Cropland mapping with UAVSAR to support the NISAR mission

Authors: HUANG, XIAODONG - Applied Geosolutions, Llc; Reba, Michele; Coffin, Alisa; RUNKLE, BENJAMIN - University Of Arkansas; Huang, Yanbo; CHAPMAN, BRUCE - National Aeronautics Space Administration (NASA) - Jet Propulsion Laboratory; ZINITI, BETH - Applied Geosolutions, Llc; SKAKUN, SERGII - University Of Maryland; KRAATZ, SIMON - University Of Massachusetts; SIQUEIRA, PAUL - University Of Massachusetts; TORBICK, NATHAN - Applied Geosolutions, Llc

Submitted to: Remote Sensing of Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/6/2020
Publication Date: 11/24/2020
Citation: Huang, X., Reba, M.L., Coffin, A.W., Runkle, B.R., Huang, Y., Chapman, B., Ziniti, B., Skakun, S., Kraatz, S., Siqueira, P., Torbick, N. 2020. Cropland mapping with UAVSAR to support the NISAR mission. Remote Sensing of Environment. https://doi.org/10.1016/j.rse.2020.112180.
DOI: https://doi.org/10.1016/j.rse.2020.112180

Interpretive Summary: Knowing what crops and how much of each are planted are essential for agricultural inventory, food security, and production forecasts. Traditionally, this information is derived using optical data from satellite missions such as Landsat and Moderate Resolution Imaging Spectroradiometer (MODIS). The use of Synthetic Aperture Radar (SAR) in agriculture has lagged behind. New opportunities with SAR now exist with recent and planned launches. SAR have advantages over optical data because it can see through clouds and is more sensitive to how the crop is oriented and crop wetness. We used the Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) platform to observe four cropland sites across the southern United States to support the development of L-band SAR prototype science products. Multiple algorithms were tested and we used ground truth data collected at sites with major crops such as corn, soybean, wheat, rice, cotton, and peanut. The findings from this study show SAR can map crop landscapes with relative high accuracy. These findings further will help prepare for future SAR Missions with goals to support agricultural monitoring.

Technical Abstract: The Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) platform was used to observe cropland sites across the southern United States to support the development of L-band prototype science products. Time series flights flew over four independent sites during the growing season of 2019 while ground truth crop training data was collected. Major crops include corn, cotton, pasture, peanut, rice, and soybean. For cropland monitoring we evaluated a set of machine learning (ML) algorithms (random forest, feedforward fully connected neural network, support vector machine), the recently developed Multi-temporal Binary Tree Classification (MBTC), and a phenology (Coefficient of Variation; CoV) approach to synergistically assess performance, scattering mechanisms, and limitations. Technical objectives of this research application included 1.) the description of scattering mechanisms and physical interpretation of cropland SAR parameters; 2.) evaluation of L-band mapping performance across multiple independent agricultural production areas with field scale training data; and 3.) assessment of the CoV approach for the generation of prototype NISAR Level 3 products. SAR terms with sensitivity to volume scattering performed well and consistently across CoV mapping experiments achieving accuracy greater than 80% for cropland vs not cropland. Dynamic phenology classes, such as herbaceous wetlands, had some confusion with CoV agriculture requiring further regionalized training optimization. Volume scattering was most useful across the different ML techniques with overall accuracy and Kappa consistently over 90% for crop type by late growth stages. Time series information proved more valuable compared to any single ML technique or site. Ultimately, as more SAR platforms launch, the user community should leverage physical contributions of different wavelengths and polarizations for efficient and meaningful agricultural products.