Security and Authentication Configured Linux (Red Hat) and Microsoft 2019 Server VM's in the Cloud Installed relevant software, server tools, server roles, Database Server and Access Privileges Participated in Weekly Trial Design Software Meetings reviewing design architecture and schema Created import process analysis for Yield, AsApplied, AsPlanted, Field Boundary, AB Lines Data and SHP File Meta-data Re-Negotiated Oracle Cloud Tenancy Asset allocation of $25,000 for the scope of one-additional year, expiringin January of 2022 Consulted with stakeholders anddevelopers on trial design software features, priority of development Worked with software developers to convert application data store from file-based to database Built Qualtrics survey to establish survey for future project cloud asset requirements, estimated compute resources for estimating budget requirements Evaluated alternative non-cloud-based IT requirements for Project Participated in regular import script developer team meetings with U of I import script development team Purchased Internet Domain (DIFM-CIG.org and.com) for project Configured DNS Records pointing domain to Oracle Cloud Servers Configured subdomains for projected software development module testing (AI.DIFM-CIG.org, TrialDesign.DIFM-CIG.org, Extension.DIFM-CIG.org, DecisionTool.DIFM-CIG.org, Reporting.DIFM-CIG.org) Configured webhosting on servers Established Backup andRetention Policies, Schedule Built Cloud Block Volumes and Attached to Servers Configured VM Remote Desktop Access Configured Firewall Port Access for Applications andThird Party Analysis Tools Hosted regular guidance meetings with Oracle for Research Managed Cloud Asset Costs by implementing budgets and cost usage notifications Managed Cloud Asset Costs by scripting standard bucket storage timelines to move to archive after 30-days Tested RocketChat team collaboration server on Linux Server Participate Regularly in Oracle for Research Webinars 2. The team at the CyberGIS Center for Advanced Digital and Spatial Studies led by ProfessorShaowen Wang continued to update advanced geospatial libraries for the online cyberGIS-based DIFM application:http://afridb.cigi.illinois.edu. CyberGIS infrastructure was improved to enhance the security and performance of the application. The online application software and associated cyberGIS capabilities were documented to help transfer the knowledge to the project team for future development. 3.The prototype of the automated system we envision building has been written using R. First, the system established a consistent data storage structure in Box that any of the team members can easily access. Second, computer programs to design on-farm trials of various forms and designs, collect publicly available soil characteristics (SSURGO) and weather data (Daymet), process harvester and as-applied input data for statistical analysis, conduct statistical and economic analysis to find economically optimal site-specific input recommendation map, and create reports to farmers. All these operations are conducted in a semi-automated fashion where minimal human interventions are necessary to complete the entire process. We are currently working on fully automating it. 4.The Agroecology group (led by Bruce Maxwell and John Sheppard, Computer Science) has been focused on precision technology-based on-farm trials and data analysis resulting from the trials to increase profitability and resilience on Montana dryland small-grain farms. We now have eightfarms enlisted in the project, fourthat are conventionally managed with herbicides and fertilizers and fourthat are organic. Each farm has at least twofields enlisted in the project with trials created with variable-rate technology. The conventional farms are all varying the rate of nitrogen fertilizer and the organic farms are varying the crop seeding rates of the current year wheat and previous year green manure crop. We met with farmers in an annual meeting held virtually in November 2020. We presented the results of trials conducted on their farms since 2015. The farmers also were provided time to talk about their own trials they placed in fields that were not part of our specific research.More details about the Montana On-Field Precision Experiment (OFPE) system can be found at: https://sites.google.com/site/ofpeframework/home. 5.In 2021, DIFM worked with the Bureau of Food and Agricultural Policy (BFAP) in South Africa to conduct five corn and five soybean on-farm precision experiments. DIFM and BFAP are currently collaborating to analyze the data from those trials. 6. We created a Multistate Research Project, titled NC-1210: Frontiers in On-Farm Experimentation.Land Grant Participating States/Institutions:CA, IA, IL, IN, KS, LA, MI, MN, MS, MT, ND, NE, NY, OH, OK, WA, WI. Non-Land Grant Participating States/Institutions:Illinois State University, Iowa Soybean Association, Purdue University, USDA-ARS. Learn more about the project milestones here:NC1210: Frontiers in On-Farm Experimentation - NIMSS. 7. The DIFM team was awarded a $4M grant to continue research at the conclusion of this current USDA grant. The project is entitled "Improving the Economic and Ecological Sustainability of U.S. Crop Production through On-Farm Precision Experimentation" and was awarded from the USDA's Natural Resources Conservation Service (NRCS). 8.The DIFM project worked closely with Austen Omer, the Illinois Farm Bureau's Associate Director of Natural Resource Policy, who persuaded five Illinois farmers to start working with DIFM to run field trials. DIFM also worked with Omer to begin to writing a grant proposal to the Illinois Nutrient Research and Education Council, to secure funding for a research project that uses DIFM field trial methods to study cover-cropping. For more updates on this reporting period, please visit:Data-Intensive Farm Management Project (illinois.edu)

