Manuel Salvador-Castillo, José, Alejandro Bolaños-González, Martín, César Rodríguez, Julio, Palacios-Vélez, Enrique, Alberto Palacios-Sánchez, Luis, Watts, Christopher, Lizárraga-Celaya, Carlos, Ortega-Farías, Samuel, and Er-Raki, Salah
Table grapes (Vitis vinifera) are one of the crops of highest economic and social relevance in Sonora, Mexico. Due to the scant precipitation and high evaporative demand of the zone, it is a crop with a high water requirement. Therefore, its production depends on the application of irrigation, and it is important to have precise and extensive estimation of the crop evapotranspiration (ETC) to improve efficiency of irrigation in the short term. This research was carried out during 2018 and 2019 with the objective of developing and evaluating a model between the normalized difference vegetation index (NDVI) calculated with Sentinel-2 images, and the crop coefficient (KC) determined with a covariance turbulence flux system (Eddy Covariance, EC) as an option for estimating the ETC of a vineyard of table grapes in Costa de Hermosillo, Sonora. Using NDVI and KC data from 2018 records, a simple regression model was constructed with the start forced to the origin at the null coordinates (KCNDVI = 0.9467 NDVI; R2 = 0.74) as a base for ETC estimation. When the regression model was validated with daily 2019 data, an R2 of 0.76 was obtained along with a CME of 0.11 by relating KC vs. KCNDVI. Whereas by relating ETC vs. estimated ETC, an R2 of 0.92 was found, and a CME of 0.67 mm d-1. The results indicated that the ETC can be estimated with adequate precision and in an opportune manner with the proposed model. However, it was found that the model may underestimate ETC during the period of maximum crop development due to NDVI saturation. However, in the winter when the values of NDVI depend on the soil structural characteristics and on the remains of winter pruning, ETC tends to be overestimated. [ABSTRACT FROM AUTHOR]