SustaiN: A Decision Support System for Sustainable Nitrogen Management in Corn and Sorghum using Satellite Remote Sensing

Project Overview

Project Type: Graduate Student
Funds awarded in 2022: $14,966.00
Projected End Date: 06/30/2024
Grant Recipient: Saint Louis University
Region: North Central
State: Missouri
Graduate Student:
Faculty Advisor:
Dr. Vasit Sagan
Saint Louis University
Faculty Advisor:
Dr. Andrea Eveland
Donald Danforth Plant Science Center
Dr. Todd Mockler
Donald Danforth Plant Science Center
Dr. Stephen Moose
University of Illinois Urbana-Champaign
Dr. Nadia Shakoor
Donald Danforth Plant Science Center

Information Products

SustaiN (Decision-making Tool)


  • Agronomic: corn, sorghum (sweet)


  • Crop Production: application rate management, cropping systems, fertilizers
  • Sustainable Communities: sustainability measures

    Proposal abstract:

    Sustainable Nitrogen (N) management in both corn and sorghum production is a crucial step for farmers. In addition to base rate of N application, in-season N ensures more N-uptake to crops and significantly increases yield. However, excessive use of N can lead to unnecessary cost for farmers and environmental threats as the excess N can result in leaching to aquifers and runoff towards nearby streams. Farmers often rely on N-sprayers equipped with sensors; however, the use of ground-based applicators often leads to crop damage and more operational cost for farmers. Therefore, we propose to utilize satellite remote sensing and other publicly available datasets (e.g., precipitation, temperature) to develop in-season N- prescription model for corn and sorghum within large-scale regions of the Midwest. The output of the model will be hosted in an interactive web platform, which can be easily accessed by farmers to make informed decisions about sustainable N-application. We are calling the web-platform as “SustaiN”. Several experimental fields will be set up with collaborators from Donald Danforth Plant Science Center and University of Illinois, where two types of plots will be established, i.e., one with 225 lb/ac N base rate (reference plots as the best quality crop) and the other with variable rate N. In addition, we will also reach out to interested farmers in both Missouri, Illinois, and Kansas for participation. From the experimental fields, we will collect simultaneous drone and non-destructive measurements (e.g., leaf chlorophyll and nitrogen content) throughout the growing season. To develop the N-prescription model, we will rely on PlanetScope satellite images for spectral data (4 bands), and gridMET for surface meteorological data. Machine learning techniques will be used to develop a plant N-prediction model, where the inputs are fused satellite datasets and output is plant N-content. Based on the N-content of the reference plots, optimized variable rate N will be prescribed. The machine learning model and satellite datasets will allow us to interpolate the results for larger regions. We will evaluate our progress by circulating the results among the participating farmers and questionnaire surveys. The outcomes of this project will increase farmer profitability by prescribing optimal in-season N application, reduce environmental stress by minimizing the use of excessive N, and increase social sustainability by disseminating the knowledge among farmers, researchers, policy planners and seed providers.

    Project objectives from proposal:

    Learning Outcomes

    The learning outcomes will be shared to farmers and researchers by peer-reviewed journal publication, conference presentation, and hosting the results in a project website.

    1. Farmers will learn about how satellite data and environmental factors can explain in-season N-requirement for corn and sorghum within a larger area. They will know if there was any environmental stress throughout the growth stages and the contribution of the stress in overall N-uptake.
    2. The project will shed light on which genotypes of corn and sorghum has better nitrogen use efficiency in terms of different environmental condition.
    3. Researchers will have access to unique crop datasets for training machine learning models. The machine learning training usually requires massive number of datasets, which is often difficult to collect in terms of agricultural context. The availability of open-source datasets and tools from this study will help researchers to develop more reproducible data-driven models for tackling other agricultural challenges.

    Action Outcomes

    1. Farmers will apply optimized level of in-season N without relying on drone-based flights or ground-based sensors. They will know when and where in-season N should be applied for achieving their yield goals. This will allow increased profitability for farmers and reduced operational cost. The outcome will also improve the overall soil quality of the farmers’ fields.
    2. Based on the knowledge of which corn and sorghum genotypes in a specific environmental condition efficiently utilize N, farmers will take informed decision on selecting the appropriate genotype for their fields in future cultivation.
    Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.