Publications

  • Type: Journal Articles Status: Other Year Published: 2021 Citation: Hoselton, G.S.W. and M.A. Boerngen. 2021. Illinois corn farmers adoption of best management practices in response to nutrient loss concerns. Journal of Environmental Quality.
  • Type: Journal Articles Status: Awaiting Publication Year Published: 2021 Citation: Hoselton, G.S.W. and M.A. Boerngen. 2021. Farmers awareness of and concerns about nutrient loss. Journal of Soil and Water Conservation.
  • Type: Conference Papers and Presentations Status: Other Year Published: 2020 Citation: Boerngen, M.A. 2020. Illinois corn farmers perspectives on nutrient loss. Greenleaf communities healthy soils for healthy waters webinar series: Agricultural management practices and data usage for soil health. September 8, 2020, virtual. (National Audience).
  • Type: Conference Papers and Presentations Status: Published Year Published: 2020 Citation: Barbosa A.O., N. Hovakimyan and N.F. Martin. 2020. Risk-averse optimization of crop inputs using a deep ensemble of convolutional neural networks. Computer and Electronics in Agriculture.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2020 Citation: Alesso C.A., R. G. Trevisan and N.F. Martin. 2020. On-farm experimentation with high-resolution technologies. American Society of Agronomy Annual Meeting, Phoenix, AR.
  • Type: Journal Articles Status: Published Year Published: 2020 Citation: Barbosa A.O., N. Hovakimyan and N.F. Martin. 2020. Risk-averse optimization of crop inputs using a deep ensemble of convolutional neural networks. Computer and Electronics in Agriculture. https://doi.org/10.1016/j.compag.2020.105785.
  • Type: Journal Articles Status: Published Year Published: 2021 Citation: Trevisan, R.G., D.S. Bullock and N.F. Martin. 2021. Spatial variability of crop responses to agronomic inputs in on-farm precision experimentation. 2021. Precision Agriculture Journal. https://doi.org/10.1007/s11119-020-09720-8.
  • Type: Journal Articles Status: Published Year Published: 2021 Citation: Gardner, G., T. Mieno and D.S. Bullock. 2021. An economic evaluation of site-specific input application Rx maps: Evaluation framework and case study. Precision Agriculture (2021). https://doi.org/10.1007/s11119-021-09785-z.
  • Type: Journal Articles Status: Published Year Published: 2021 Citation: Mandrini, Germ�n, David S. Bullock and Nicol�s F. Mart�n. 2021. Modeling the economic and environmental effects of corn nitrogen management in Illinois. Field Crops Research 261(2021). https://doi.org/10.1016/j.fcr.2020.108000.
  • Type: Journal Articles Status: Published Year Published: 2021 Citation: Trevisan, Rodrigo G., David S. Bullock and Nicol�s F. Mart�n. 2021. Spatial variability of crop responses to agronomic inputs in on-farm precision experimentation. Precision Agriculture 22(2021): 342-363. https://doi.org/10.1007/s11119-020-09720-8.
  • Type: Journal Articles Status: Published Year Published: 2021 Citation: Paudel, J and Crago, C.L. 2021. Environmental externalities from agriculture: Evidence from water quality in the united states. American Journal of Agricultural Economics, 103(1), 185-210.
  • Type: Journal Articles Status: Submitted Year Published: 2021 Citation: Paudel, J. and Crago, C.L. 2021. Agricultural adaptation to climate change: Implications for fertilizer use and water quality in the United States. Journal of the Association of Resource and Environmental Economists.
  • Type: Journal Articles Status: Published Year Published: 2020 Citation: P. Paccioretti, M. C�rdoba and M. Balzarini. 2020. FastMapping: Software to create field maps and identify management zones in precision agriculture. Volume 175, 2020, 105556, ISSN 0168-1699, https://doi.org/10.1016/j.compag.2020.105556. Computers and Electronics in Agriculture.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2021 Citation: Md. Asaduzzaman Noor and John W. Sheppard. 2021. Evolutionary grain-mixing to improve profitability in farming winter wheat. In: Applications of Evolutionary Computation, Lecture Notes in Computer Science, LNCS 12694, Springer, 2021, pp. 113-129.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2021 Citation: Giorgio Morales, John Sheppard, Riley Logan and Joseph Shaw. 2021. Hyperspectral band selection for multispectral image classification with convolutional networks. To appear in Proceedings of the IEEE International Joint Conference on Neural Networks (IJCNN), July 2021.
  • Type: Journal Articles Status: Published Year Published: 2021 Citation: Bullock, David S., Taro Mieno and Jaeseok Hwang. 2021. The value of conducting on-farm field trials using precision agriculture technology: A theory and simulations. Precision Agriculture 21(2020): 1027-1044. DOI: 10.1007/s11119-019-09706-1.


Progress 03/01/19 to 02/29/20

Outputs
Target Audience:The target audience reached by our efforts during this reporting period included farmers, crop consultants, Extension educators, environmentalists, researchers, educational professionals, software developers, and companies and consumers interested in optimal fertilizer application. Changes/Problems:In the last reporting period, the DIFM project experienced a difficult time in finding Spanish-speaking American students to train for research in South America and in developing software for the DIFM infrastructure and database. We have successfully overcome both of these challenges in the 2019 year by recruiting student Francisco Gamino and developing a relationship/collaboration with Oracle Corporation (detailed in the above). Weather was the main problem for the project in the 2019 year. Weather was not ideal and excessive rainfall greatly affected our ability to complete field trials as designed. We did receive data from several trials, however, not as many as the project planned for. Receiving a no-cost extension on the DIFM project will allow us to generate more data by completing another year of on-farm field trials. What opportunities for training and professional development has the project provided?DIFM continued to support the fiveundergraduate FAB Fellow students and threegraduate students in the 2019 year. In addition to the students previously supported, DIFM recruited and offered support to undergraduate student Francisco Gamino. Gamino is studying Computer Science + Crop Science at the University of Illinois and has over three years of experience working as an independent contractor in programming in web development. Gamino can speak, read, and write in English and Spanish fluently and has been a tremendous asset to the DIFM project working with Oracle Corporation to build the DIFM database system. Emanuel Hernandez Cornejo, an undergraduate student from Zamorano University in Honduras, completed a fifteen-week internship with the DIFM project in Spring of 2020. Emanuel primarily worked with the other undergraduate and graduate students to learn about our precision agriculture technology research methods, data management, and assist in various tasks related to the management of DIFM trials. Additional students participating in the DIFM project who have received opportunites for training and professional development: George Hoselton, MS. Human subjects research in farm management. Illinois State University. (Dr. Boerngen). December 2017-May 2019. Kelsey Schreiber, MS. FAB Fellow. International engineering efforts targeting water and irrigation systems to increase rural water access. University of Illinois, Urbana-Champaign. (Dr. Rodriguez). August 2017-May 2019. Brittani Edge, Ph.D. Management zone delineation. University of Illinois, Urbana-Champaign. (Dr. Bullock). August 2016- May 2020. Aolin Gong, Ph.D. Optimal length of plot trials for experiment design. University of Illinois, Urbana-Champaign. (Dr. Bullock). August 2016-May 2020. Rodrigo Trevisan, Ph.D. Crop Sciences. University of Illinois, Urbana-Champaign. (Dr. Nicolas Martin). August 2018-May 2021. Felippe Karp, MS. Agroecology and Sustainable Agriculture (Research focused on precision agriculture). Louisiana State University (Dr. Luciano Shiratsuchi). January 2019-December 2020. Jaeswok Hwang, MS. Agricultural Economics and Precision Agriculture. University of Illinois, Urbana-Champaign. (Dr. David Bullock). August 2018-May 2020. Amy Peerlinck, MS. Computer Science - Machine Learning. Spring 2017-May 2019. Paul Hegedus, PhD. Addressing the ecological effects of nitrate pollution from agricultural fields by investigating nitrate loss and crop response to variable rate nitrogen fertilizer management on a subfield scale. (Dr. Bruce Maxwell and Dr. Stephanie Ewing). January 2018 - April 2022. Murilo Martins, PhD. Precision Agriculture. (Dr. Luciano Shiratsuchi). January 2018 - December 202 How have the results been disseminated to communities of interest?Results have been disseminated to communities of interest using various communication platforms, includingpeer-reviewed publications, conferences, and symposium presentations, Extension publications and events, interviews, news and magazine articles,quarterly newsletters, a frequently updated website, field days, international and national meetings, public interviews via agricultural press, university events, and several agricultural and university workshops. As Principal Investigator, David Bullock has participated in several presentations to expand awareness over the past year, including the following: Title: The Value of Conducting On-Farm Field Trials Using Precision Agriculture Technology. Audience/Group: Big Ag Data Conference.Location: Davis, California. Title: Estimating Yield and Water-Quality Response Functions Using On-Farm Precision Experimentation, Spatially-Intense Soil Sampling, and Hyperspectral Imagery.Audience/Group: University of Illinois Center for Digital Agriculture Workshop. Location: Urbana, Illinois. Title: Using Precision Technology to Conduct On-Farm Research Trials for Data-Intensive Farm Management.Audience/Group: 56th Annual Illinois Corn Breeders' School. Location: Champaign, Illinois. Title: Working with the Data-Intensive Farm Management Project. Audience/Group: Christiansen Land & Cattle Company. Location: Remote. Upcoming: Invited speaker at ASA/CSSA/SSSA Annual Meetings. Symposium-Coupling the Power of Digital Agriculture, Experimental Design, and Modeling.Location: Phoenix, Arizona. What do you plan to do during the next reporting period to accomplish the goals?1) The project recently recruited a few students who fit the qualifications, and will make a significant impact, in the FAB program. A portion of the unobligated funds will be used to support these students through their graduation in the Spring of 2021. The education and training of the students in the FAB program is essential in the project's goal to improve agricultural research and agribusiness collaborations between the U.S. and Latin America. 2) Working with Oracle, we plan to complete the software development and data analysis. 3) Continue expanding on-farm field trials and improving the DIFM system to create an infrastructure that is capable of running hundreds of trials each year, all over the world.

