Scalable Early Detection of Grapevine Viral Infection with Airborne Imaging Spectroscopy

Area of Science:

• Agricultural remote sensing
• Plant pathology
• Spectroscopy

Background:

• Grapevine viral diseases, like grapevine leafroll-associated virus complex 3 (GLRaV-3), cause significant annual losses in the U.S. wine and grape industry.
• Current detection methods are costly and labor-intensive, hindering effective disease management.
• GLRaV-3's latent period, where infected vines show no visible symptoms, presents a challenge for early detection.

Purpose of the Study:

• To evaluate the scalability of imaging spectroscopy for detecting GLRaV-3 in grapevines.
• To assess the potential of airborne imaging spectroscopy for identifying latent GLRaV-3 infections.
• To lay the foundation for using future hyperspectral satellite data for regional grapevine disease monitoring.

Main Methods:

• Deployment of NASA's Airborne Visible and Infrared Imaging Spectrometer Next Generation (AVIRIS-NG) over Cabernet Sauvignon vineyards in Lodi, CA.
• Ground-truthing through vine-by-vine scouting for viral symptoms and molecular confirmation testing over 317 hectares.
• Training random forest models on spectroscopic data, incorporating synthetic minority oversampling to balance datasets.

Main Results:

• Imaging spectroscopy models successfully differentiated between noninfected and GLRaV-3 infected vines, both pre- and postsymptomatically, at 1 to 5 m resolution.
• The best models achieved 87% accuracy in distinguishing asymptomatic infected vines and 85% accuracy for asymptomatic + symptomatic vines.
• The analysis highlighted the importance of nonvisible wavelengths, indicating disease-induced physiological changes in plants.

Conclusions:

• Airborne imaging spectroscopy is a scalable and effective method for detecting GLRaV-3, even during its latent phase.
• This technology offers a promising tool for early and accurate disease detection in vineyards, potentially reducing crop losses.
• The findings support the use of future hyperspectral satellite missions for broad-scale crop disease surveillance.