Eve Laroche-Pinel, Vincenzo Cianciola, Khushwinder Singh, and Luca Brillante*
*Department of Viticulture and Enology, California State University Fresno, 2360 E Barstow Ave, Fresno, CA, 93740 (lucabrillante@csufresno.edu)

Satellite imagery is a powerful tool to assess vineyard characteristics such as vigor or water content. To explore the ability to link satellite information with vineyard characteristics, we set up an experiment in 2020 and 2021 in a Merlot vineyard in the Bakersfield area. A total of 24 experimental units were distributed spatially according to the grid created from the pixels of a Landsat 8 image. Water status was assessed every two weeks in both years, from June to August, measuring midday stem water potentials (Ψstem); we also measured leaf stomatal conductance (gs) and net carbon assimilation rate (AN). Berry weight, pH, total soluble solids (Brix), and titratable acidity (TA) were measured every two weeks from July to August. Landsat 8 images of the same periods were downloaded and reflectance values of each experimental unit were extracted and averaged. Maximum temperature (Tmax) and minimum relative humidity (RHmin) were also extracted for the location of the vineyard. Machine learning models were applied to predict water status or berry information using band reflectance values alone or with Tmax and RHmin, using a block-out validation method. The first results show an R2 of 0.8 for predictions of Ψstem, AN, and gs using band values. Adding Tmax and RHmin as predictors increases R2 to 0.9 for AN and 0.85 for gs. Using feature importance extraction, it was determined that the near-infrared (NIR) and shortwave-infrared (SWIR) spectral domains are most effective for predicting water status or grape characteristics. Previous research established that these spectral domains are closely associated with plant water status or cellular composition. These findings validate the potential of satellite imagery as a means of monitoring vineyards on a broad scale and with a high level of temporal resolution.

Funding Support: American Vineyard Foundation, California State University – Agricultural Research Institute