Impacts
What was accomplished under these goals? 1. From 2014 to 2018 the Advanced Ag Alliance (AAA), a non-profit organization consisting of farmers in New York state, had performed hybrid and variable rate seed experiments in 65 corn and soybean fields with 2,800 acres scale total. Researchers in the DIFM project at Cornell University have evaluated the performance of this trial by applying their own statistical model and provided farmers with an optimal seeding rate recommendations for each field. However, AAA was not satisfied with theanalysis results since their profits in the following year with this recommended planting werenot improved. For this reason, AAA asked DIFM to re-evaluate their VR seed experiments and set a new 2019 trial under the DIFM model. By investigating the AAA model which was designed by the project team at Cornell, DIFM found that their econometric model had derived optimal seed rate by not taking appropriate profit (net revenue) maximization process into account. By applying the DIFM analysis model to re-evaluate the experiments for four years (2014~2018), we found the significant overseeding problems through almost all soybean and corn fields. In designing 2019 VR seed trials, DIFM set totally randomized checkerboard trials for AAA to compare the performance of VR seeding rate with farmer's status quo planting rates drawn from their historical experiences. 2. In the 2019 year, we expaned and conducted the following on-farm field trials: Argentina (30 - in collaboration with Clarion, the Faculty of Agricultural Sciences at the National University of Cardoba, and Cresud, S.A.), Illinois (24), Brazil (2 - In collaboration with EMBRAPA), Louisiana (3), Montana (6), Nebraska (8), New York (35 - in collaboration with Advanced Ag Alliance), Ohio (2), South Africa (4 - in collaboration with the University of Pretoria and the University of KwaZulu-Natal), Texas (1), and Washington (8). 3. In the 2019 year, DIFM began collaborating with Oracle Corporation, an American multinational computer technology corporation that sells database software and technology, cloud engineered systems, and enterprise software products. Oracle gifted the project one year of Phd funding and is working with DIFM to accomplish the following milestones: 1) Integrate DIFM's current data-cleaning R programs into the Oracle system, 2) Develop basic components for uploading the raw datasets from several DIFM trials, including as-planted, as-applied, yield, electroconductivity, LiDAR elevation, Sentinel 2 NDVI data, and other spatially-oriented data, 3) Develop algorithms for recognizing data, sorting it, and assigning common names to label as appropriate data,4) Build a system with which different types of data can be efficiently queired and received, and5) Write and submit two academic journal articles, one describing the cyber-infrastructure created and its useand another describing research findings on the value-of-electroconductivity data. 4. Data workflows between researchers and farmers were created, along with methods of analysis for on-farm research and landscape analysis at the continental scale. A significant amount of effort was invested spatially integrating dataset from on-farm research from more twenty-crop production fields participating in the DIFM project from the 2017, 2018 and 2019 seasons. These efforts resulted in new R code libraries, spatial databases in PostGIS, abd repositories in R and Python code in GitHub and Bitbucket. Because of this effort, three publications were published in academic journals and twostudies published in peer-review conference proceedings.

Publications

  • Type: Journal Articles Status: Published Year Published: 2020 Citation: Barbosa A.O., R.G., Trevisan, N. Hovakimyan and N.F. Martin. 2020. Modeling Yield Response to Crop Management Using Convolutional Neural Networks. Computer and Electronics in Agriculture (170): 105197.
  • Type: Journal Articles Status: Published Year Published: 2019 Citation: Bullock, David S., Maria Boerngen, Haiying Tao, Bruce D. Maxwell, Joe D. Luck and Nicolas Martin. 2019. The Data-Intensive Farm Management Project: Changing Agronomic Research through On-Farm Experimentation. Agronomy Journal 111(2019): 725-735.
  • Type: Journal Articles Status: Published Year Published: 2020 Citation: Bullock, David S., Taro Mieno and Jaeseok Hwang. 2020. The Value of Conducting On-Farm Field Trials Using Precision Agriculture Technology: A Theory and Simulations. Precision Agriculture.
  • Type: Journal Articles Status: Published Year Published: 2019 Citation: Marks, B. and M.A. Boerngen. 2019. A Farming Communitys Perspective on Nutrient Loss Reduction. Agricultural & Environmental Letters. 4:190004.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2019 Citation: Martin, N.F., D.S. Bullock and R.G. Trevisan. 2019. Spatial Variability of Crop Responses to Nitrogen in On-Farm Precision Experimentation. 2019 Nitrogen Use Efficiency Workshop.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2019 Citation: Martin, N.F., D.S. Bullock and R.G. Trevisan. 2019. Understanding the Spatial Variability of Optimum Nitrogen Rates Using Remote Sensing and on-Farm Precision Experimentation. 2019 ASA-CSSA-SSSA International Annual Meeting.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2019 Citation: Martin, N.F., D.S. Bullock and R.G. Trevisan. 2019. Deep Learning to Predict Optimum Crop Management Decisions. 2019 ASA-CSSA-SSSA International Annual Meeting.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2019 Citation: Martin, N.F., L. Shiratsuchi, B. Maxwell, J.D. Luck, H. Tao, M. Boerngen and D.S. Bullock. 2019. Changing Agronomic Research Through On-Farm Precision Experimentation. Symposium Making the Most of On-Farm Trials Using Spatial Statistics. 2019 ASA-CSSA-SSSA International Annual Meeting.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2019 Citation: Martin, N.F., L. Shiratsuchi and R.G. Trevisan. 2019. Cotton Yield Monitor Values Drift Over Time. Cotton Agronomy, Physiology and Soil Conference. 2019 Beltwide Cotton Conference.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2019 Citation: Paccioretti P., M. C�rdoba, C. Bruno, D.S. Bullock and M. Balzarini. 2019. Statistical Modeling For On-Farm Experimentation Using Precision Agricultural Technology. Precision Agriculture 2019. Proceedings of the 12th European Conference on Precision Agriculture.
  • Type: Journal Articles Status: Published Year Published: 2019 Citation: Tao, Haiying and David S. Bullock. 2019. Using Digital Agriculture Technologies to Improve Nitrogen Management and Wheat Yield. Cereal Foods World 64(6).
  • Type: Journal Articles Status: Published Year Published: 2020 Citation: Trevisan, Rodrigo G., David S. Bullock and Nicol�s F. Mart�n. 2020. Spatial Variability of Crop Responses to Agronomic Inputs in On-Farm Precision Experimentation. Precision Agriculture.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2019 Citation: Bullock D.S. and T. Mieno. 2019. The Economic Value of On-Farm Experimentation. Precision Agriculture 2019. Proceedings of the 12th European Conference on Precision Agriculture. pp. 1030.


