Utilization of unmanned aerial vehicles and proximal sensing to detect Riesling vineyard variability
A single vineyard block can consist of significant spatial variability for several grape-growing attributes. The ability to detect and subsequently respond to this variation can lead to improved vineyard management, a growing practice termed precision viticulture. The overall goal of this research study was to determine if remote-sensing technologies could be used to detect Riesling vineyard variability, thus enhancing precision viticulture implementation. Approximately 80 grapevines in a grid pattern were geo-located within each of six commercial Riesling vineyards across the Niagara Peninsula in Ontario. From these grapevines the following variables were measured to determine their vineyard variation: soil and vine water status, vine size/vigor, winter hardiness, virus titer, yield components, and berry composition. Subsequently, remote-sensing technologies collected thermal [by unmanned aerial vehicle (UAV)] and multispectral (by UAV and ground-based proximal sensing technology GreenSeeker™) data from each block. Multispectral data were transformed into the Normalized Difference Vegetation Index (NDVI). Vineyard UAV NDVI maps were further used for selective harvesting of areas corresponding to low vs. high NDVI and wines made from these two zones were compared chemically and sensorially. The hypothesis was that remote and proximal sensing technologies could accurately detect vineyard variation for manually collected variables and further implicate differences in wine attributes upon zonal harvesting. Direct positive correlations were observed between remotely and proximally sensed NDVI vs. vine size, leaf stomatal conductance, leaf water potential, virus infection, yield, berry weight, and titratable acidity and inverse correlations with Brix and potentially-volatile terpene concentration. Maps created from remotely and proximally sensed data demonstrated similar spatial configurations to interpolated maps of these variables. In general, GreenSeeker NDVI demonstrated the most significant relationships with measured variables compared to UAV NDVI and UAV thermal data. Wines created from areas of low vs high NDVI differed inconsistently in their wine pH. Sensorially, in certain sites and vintages, panelists were able to distinguish between wines made from low vs high NDVI zones. Overall, remote sensing demonstrates the ability to detect vineyard areas differing in measures of vine health, size, yield, berry composition, and wine attributes, though more research is needed to understand the inconsistent results observed between vineyard sites and vintages.