"Skyborne Surveillance: NASA's Role in Identifying Wine Grape Disease in California"

In a scientific investigation, researchers identified a costly infection in cabernet sauvignon grapevines before any visible symptoms emerged. Plant diseases like molds, bacteria causing root rot, viruses, and other pathogens result in substantial losses of 15% to 30% of global crop yields annually. Timely detection can determine the fate of a harvest, whether it fails entirely or can be salvaged.

By utilizing an airborne scientific device created at NASA's Jet Propulsion Laboratory in Southern California, scientists have effectively pinpointed subtle indications of a grape disease that causes billions of dollars in crop damage each year. This remote sensing technique has the potential to assist in monitoring various crops from the ground. In two recent studies, experts from JPL and Cornell University concentrated on a viral ailment called GLRaV-3 (short for grapevine leafroll-associated virus complex 3). Mainly transmitted by insects, GLRaV-3 decreases yields and spoils developing fruit, leading to annual losses and damages of around $3 billion in the U.S. wine and grape sector. The usual method of detection involves labor-intensive manual inspection of each vine and expensive molecular testing. The first of these studies has been published in the journal Phytopathology.

The research team aimed to determine whether they could use machine learning and NASA's advanced Airborne Visible/InfraRed Imaging Spectrometer (AVIRIS-NG) to aid growers in identifying early GLRaV-3 infections from the air. This optical instrument, which captures interactions between sunlight and chemical bonds, has been employed to monitor various hazards like wildfires, oil spills, greenhouse gases, and pollution associated with volcanic eruptions.

In 2020, while conducting a campaign to map methane leaks in California, plant pathologist Dr. Katie Gold and her team seized the chance to explore a different question: Could AVIRIS-NG uncover hidden crop infections in one of the state's primary grape-producing regions? According to Gold, sick plants, like humans, might not exhibit immediate outward symptoms, making early detection a significant challenge for growers. The signs of grapevine leafroll virus infection might take up to a year to become visible, such as changes in foliage color and stunted fruit growth. However, at the cellular level, stress is already underway, altering how sunlight interacts with plant tissue.

Installed in a research plane, AVIRIS-NG observed approximately 11,000 acres of vineyards in Lodi, California, a major wine grape producer in the heart of the state's Central Valley. The team used these observations to train computer models to differentiate between infected and non-infected vines. To validate the results, industry collaborators examined over 300 acres of vineyards on the ground for visible viral symptoms while collecting vine samples for molecular testing.

Although a labor-intensive process, this approach yielded positive outcomes, achieved during a California heatwave. Gold acknowledged the effort put in by the growers, collaborators, and scouting teams, emphasizing the collaboration's essential role. Similar initiatives will continue under the NASA Acres Consortium, led by Gold.

The researchers discovered that they could successfully distinguish between infected and non-infected vines both before and after symptoms emerged, with the most accurate models achieving an 87% accuracy rate. Detecting GLRaV-3 early could offer grape growers up to a year to take action.

In an accompanying article published in the Journal of Geophysical Research: Biogeosciences, the researchers noted how advancements in air and space capabilities can support ground-based pathogen surveillance. These capabilities, including upcoming missions like NASA's Surface Biology and Geology, could combine with machine learning to aid agricultural decision-making globally.

Fernando Romero Galvan, a doctoral candidate and co-author of both studies, stressed the importance of sustainable farming practices in the face of climate change. He expressed enthusiasm about the potential of remote sensing and plant disease detection, highlighting scalable solutions that could enable data-driven and sustainable crop management choices.

Co-author Ryan Pavlick, a research technologist at JPL, emphasized the broader vision of this study—enabling such detection and monitoring not just in one region of California for a single disease, but across the world for multiple crop diseases and growers everywhere.