S125 The Dionysus Project: Classifying and Monitoring Vineyards with Satellite Remote Sensing & Image Analysis

Sunday, 6 January 2019
Hall 4 (Phoenix Convention Center - West and North Buildings)
Nicole Flores, The New York City College of Technology, Brooklyn, NY; and E. Perez-Flores

Abstract

Climate change is expected to impact the wine industry by shifting suitable growing regions away from established regions, and increasing the demand for freshwater supplies to maintain vineyard health. This creates challenges for maintaining vineyards since the grapevines need to be appropriately stressed to produce quality grapes for quality wine. This requires precision management of freshwater application and canopy management in order to produce a grape that has the appropriate concentration of flavors and sugars. Vineyards in the U.S. are typically monitored by vineyard managers at the large scale, lacking the resolution needed identify grapevine growth at the subfield level. Vineyard managers make decisions from planting and fertilizing to water schedules and harvesting. Unfortunately since vineyards are large areas, it will require more vineyard managers to monitor and identify grapevine health which will cost more for resources and labor. Remote sensing provides coverage over large areas, a cost-effective tool to monitor large vineyard areas, giving growers informed decisions on irrigation timing and amounts related to vineyard grape development. By using remote sensing, it will allow vineyard managers to improve in decision-making to maintain grapevine health.

For our research, satellite remote sensing data was used to analyze two varietals - Chardonnay and Cabernet Sauvignon - at a vineyard in the North Fork of Long Island, NY during the 2017 and 2018 growing season. This research used Landsat-8 and Sentinel-2 satellite data to generate Normalized Difference Vegetation Index (NDVI), an indicator of vegetation “leafiness”, and Normalized Difference Water Index (NDWI), an indicator of plant canopy water content, for the following research goals. First, to identify how well each satellite tracks grapevine growth and health between the two growing season (April to September 2017 and 2018) using differences maps and time-series analysis. Second, to use image classification to determine how well each satellite dataset identifies the vineyard by varietal type in terms of location and through the growing season. Ground data collected during each growing season will verify the accuracy of Landsat-8 and Sentinel-2 observations.

Preliminary results indicate NDVI and NDWI measurments accurately track grapevine for each growing season, as verified by ground data. Preliminary results also indicate that the growing season in 2018 had a higher NDWI, signifying that the grapevines had a higher water content. Whereas, the growing season in 2017 had a higher NDVI.

In all, this research seeks improve our understanding on the ability of satellite remote sensing to observe and monitor vineyards and inform vineyard managers on the state of their grapevines. Improving our understanding of vineyard monitoring can lead to better identification of vineyard fields that need more management, such as disease monitoring or more irrigation of specific areas to maintain grapevine health.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner