Estimativa da evapotranspiração em regiões do semiárido brasileiro utilizando o algoritmo Sebal.

Abstract

The objective of this work was to estimate the evapotranspiration using satellite image spectral data and meteorological elements, using the SEBAL model, which has the principle of using the energy balance equation and comparing it with the evapotranspiration estimated by the method of Penman-Monteith. The study areas are from Sousa-PB and Barbalha-CE, the Sousa area being cultivated with dwarf coconut and the Barbalha area cultivated with Banana nanica. The Landsat-8 OLI / TIRS satellite images were used to perform the study, and to use complementary surface data such as: average air temperature, relative air humidity, atmospheric air pressure, wind speed, global solar radiation, obtained from automatic weather stations located near the areas. The images were processed from the software ERDAS imagine version 9.1, in which the processes such as: stacking of the bands, cutting of the stacked images, followed with the image processing, using the Model Maker tool were performed mathematical operations, and evapotranspiration determination. The daily actual evapotranspiration of the coconut in the Souza county and the banana culture at Barbalha county, estimated by the SEBAL Algorithm are in agree with those found in the bibliography. For the coconut area were found values between 4,2 e 5,9 mm dia-1, for the banana area values between 4,40 e de 6,5 mm dia-1. When comparing these evapotranspiration values with those obtained by the Penman-Monteith method at the coconut area it was observed high differences among the obtained values for the different methodologies, mainly for the May 15th and June 16th dates (16,6 23,8 e 3,3%). When the comparison was conducted for the banana area, the differences found between the methods were 6,0, 7,4 e 8,4 %, intervals registered and accepted in the Bibliography. The SEBAL algorithm is presented as a promissory tool for studies related with irrigation and water resources management.