Progress 03/01/18 to 02/28/19

Outputs
Target Audience:The target audience reached by our efforts during this reporting period included farmers, crop consultants, Extension personnel, environmentalists, researchers, educational professionals, and companies interested in optimal fertilizer application. Changes/Problems:1. Brendan Kuhns, agraduate research assistantat the University of Illinois pursuing his master's degree in the Department of Agricultural and Biological Engineering, will be graduating in May 2019 and has accepted a position with John Deere. Brendan has greatly contributed to cotton research within the DIFM project and has been our main technical support when it comes to troubleshooting equipment and machinery issues with our trials. 2. The project has experienced adifficult time finding Spanish-speaking American students to train for research in South America. Our approach for the next year is to allocate more funds into recruitment for Spanish speaking American students. 3. Another problem we faced waswith the National Center for Supercomputing Applications (NSCA) at the University of Illinois. We hoped to work with NCSA to develop software. In the middle of our development, our main software engineer at NSCA left for a job in San Francisco,which set us back in terms of software creation and forced us to find new computer programmers to work with the project and help create software platforms. What opportunities for training and professional development has the project provided?DIFM supported 5 undergraduateFAB Fellow students in the 2018 year, under the direction and work of Luis Rodriguez. We currently estimate that we have space to recruit additional undergraduate students in Fall 2019 for approximately 3 semesters of support. At the graduate level, we supported one master's student who is approaching graduation in 2019. In January of 2018, a second master's student was offered and accepted support, and thus just started their program. One offer is pending to a PhD student, and depending on the success or failure of this recruitment, we have two potential candidates in waiting. Students participating in the DIFMproject who have received opportunites for training and professional development: George Hoselton, MS. Human subjects research in farm management. Illinois State University. (Dr. Boerngen). December 2017-May 2019. Kelsey Schreiber, MS. FAB Fellow. International engineering efforts targeting water and irrigation systems to increase rural water access. University of Illinois, Urbana-Champaign. (Dr. Rodriguez). August 2017-May 2019. Brittani Edge, Ph.D. Management zone delineation. University of Illinois, Urbana-Champaign. (Dr. Bullock). August 2016-May 2020. Alexandre Ormiga G. Barbosa, Ph.D. A software tool for field trials setup with an optimal path generation for complex terrain field coverage. University of Illinois, Urbana-Champaign. (Dr. Hovakimyan). August 2017-May 2020. Aolin Gong, Ph.D. Optimal length of plot trials for experiment design. University of Illinois, Urbana-Champaign. (Dr. Bullock). August 2016-May 2020. German Mandrini, Ph.D. Optimal stopping time for field trial design. University of Illinois, Urbana-Champaign. (Dr. Bullock). June 2017-May 2020. Laila Puntel, Ph.D. Trial implementation, data collection, and data analysis in Argentina. Iowa State University. (Dr. Archontoulis). August 2016-May 2020. Robert James Reis, Ph.D. Development of a sensor system to measure soil type and compaction. University of Illinois, Urbana-Champaign. (Dr. Tony Grift). June 2017-May 2019. Brendan Kuhns, MS. On-the-go soil physical properties characterization using acoustic emission detection. University of Illinois, Urbana-Champaign. (Dr. Tony Grift) January 2018-May 2019. Grace Nystrom, BS. Computer Science and Crop Science. University of Illinois, Urbana-Champaign. August 2018-May 2022. Rodrigo Trevisan, Ph.D. Crop Sciences. University of Illinois, Urbana-Champaign. (Dr. Nicolas Martin). August 2018-May 2021. Felippe Karp, MS. Agroecology and Sustainable Agriculture (Research focused on precision agriculture). Louisiana State University (Dr. Luciano Shiratsuchi). January 2019-December 2020. Joshua Babes,BS.ACE Finance in Agri-Business.University of Illinois, Urbana-Champaign. August 2017-May 2021. Jaeswok Hwang,MS. Agricultural Economics and Precision Agriculture. University of Illinois, Urbana-Champaign. (Dr. David Bullock). August 2018-May 2020. Amy Peerlinck,MS. Computer Science - Machine Learning. Spring 2017-May 2019. Paul Hegedus, PhD.Addressing the ecological effects of nitrate pollution from agricultural fields by investigating nitrate loss and crop response to variable rate nitrogen fertilizer management on a subfield scale. (Dr. Bruce Maxwell and Dr. Stephanie Ewing). January 2018 - April 2022. Murilo Martins, PhD. Precision Agriculture. (Dr. Luciano Shiratsuchi). January 2018 - December 2021. How have the results been disseminated to communities of interest?Results have been disseminated to communities of interest using various communication platforms. Results were disseminated via peer-reviewed publications, conferences, and symposium presentations, Extension publications and events, quarterly newsletters, a website, field days, meetings, public interviews via agricultural press, university events, and several workshops. As Principal Investigator, David Bullockhas participated in several presentationsto expand awareness over the past year, including the following: Title:Data-Intensive Farm Management: On-farm Field Trials Using Precision Agriculture Technology.Audience/Group:U of I Data Science Day Lightening Talk.Location:Urbana, IL. Title:El Valor de la Infromacion de Ensayos Agronomicos de Gran Escala.Audience/Group:First Latin American Conference on Precision Agriculture.Location:Santiago, Chile. Title:Some Preliminary Results from the 2017 DIFM Field Trials.Audience/Group:U of I ACE Commercial Ag Seminar.Location:Urbana, IL. Title:Overview of the Data-Intensive Farm Management Project.Audience/Group:San Patricio County Crops Committee.Location:Sinton, TX. Title:Overview of the Data-Intensive Farm Management Project.Audience/Group:Nueces Country Crops Task Force.Location:Robstown, TX. Title:Large-Scale, On-farm Field Trials: Opportunities for Collaboration.Audience/Group:Texas A&M AgriLife Research and Extension Center.Location:Corpus Christi, TX. Title:Overview of the Data-Intensive Farm Management Project.Audience/Group:ACES Office of Corporate Engagement (AGCO).Location:Champaign, IL. Title:Using Precision Agriculture Technology in On-farm Field Trials to Enable Data-Intensive Fertilizer Management.Audience/Group:USDA-NIFA-AFRI Project Directors' Meeting.Location:Washington, DC Title:El Proyecto DIFM: Usando la Tecnologia de Precision para Conducir Ensayos Economicos de Gran Escala.Audience/Group:17o Curso Internacional de agricultura y Ganaderia de precision.Location:Manfredi, Argentina Title:A Better Model of Nutrient Loss and Yield in the Mississippi River Watershed.Audience/Group:2018 Illinois Nutrient Loss Reduction Strategy Workshop.Location:Urbana, IL. Title:Using Precision Technology to Conduct On-Farm Research Trials for Data-Intensive Farm Management.Audience/Group:2018 Illinois Ag Masters Conference (conference of Certified Crop Advisors).Location:Springfield, IL. Title:Using Precision Technology to Conduct On-Farm Research Trials for Data-Intensive Farm Management.Audience/Group:Delegation from the Ministry of Agriculture and Rural Affairs, China.Location:Urbana, IL. Title:The Value of Conductin On-farm Field Trials Using Precision Agriculture Technology: A Theory and Simulations.Audience/Group:All-DIFM Annual Meeting.Location:Iowa City, IA. Title:Input Choice under Risk: What's the Optimal Choice When the Farmer Doesn't Know the Weather?Audience/Group:All-DIFM Annual Meeting.Location:Iowa City, IA. Title:El Proyecto DIFM: Usando la Tecnologia de Precision para Conducir Ensayos Economicos de Gran Escala.Audience/Group:Fermers, Ag Scientists, and Technical Ag Educators.Location:Menteria, Columbia. Title:Nitrogen Loss and Water Quality Research in the Data-Intensive Farm Management Project.Audience/Group:U of I ACE PERE Workshop.Location:Urbana, IL. What do you plan to do during the next reporting period to accomplish the goals?Objective 1:Increase amount of trials in South America.We plan to work with the Food and Agriculture Organization of the United Nations (UNFAO) to increase the amount of corn and cottontrials in South America. Objective 2: Createsoftware capable of automatingthe data analyzationprocess to quickly and efficiently produce management reccommendations.We want to create an infrastructure so that thousands of trials can be run and analyzedper year. Software development will be our main focus for the next year.

