Using Sentinel-2 Images to Delineate Management Zones for Variable Rate Drip Irrigation in Vineyards

Adoption of variable rate irrigation (VRI) practices in precision agriculture requires describing the variability of the crop-soil system. Geophysical survey and multispectral imagery (acquired through satellite or drone platforms) are usually employed to investigate the variability of soil and canopy respectively. The use of Sentinel-2 (S2) satellite data is of particular interest, as they are free and readily available, although their spatial resolution is not suitable to accurately describe the variability of crops in rows as vines. On the other hands, for vineyards the S2 imagery mostly describes the vigour of the rainfed inter-row vegetation rather than of the drip irrigated vines. Consequently, S2 imagery is mainly correlated to soil properties. Moreover, since the management zones (MZ) for VRI mainly depend on soil variability, time-series of S2 images can be used to delineate the MZs. In this study, time-series of S2 images were analysed by principal component analysis to extract time-stationary components to be used for delineating MZs. The identified MZs were found to be functional in explaining the variability of the vine vigour, grape production and quality data measured at harvest. Moreover, the main factors influencing the vegetation vigour were analysed.