Scalable early detection of grapevine viral infection with airborne imaging spectroscopy
Fernando Romero Galvan, Cornell University, fer36@cornell.edu (Presenter)
Ryan Pavlick, Jet Propulsion Laboratory, rpavlick@jpl.nasa.gov
Somil Aggarwal, Cornell University, sa748@cornell.edu
Graham Trolley, University of Connecticut, grt38@g.cornell.edu
Daniel Sousa, San Diego State University, dan.sousa@sdsu.edu
Charles Starr, Viticultural Services, cstarriv@gmail.com
Elisabeth Forrestel, University of California, Davis, ejforrestel@ucdavis.edu
Stephanie Bolton, Lodi Winegrape Commission, stephanie@lodiwine.com
Maria del Mar Alsina, E. & J. Gallo, mariadelmar.alsina@ejgallo.com
Nick Dokoozlian, E. & J. Gallo, nick.dokoozlian@ejgallo.com
Kaitlin Gold, Cornell University, kg557@cornell.edu
Viral disease, including that caused by Grapevine Leafroll Virus Complex 3 (GLRaV-3), causes $3 billion in damages and losses to the US wine industry annually. GLRaV-3 has a long latent period in which vines are infectious but do not yet display visible symptoms, making it an ideal model pathosystem to evaluate symptomatic and asymptomatic remote sensing-based disease detection. The ability to detect plant disease during the latent period at scale would greatly reduce management costs, current detection methods are labor-intensive, expensive, and non-scalable. Here, we use airborne imaging spectroscopy data collected in September 2020 by NASA's Airborne Visible/Infrared Imaging Spectrometer Next Generation (AVIRIS-NG) to detect GLRaV-3 in red grapevines Lodi, CA. In September 2020 and 2021, industry partners scouted 280 acres of Aglianico, Cabernet Sauvignon, and Petite Sirah grapes for visible disease symptoms, and a subset was collected for confirmation testing. We combined random forest with SMOTE oversampling to train a spectral model able to distinguish between healthy and GLRaV-3 infected grapevine. We observed clear spectral differences that allowed for differentiation between healthy and GLRaV-3 infected vines at 1m resolution with 82%/0.65 accuracy/kappa. We hypothesize these spectral differences are linked to changes in overall plant physiology induced by disease, as visible foliar symptoms were restricted to the lower canopy. The next steps for this work are to assess scalability to spaceborne resolution for use with NASA's forthcoming Surface Biology and Geology satellite.
Associated Project(s):
Poster Location ID: 3-36
Presentation Type: Poster
Session: Poster Session 3
Session Date: Thu (May 11) 3:00-5:00 PM
CCE Program: BDEC