Grapevine water status strongly impacts vine yield and berry quality, hence the importance of a cost- and time-effective methodology to estimate it in commercial vineyards. A two-year experiment was carried out in a rainfed commercial vineyard (cv. ‘Vermentino’ grafted on ‘1103P’) to test the ability of vegetation indices (VIs) derived from Sentinel-2 (S2) multispectral imagery in estimating the stem water potential (Ystem). The S2 VIs calculated including NIR and SWIR bands (NDMI, MSI and NMSI), performed better (R2 values of 0.66, 0.66 and 0.59, respectively) than those calculated using the bands in the VIS-NIR region (R2 values 0.47, 0.58 and 0.58 for NDVI, GNDVI and NDWI, respectively). The slope between B8a and B11 bands was also significantly affected by grapevine water status. The different weather conditions occurred during the two experimental years also allowed to observe as the S2 VIs performed better in estimating Ystem under prolonged drought conditions. Moreover, S2 VIs were not able to detect differences in grapevine water status during the onset of water stress condition.
Keywords: Remote sensing, Stem water potential, Vegetation indices, Vitis vinifera L.