Impacts
What was accomplished under these goals? Objective 1:Generate and gather data using precision agriculture technology to run coordinated long-term, large-scale, on farm field trials. 1. Last year we removed the University of Kentucky from the project and replaced the subawardee with Montana State University. By partnering with MSU, led by Dr. Bruce Maxwell, we have expanded our field trials andincreased the amount of data generated. DIFM successfully generated and gathereddata from 6 different field trialsin Nebraska for 2018. In addition to generating an increased amount of data in new locations, our collaboration with MSU has led to significant software developments. The AG Data Management Tool Development Team, led by Dr. Mary Ann Cummings at MSU, is developing a box repository (with an interface) for farmers to input their data files. In addition, the team is also developing a single PostGIS database (with an interface) for researchers to use for cleaning and analysis purposes. The first release of this software included the interface to the Box repository and the schema for the PostGIS database. The second release is scheduled to be completed in May 2019 and will include a single server database and Box repository. 2. DIFMbegan workingwith the Advanced Ag Alliance (AAA), a non-profit organization dedicated to running on-farm whole-field agronomic trials, to design approximately 20 field trials in New York State between 2019-2020 as well as analyzing and reporting on data from the New York field experimentsbetween 2015-2018. In addition to providing DIFM with additional trial locationsand data in the New York State area, the Advanced Ag Alliance is funding graduate student, Jaeseok Hwang. Hwang will be traveling to New York State and driving the Verismachine to collect Electro Conductivity(EC) data in the New York fields this upcoming Spring, research that will enable in-depth study of interaction between soil characteristics and nitrogen fertilizer. DIFM's work with AAA has provided an opportunity forgathering assorted data and excellent student training and involvement within the DIFM project. 3. In 2018, the Data-Intensive Farm Management Project has generated and collected electroconductivity, pH, and organic matter data using the Veris U3 machine on approximately 1,205 acres. 4. DIFM and the University of Illinois signed an agreement for collaborative research with the National University of Cordoba, Argentina. Objective 2:Estimate the relationship between fertilizer use and water quality. 1. With the Illinois Corn Growers Association we have successfullygathered 2018 data from theField Lab and this will be our first year analyzing the data from the nutrient loss monitoring system. Objective 3: Expand awareness of optimal fertilizer management practices to growers. 1. Researcher from Illinois State University, Maria Boerngen, and graduate student, George Hoselton, conducted a professional research survey with the help of the Illinois Corn Growers Association (ICGA). This survey sent out to their membership base helped us gain a more detailed understanding of how farmers view and are responding to concerns about nutrient loss, and their willingness to voluntarily comply with nutrient loss reduction goals. This survey was designed using theInternet, Phone, Mail, and Mixed-Mode Surveys: The Tailored Design Methodhandbook (Dillman, Smyth, and Christian, 2014). The survey by Boerngen and Hoselton was sent out the first week of July 2018 and reached roughly 3,850 members of the ICGA and was closed the first week of November 2018. The survey received a 19.9% response rate, which is a high response rate for external surveys. Key findings of the survey include: Over 65% of respondents indicated that they are very familiar or somewhat familiar with the Illinois Nutrient Loss Reduction Strategy. 80% of the respondents indicated that they are very concerned or somewhat concerned about nutrient loss. The majority of respondents (90.9%) believe that nutrient loss negatively impacts the environment, and 88.1% are very or somewhat concerned about the implementation of regulation because of nutrient loss. Nearly 80% of respondents had already made changes in their farming practices due to nutrient loss. There are currently two manuscripts in preparation, one for the Journal of Soil and Water Conservation and one for the Journal of Environmental Quality. 2. DIFM was awarded the Interdisciplinary Collaboration in Extension (ICE) grant: "Putting UIE at the Forefront of the Coming Data-Intensive Farm Managment Revolution: A Tool to Help Farmers Turn On-Farm Experiments into Profitable Decisions". PI/Co PIs: Bowman, N., D.S. Bullock, N. Martin, S. Wang, P. Alberti, T.M. Becker, R. Higgins, and J. Soule. Through this grant DIFM hascollaborated with Extension and Certified Crop Advisors (CCAs)in efforts to expand field trials and set a communication pathway to farmers. Each CCA will find two farmers to participate in field trials, adding approximately 10 more field trials to the 2019 year. CCAs work closely with each of their farmer participantsand provide support throughout their involvement of the project.

Publications

  • Type: Journal Articles Status: Published Year Published: 2019 Citation: Rodriguez, Divina Gracia & S. Bullock, David & Boerngen, Maria. (2019). The Origins, Implications, and Consequences of Yield-Based Nitrogen Fertilizer Management. Agronomy Journal. 10.2134/agronj2018.07.0479.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2019 Citation: Trevisan, R.G., D.S. Bullock, and N.F. Martin. 2019. Site-specific treatment responses in on-farm precision experimentation. Preprints. doi: 10.20944/PREPRINTS201902.0007.V1.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2019 Citation: Trevisan, R.G.; Shiratsuchi, L..S.; Bullock, D.S.; Martin, N.F. Improving Yield Mapping Accuracy Using Remote Sensing. Preprints 2019, 2019010287 (doi: 10.20944/preprints201901.0287.v1).


