Skip to main content
ARS Home » Southeast Area » Tifton, Georgia » Southeast Watershed Research » Research » Publications at this Location » Publication #410547

Research Project: Shifting the Balance of Water Resources and Interacting Agroecosystem Services Toward Sustainable Outcomes in Watersheds of the Southern Coastal Plain

Location: Southeast Watershed Research

Title: Predicting Cotton Above Ground Biomass Using Multispectral Imagery from Uncrewed Aerial Systems

Author
item STEEN, KYLE - University Of Georgia
item Coffin, Alisa
item MARGUERITE, MADDEN - University Of Georgia
item SEYMOUR, LYNNE - University Of Georgia

Submitted to: Association of American Geographers
Publication Type: Abstract Only
Publication Acceptance Date: 2/29/2024
Publication Date: 4/16/2024
Citation: Steen, K., Coffin, A.W., Marguerite, M., Seymour, L. 2024. Predicting Cotton Above Ground Biomass Using Multispectral Imagery from Uncrewed Aerial Systems. Association of American Geographers. Abstract.

Interpretive Summary:

Technical Abstract: Predicting above ground biomass (AGB) is beneficial to a multitude of parties including agriculturalists, government, and consumers. However, collecting AGB data is labor intensive and costly. Remote sensing images acquired via uncrewed aerials systems (UAS), or drones, can provide high resolution, multispectral data for the analysis of ABG. In 2018 and 2019, ABG was measured several times throughout the growing season from plots within a commercial cotton farming operation near Ashburn, Georgia. Drone-based multispectral imagery was also collected co-incident with the ABG measurements. The imagery was analyzed with a random forest model to predict AGB at plot-level and field-level scales, and the results were compared across years. Precipitation and Normalized Difference Vegetation Index (NDVI) values were also analyzed to visualize any existing correlation between these variables and AGB. Results detailed that 2018 had cumulatively more precipitation than 2019, but the monthly precipitation values varied across the growing season, May through October, for both years. The NDVI values for 2019 never surpassed the threshold value of 0.9; alternatively, a majority of NDVI values for 2018 were above 0.9. Plot 39 was an area of interest during the analysis, and predicted AGB for Plot 39 details a higher maximum value for 2018 in comparison to 2019. Additionally, for 2019, Plot 39 is the sole biomass plot to display an increase in biomass from September 18th to October 24th. Field level predictions provided for 361,000 kg of AGB at time of harvest in 2019. Future work can include the analysis of precipitation at weekly intervals, and the comparison of soil indices.