Remote Sensing and Multivariate Analysis for the Characterization of a Vineyard Area

Abstract

The Recas wine-growing area, Romania, was analyzed and characterized through remote sensing and multivariate analysis. ASPECT, SLOPE and DEM were the parameters determined for the morphological characterization of the surface in the considered area. The parameter values were classified into nine classes characterizing the vineyard area (VLC1 to VLC9). For the vineyard areas ASPECT is a very important parameter. Classes VLC4, VLC5, VLC6 (S-E, S, and S-V) recorded 713.94 ha, which represented 58.85% of the total ASPECT parameter. To these classes can be added the VLC3 (10.10%) and VLC7 (15.59%) classes. In the case of the SLOPE parameter, values between 13.19 ha (VLC9) and 238.31 ha (VLC3) were recorded. In the case of the DEM parameter, values between 51.67 ha (VLC9) and 191.52 ha (VLC6) were recorded. Based on multivariate analysis, the principal components explained 95.983% of the total variance, and the VLC were distributed within the PCA diagram correlated with territorial parameters. The cluster analysis led to the grouping of VLC classes based on similarity, into three distinct subclusters (Coph.corr.=0.785). Cluster C1 included the first two subclusters, C1-A ((VLC2,VLC3), VLC1) and C1-B ((VLC4,VLC5), (VLC6,VLC7)). Cluster C2 included the classes (VLC8, VLC9). A high level of similarity was recorded between classes VLC6 and VLC7, SDI=47.46. Mathematical models described the variation of parameter values in relation to the classes position for the vineyard area.