Progress 03/01/17 to 02/28/18

Outputs
Target Audience:The target audience included farmers, crop consultants, Extension personnel, environmentalists, and companies interested in optimal fertilizer application. Changes/Problems:Dr. Donald Bullock retired from the University of Illinois and the Data-Intensive Farm Management project in 2017. We have since brought on Dr. Nicolas Martin, an assistant professor in Crop Sciences at the University of Illinois, Dr. Bruce Maxwell, a professor of Agroecology and Applied Plant Ecology at Montana State University, and Dr. Haiying Tao, an assistant professor of Soil Fertility and Residue Management at Washington State University. The University of Kentucky has been removed from the project as a subawardee, to be replaced by Montana State University in 2018. Dr. Maxwell heads the On-Farm Precision Experiments project at Montana State University, and has significant experience with on-farm research experience. By partnering with MSU, we hope to take advantage of the overlap in our research interests by expanding the data generated in this project. What opportunities for training and professional development has the project provided?Students participating in this project We have six undergraduate students andnineteengraduate students working on the project (twelvePhD andsevenMS). R.J. Baldwin, BS. Soil sample data and pre-processing as-applied data and harvest data. University of Nebraska, Lincoln. (Dr. Luck) August 2017-May 2018. RaaginiGupta, BS. FAB Fellow, studying computer science, natural resources, and environmental science. University of Illinois, Urbana-Champaign. (Dr. Rodriguez). August 2017-May 2020. Aiden Kamber, BS. Big data and agriculture.University of Illinois, Urbana-Champaign. (Dr. Rodriguez). August 2017-May 2020. Elisa Kim, BS. Flow cytometry.University of Illinois, Urbana-Champaign. (Dr. Rodriguez). August 2017-May 2020. Benjamin Marks, BS. Human subjects research and interview design. Illinois State University. (Dr.Boerngen) August 2016-May 2017. John Winkler, BS. GIS tostudyviable plots for biomass crops along state roadways. August 2017-May 2020. Nidhi Adhikari, MS. Hydrology model for evaluating drainage systemresponseto future climate scenarios.University of Illinois, Urbana-Champaign. (Dr.Cooke).May 2016-May 2018. ZhengzhengGao, MS. Spatial data management.University of Illinois, Urbana-Champaign. (Dr.Mieno). August 2017-May 2019. Paula Giron, MS. Cleaning yield and electroconductivity maps, designing field experiments.University of Illinois, Urbana-Champaign. (Dr.Bullock).September 2017-December 2020. George Hoselton, MS. Human subjects research in farm management.University of Illinois, Urbana-Champaign. (Dr.Boerngen).December2017-May 2019. Shreya Khurana, MS. Statistical data analysis of crop and field trials.University of Illinois, Urbana-Champaign. (Dr.Bullock). August 2017-May 2018. Kelsey Schreiber, MS. International engineering efforts targeting water andirrigation systems to increase rural water access.University of Illinois, Urbana-Champaign. (Dr. Rodriguez). August 2017-May 2019. Gabe Stoll, MS. Coordination of data collection with four Nebraskan collaborators.University of Nebraska, Lincoln. (Dr.Luck).June2017-December2018. Brittani Edge, Ph.D. Management zone delineation.University of Illinois, Urbana-Champaign. (Dr.Bullock). August 2016-May 2020. AlexandreOrmigaG. Barbosa,Ph.D.A software tool for field trials setupwith an optimal path generation for complex terrain field coverage. University of Illinois, Urbana-Champaign. (Dr.Hovakimyan). August 2017-May 2020. AolinGong, Ph.D. Optimal length of plot trials for experiment design.University of Illinois, Urbana-Champaign. (Dr.Bullock). August 2016-May 2020. Hao Hu, Ph.D. Spatial regression models regarding the relationship between crop fieldand farm management practice.University of Illinois, Urbana-Champaign. (Dr.Wang). August 2017-May 2018. WilmanIglesias, Ph.D. Models toevaluate variable application rate payoff. University of Nebraska, Lincoln. (Dr.FulginitiandDr. Perrin).August 2017-May 2018. GermanMandrini, Ph.D. Optimal stopping time for field trial design.University of Illinois, Urbana-Champaign. (Dr.Bullock).June2017-May 2020. JayashPaudel, Ph.D. Fertilizer effect on watershed level. University of Massachusetts, Amherst. (Dr. Crago). May 2016-December 2018. LailaPuntel, Ph.D. Trial implementation, data collection, and data analysisin Argentina.Iowa State University. (Dr.Archontoulis). August 2016-May 2020. Pedro Queiroz, Ph.D.Models toevaluate variable application rate payoff. University of Nebraska, Lincoln. (Dr. Fulginitiand Dr. Perrin).August 2017-May 2018. Robert James Reis, Ph.D. Development of a sensor system to measure soil type and compaction.University of Illinois, Urbana-Champaign. (Dr.Grift).June2017-May 2019. Felipe Silva, Ph.D.Models toevaluate variable application rate payoff. University of Nebraska, Lincoln. (Dr. Fulginitiand Dr. Perrin).August 2017-May 2018. Brandon Tate, Ph.D.Soil sensing technologies.University of Illinois, Urbana-Champaign. (Dr. Rodriguez). August 2017-May 2020. How have the results been disseminated to communities of interest?Results were disseminated viapeer-reviewed publications, conferences and symposium presentations, Extension publications, quarterly newsletters, a website, phone calls, field days, spring talk meetings, workshops, and field days. What do you plan to do during the next reporting period to accomplish the goals?Objective 1: Generate and gather data using precision agriculture technology to run coordinated long-term, large-scale, on-farm field trials. 1. At present, we currently estimate that in 2018we will run 36 trials in Argentina, 8 trials in Brazil, 3 in Colombia, 2 in Canada, 6 in Montana, 25 in Washington, 10 in New York, 1 in Idaho, 12 in Iowa, 2 in Kansas, 5 in Nebraska, 2 in Ohio, 3 in Indiana, 20 in Illinois, and 3 in Louisiana. Our research partnerships with Clarion in Argentina, Embrapa in Brazil, the Ontario Ministry of Agriculture, Grain Growers of Ontario, Guelph University in Canada, the Food and Agriculture Organization of the United Nations in Colombia, the New York Corn and Soybean Growers Association, and three crop consultants, as well as with Washington State University, Louisiana State University, and Montana State University have allowed us to expand our research more quickly than anticipateddue to a reduction in time spent in recruitment. 2. The plan for the year is to complete the EVSI model for evaluating expected payoff from VAR and apply it for EC and other soil signals on data from as many project fields as possible from the first two years of experimental results. Objective 2: Estimate the relationship between fertilizer use and water quality We will refine preliminary estimates of the impact of changes in fertilizer use on water quality at the watershed level. We will present results of the analysis at various academic conferences, as well as collaborators on this NIFA grant. Based on feedback we receive, we may collect additional data or make changes to our analysis.We will continue data collection that will allow us to estimate the relationship between temperature and rainfall and fertilizer use by US farmers. Once data collection is complete, we will conduct empirical analysis. Objective 3: Provide training to students who speak Spanish or Portuguese. Despite current successes in recruitment of Spanish- and Portuguese-speaking students, in the final year of the project, we will intensify our efforts to bring on additional Spanish-speaking undergraduate students (from Puerto Rico in particular). Objective 4: Expand awareness of optimal nitrogen management practices. 1. We will continue our schedule of Extension activities and workshops with farmers to continue spreading awareness of optimal nitrogen management practices. 2. In November 2019, we will produce a Data-Intensive Farm Management software training course that can be used by Extension educators and consultants to help farmers implement field trials on their farms. We will develop and conduct a course to train Extension educators and consultants to use our software and we plan to train about 10 people at the University of Illinois and the University of Nebraska, Lincoln. We plan to extend this training course, as well as the DIFM software, to university-based Extension educators and private consultants who will use it to help farmers make better management choices. 3. The project's results and discoveries will be distributed through a variety of both research and Extension materials, including traditional and web-based media, workshops, and presentations. Extension materials will take several forms, ranging from written circulars or workshops to innovative web-based videos and interactive media. The goal is to make this informal educational material as accessible as possible to producers, private consultants, agricultural product retailers, dealers, and other Extension professionals. The intention is that other educators are able to download and use the material to instruct their constituents outside of Nebraska where the primary Extension material development effort will take place. Extension materials will be published through the UNL Cooperative Extension Service with the assistance of IANR Media, a multimedia communications unit within UNL. The material will be made available via the project's website and the UNL Precision Agriculture and On-Farm Research Network websites. The Extension materials developed from this project will be most suitable for delivery through eXtension, a partnership that UNL strongly supports.

