Remote sensing in the estimation of evapotranspiration of tomato cultivation for industrial processing

Abstract

This study evaluated the performance of the SAFER and METRIC algorithms to estimate the actual evapotranspiration (ETa) of irrigated tomato crops for industrial processing in the south-central region of Goiás, Brazil. The research was conducted in eight tomato-producing areas using center-pivot irrigation during the 2018 and 2019 harvests. Landsat 8 OLI/TIRS satellite images (temporal resolution of 16 days) helped estimate ETa through the SAFER e METRIC models compared with FAO methods, using the single crop coefficient (Kc) of the FAO-56/Embrapa and the soil water balance (BHS) method based on statistical indices. The analyzed algorithms presented spatiotemporal variations for ETa during the tomato crop cycle for industrial processing. The maximum evapotranspiration estimated by SAFER was 5.20 mm d-1, and by METRIC was 5.00 mm d-1. The algorithms were accurate compared with the standard methods, mainly the FAO using Embrapa’s Kc. The mean squared error was lower than 0.59 mm d-1 for SAFER and lower than 0.73 mm d-1 for METRIC. The ETa estimated by both models in the vegetative and fructification phases was lower than the mean absolute error of 0.24 mm d-1 compared with the standard methods. The SAFER model showed higher agreement with standard practices than the METRIC model, with an index between 0.64 and 0.99. This study demonstrated that algorithms may effectively estimate ETa in tomato crops for industrial processing in the analyzed region.