Issue link: http://winesandvines.uberflip.com/i/776679
February 2017 WINES&VINES 87 WINE EAST GRAPEGROWING the first two PCs; leaf ψ and TA were inversely correlated with FVT, potentially volatile terpenes (PVT), Brix and pH (PC1), while NDVI and berry weight were inversely cor- related with SM, yield and cluster number (PC2). The non-hierarchi- cal classification algorithm k- means was conducted for three clusters and supplemented the PCA as a qualitative variable with ob- servation biplots projected in dif- ferent colors based on low, medium and high NDVI values. The low NDVI observations appeared closer to FVT, PVT and SM, whereas the high NDVI ones were closer to berry weight, NDVI and TA. Maps (see "Lam- bert Riesling Vineyard") showed that NDVI demonstrated very temporally consistent patterns in both vintages with high values in the north side and low values in the south; correlations were established with berry weight, cluster weight and TA, while in- verse relationships were exhib- ited with SM and FVT. Moran's I indicated a 100% clustering pat- tern, further confirming results obtained from statistical and spatial analysis. Overall, high NDVI was associated with yield components and vine size, while low NDVI was associated with higher anthocyanins, phenols and color for the red cultivars and terpenes for Riesling. Summarized statistics were conducted for all the variables, and the most notable coefficient of variation was observed for yield (CV% 23.5-62.8 across sites and years), followed by high variability in vine size (CV% 22.9-38.9). These basic statistics for one of the vine- yards are presented in "Results for Yield Components, Berry Composition and NDVI." This underscores the potential of a precision viticulture approach for zonal management and/or selective harvesting. 29 Likewise, variables associated with yield exhibited similar within-field variation. With respect to berry compo- sition, the most variability was exhibited by anthocyanins, color and phenols in Pinot Noir and Cabernet Franc and monoter- penes in Riesling. Aside from establishing relationships among NDVI and other variables, it was essential to show that spatial variability patterns remained stable over time and that the spatial variability in yield com- ponents could be reflected in fruit composition. Despite the small vineyard sites (≈2 to 3 acres), spatial variability was observed for vine water status, NDVI and berry composition. LAMBERT RIESLING VINEYARD This PCA observation biplot (a) shows a Lambert Riesling vineyard in 2015 with NDVI values subjected to k-means clustering analysis. Colored dots correspond to vine groupings of high (green), medium (blue) and low (red) NDVI. Maps (b-h) likewise depict Lambert Riesling 2015. Individual points on map h represent the sentinel vines. Abbreviation: FVT: free-volatile terpenes. BIPLOT (AXES F1 AND F2: 45.46%) F2 (19.27%) 3 2 1 0 -1 -2 -3 3 2 1 0 -1 -2 -3 F1 (26.20%) n High n Medium n Low n Centroids n 0.76-0.77 n 0.78-0.78 n 0.79-0.78 n 0.79-0.79 n 0.8-0.79 n 0.8-0.8 2014 NDVI MEAN n 0.77-0.78 n 0.79-0.78 n 0.79-0.78 n 0.79-0.79 n 0.8-0.79 n 0.8-0.79 2015 NDVI MEAN n 130.54- 155.04 n 155.05- 159.25 n 159.26- 163.21 n 163.22- 166.43 n 166.44- 170.63 n 170.64- 193.65 2015 BERRY WEIGHT (G) n 6.22-8.11 n 8.12-8.33 n 8.34-8.51 n 8.52-8.72 n 8.73-8.97 n 8.98-10.55 2015 TITRATABLE ACIDITY (G/L) n 16.1-18.1 n 18.11-19.52 n 19.53-20.67 n 20.68-21.75 n 21.76-23.17 n 23.18-25.9 2015 SOIL MOISTURE (%) MEAN n 0.38-0.5 n 0.51-0.57 n 0.58-0.65 n 0.66-0.71 n 0.72-0.78 n 0.79-0.89 2015 FVT (MG/L) n 0.05-0.12 n 0.13-0.13 n 0.14-0.14 n 0.15-0.15 n 0.16-0.19 n 0.2-0.35 2015 CLUSTER WEIGHT (KG) (b) (c) (d) (e) (f) (g) (h) (a) PVT FVT Brix Soil moisture pH Berry weight NDVI TA Cluster number Vine size Yield Leaf Ψ