Impacts
What was accomplished under these goals? Objective 1: Generate and gather data using precision agriculture technology to run coordinated long-term, large-scale, on-farm field trials to determine yield response. 1. We initially intended to perform 22 trials in 2017; however, we conducted 10 trials in Illinois, 1 trial in Kansas, 3 trials in Nebraska, 2 trials in Ohio, 3 trials in Ohio, and 18 trials in Argentina. In total, we generated and gathered data using precision agriculture technology to coordinate long-term large-scale, on-farm trials on 34 fields. 2. The NCSA successfully completed a beta-version of the cloud-based On-farm Research Design system, which will allow other research teams to collaborate in the design of randomized large-scale field trials. 3. Alexandre Ormiga Galvao Barbosa, a Ph.D. student in Mechanical Engineering from the University of Illinois, has created software that designs field trials that follow "as-applied" paths made by farm equipment when applying inputs in an earlier year. This software will improve field trial design particularly in areas outside of the Midwest, where slopes can affect trial implementation. As a result of this software, we have begun designing trials in Washington, Idaho, and Montana, where highly variant topography prevents farmers from driving machinery in "straight lines". 4. Dr. Tony Grift ran experiments on three farms to measure the yield from experimental plots with a weigh wagon that has a two pound resolution. This data was used to compare with the output of the yield monitors in the combine harvesters and to determine the latency time of the yield monitor. Measured yield on the three farms ranged from 150 to 276 bu/ac, with a 4.15-4.37% error between the yield monitor and the weigh wagon yield. In this experiment, it became clear that the machine ground speed needs to be taken into account when calculating latency distances in plots. Furthermore, we found that on modern machines with accessible CAN data (e.g. RTK-GPS location, engine speed, engine torque, and fuel consumption), we can access power data to monitor how much energy is being spent in harvesting a single plot, which can be correlated to yield. If this assumption is valid, this could be used as a proxy for yield in next year's experiments. 5. In 2017, the Data-Intensive Farm Management project began generating and collecting electroconductivity, pH, and organic matter data using its newly purchased Veris U3. To date, the U3 has been used on two fields in Illinois, with many more planned for the spring of 2018. Veris Technologies has volunteered to take electroconductivity readings of farms in Iowa. 6. The Fulginiti-Perrin research team completed preliminary evaluation of the payoff from VAR (variable application rates) for N at two project fields in Illinois. We conceptualize that payoff as the expected value of obtaining a soil characteristic signal for a grid prior to choosing the N rate. The ex-post value of obtaining the soil EC (electrical conductivity) signal was low, averaging $1.00/acre on one field, $0.58/acre on the other. Whether this disappointing result was due to relatively uniform grids across the fieldor to weak correlation between EC and N responseremains to be determined. Subsequent team activity was directed toward the adaptation of a Bayesian model for determining the expected value of the signal (rather than ex-post), following classic development of models of EVSI (expected value of sample information).This will allow us to evaluate the expected payoff of VAR on a particular field, given some characteristics of the field. Objective 2: Estimate the relationship between fertilizer use and water quality 1. Three experimental sites have been instrumented with water collection systems. The water collection systems consist of control structures installed in-line with the subsurface drainage tile lines. One or more control structures have been installed on each to collect the majority of the drainage water from the experimental area. In some cases, the drainage networks overlap with adjacent fields. In these cases, control structures have been installed at both the upslope and downslope ends of the experimental areas. Each water collection system is also equipped with a pressure transducer and v-notch weir, allowing the depth of water to be measured (pressure transducer) and the water flow rate to be calculated (from the v-notch weir calibration). Grab samples of water are collected multiple times from each field by a field technician and analyzed for nitrate-N and orthophosphate. The first field was installed in December 2016. The second and third fields were installed in December 2017. 2. To date, data has been collected on1,729,307 nitrogen readings from 87,883 water sites in the US from 1987-2006(Source: Water Quality Portal, United States Geological Survey),1,246,363 phosphorus readings from 77,742 water sites in the US from 1987-2006 (Source: Water Quality Portal, United States Geological Survey), county-level fertilizer application from 1987-2006 (Source: National Water-Quality Assessment Program), county-level weather data from 1987-2006 (Source:Agro-Climatic Data by County),andhydrologic maps of the United States (Source: United States Geological Survey).The results from these research projects are expected to provide information useful to administrators of programs to limit runoff from agricultural production. In addition, results will provide insight on the impact of climate change on fertilizer use and subsequently water quality. This will aid in designing national policies for climate change mitigation and adaptation in the agricultural sector.Results of preliminary analysis were presented at the University of Massachusetts Environmental Working Group on December 1, 2017. 3. With the Illinois Corn Growers Association, Field Lab was completed, and nutrient flow data needed to calibrate the nutrient loss monitoring system was gathered, allowing for initial use of the field lab during the 2018 crop year. Objective 3: Provide training to students who speak Spanish or Portuguese. At present, we have brought on ten Spanish- or Portuguese-speaking undergraduate and graduate students to the project. Two Portuguese-speaking Ph.D. students are Brazilian native speakers, and one undergraduate is learning Portuguese as a condition of the undergraduate FAB program. Five M.S. or Ph.D. students are Spanish native speakers, and one undergraduate is learning Spanish as a condition of the undergraduate FAB program. A final Ph.D. student is learning both Spanish and Portuguese. Objective 4: Expand awareness of optimal nitrogen management practices. 1. Team members at Illinois State University have worked in collaboration with the McLean County (IL) Soil and Water Conservation District Board of Directors to interview local farmers about their awareness of, and attitudes toward, issues of nutrient loss reduction and the impact that nutrient loss has on their farming operations. This phase of the project proceeded on schedule, with interviews completed in 2016 and two public presentations of research findings in 2017. 2. In December of 2016, DIFM hosted its annual DIFM Field Day, which was organized by DIFM's Gary Letterly. Dr. David Bullock and Dr. Laura Gentry gave detailed presentations and held question-and-answer sessions at the field day, discussing both the checkerboard trials and the field laboratory. Approximately 30 people were in attendance. DIFM project members Keith Glewen and Laura Thompson worked closely with farmers through their positions at the University of Nebraska's On-Farm Research Network and presented the DIFM project to groups in the state.

