Geospatial Modeling of Wine Grape Quality (Anthocyanin) for Optimum Sampling Strategy in Mechanized On-The-Go Differential Harvesting Programs

摘要

Site-specific harvest of wine grapes based on quality and segregation before delivery to winery is a profitable cultural practice in vineyard management. Wine grape segregation based on quality parameters like anthocyanin to delineate quality zones in vineyards aids differential harvest. However, capturing vineyard variability with optimal sampling strategies is essential for economical feasibility of differential harvest. In this study, anthocyanin (mg/g fruit) data was collected on two production vineyards for geo-statistical analysis of spatial variability and determination of optimum samples/acre (SPA) for differential harvesting programs. Geo-referenced field samples of wine grapes were measured for anthocyanin using near-infrared sensors (NIR) in two vineyards of San Joaquin Valley California (Twin Creeks & Merjan). Two strategies of sampling, strategy I & II were utilized for 3, 5, 7 and 10 (reference) SPA. While strategy I selected 3, 5 or 7 random SPA in whole vineyard, strategy II did the same from every 1 acre block of the vineyard. Geo-spatial interpolations using ordinary kriging prediction of anthocyanin evaluated through cross validation parameters resulted in determining applicability of strategies in capturing vineyard field variability for differential harvesting. Strategy II outperformed strategy I in predictions with 5 and 7 SPA predicting vineyard spatial variability.