Publications

  • Type: Conference Papers and Presentations Status: Accepted Year Published: 2018 Citation: Puntel, L. 2018. A Crop Simulation Approach to Estimate The Value Of On-Farm Field Trials. International Society of Precision Agriculture.
  • Type: Conference Papers and Presentations Status: Accepted Year Published: 2018 Citation: Edge, B. 2018. An Economic Theory-Based Approach To Management Zone Delineation. International Society of Precision Agriculture.
  • Type: Conference Papers and Presentations Status: Submitted Year Published: 2018 Citation: Crago, C. and J. Paudel. 2018. U.S. Agricultural Adaptation to Climate Change: Implications For Fertilizer Use. World Congress of Environmental Economists. (Submitted).
  • Type: Conference Papers and Presentations Status: Submitted Year Published: 2018 Citation: Crago, C. and J. Paudel. 2018. Fertilizer Use And Water Quality In The United States. Agricultural and Applied Economics Association. (Submitted).
  • Type: Journal Articles Status: Submitted Year Published: 2018 Citation: Bullock, D. and T. Mieno. 2018. Assessing The Value Of Information From On-farm Field Trials. American Journal of Agricultural Economics. (Submitted).
  • Type: Journal Articles Status: Submitted Year Published: 2018 Citation: Bullock, D. and D.G.P. Divina. 2018. The Origins, Consequences, And Implications of Stanfords 1.2 Rule For Nitrogen Fertilizer Application. Applied Economics and Public Policy. (Submitted).


Progress 03/01/16 to 02/28/17

Outputs
Target Audience:We reached out and worked with dozens of midwestern corn farmers. We ran eight large-scale on-farm field trials, and also lined upover thirty farmersto paticipate in our project in 2017. To publicize and promote our project, we made numerous presentations to farmer groups. We communicated with environmental advocacy groups,such as thePrairie Rivers Network, which focuses its efforts on promoting improvements in Illinois water quality. We dealt extensively with agribusinesses, having lengthy discussions and negotiations with agriculutral software developers, and with corporations with signicant research and development efforts. Changes/Problems:We made unsatisfactory progress in our research at the University of Kentucky. The scientist there did not follow project protocol when he designed the experiment, and as a result the data generated was not useful. We are working harder with that scientist this year, to keep him on track, and help him to better understand what is required. What opportunities for training and professional development has the project provided?We gave academic scholarships to five graduate students They have all begun their education. We funded and supervised five FAB Fellows, who have begun studying in a program designed to help them learn quantitative and statistical agricultural analysis. How have the results been disseminated to communities of interest?We participated in approximately ten public interviews on radio and in the agricultural (paper and on-line) press. We reported our results at the DIFM Field Day in Blue Mound, Illinois, and we made a presentation of our results in front of an audience of hundreds at the Farm Assets Conference in Normal, Illinois. What do you plan to do during the next reporting period to accomplish the goals?We plan to run around 50 field trials in Illinois, Nebraska, Kansas, Ohio, Iowa, and Argentina. We will continue to work with the Illinois Corn Growers' Association in readying the Field Laboratory for future research. We will continue to monitor nutrient losses on our experiemental fields. We will finish developing both the On-Farm Research Design and the Data-Intensive Farm Management software packages. We will continue to recruit and educate undergraduate and graduate scholars through our FAB Fellowship program. In more detail, we anticipate continuing to develop decision support and field trial design software, as well as a central databank. We will also be recruiting and training Puerto Rican undergraduate students, maintaining and updating the project website, and using media outlets to disseminate results. We will continue conducting team conference calls and sending out a non-technical newsletter to team members and stakeholders, and we will analyze data, attend conferences, and will have several years worth of research from which we will produce a greater amount of scholarly output. Our student advising and review of progress will also continue throughout the year as planned. During the winter of our third year, we will plan meetings to design trials for the following year, hold our annual advisory board and whole project meeting, and perform face-to-face reviews of undergraduate and graduate student progress. We will also meet with participating and additional interested farmers. These are some of our revised goals as a consequence of our experiences and successes during our first year: Learn More About Yield Response: In the winter of our third year, we will be holding our meetings with new and participating farmers, and we will plan a meeting to design trials for the following year; however, we plan to use the $20,000 grant from ACES as well as our unused travel funds to recruit far more farms than we originally planned. We are considering adding farmers to our project by including winter wheat. More Complete Characterizations: Our collaboration with Intelinair will provide us with pictures of fields from the air, which will help to create more complete characterizations of our farmers' fields. We are also seeking methods of accounting better for in-soil nitrogen in future years. Increase Foreign Component: Lead PI David Bullock will be attending a Big Farmer/agribusiness meeting in Argentina in July 2017 in order to gain more contacts in South America. Our $20,000 grant will allow us to do more work in Brazil with Embrapa. This will help to increase our data output as a consequence of the reversed growing seasons in the southern hemisphere. Reduce Administrative Costs: In our first year, we discovered that working with bigger farms has been a tremendous help in reducing the overhead cost and time of working with smaller farms. In year 3 of our project, we will continue to aim for these big farms. Learn More About Water-Quality Response: We have already prepared a grant proposal for next year's USDA INFEWS grant application, in conjunction with the DIFM program. We will be working with cover cropping and water technology in order to better understand the role water plays in this system. By including water monitoring in our field characterization costs, we can keep farmer compensation costs low. Make Better Use of Students in Training: By better incorporating our FAB Fellows and graduate research assistants into our workflow, we will be helping to create a well-trained future workforce while also optimizing the resources we have available. Increase Scholarly Output: By working on scholarly research writing in tandem with increased data over the course of three years, we will use the growing season in summer when the DIFM project is less focused on other tasks to produce a greater amount of scholarly output than was previously possible. More Collaboration with the National Center for Supercomputing Applications: We hope to have solid versions of OFRD and DIFM software completed by the harvest season in 2018, a goal which is made possible by a graduate student at the NCSA.

Impacts
What was accomplished under these goals? 1. We have developed a beta version of the data-driven on-farm precision-agriculture technology. 2. We cooperated with farmers in Illinois and Nebraska to conduct eight on-farm field trials, as described in our original proposal. 3. We collected the experimental data, and are in the process of analyzing it, with the aim of providing farmers with profitable management advice. 4. We worked with the Illinois Corn Growers' Association to oversee the construction of a 160-acre field laboratory, which was compeletely re-tilled for our research purposes. 5. We developed relationships with Argentine agribusinesses, and are currently working with them, running thirteen field trials. 6. We recruited under-represented students from Puerto Rico to participate in our research.

Publications