Determining large scale actual evapotranspiration using agro-meteorological and remote sensing data in the northwest of Sao Paulo State, Brazil

Abstract

The best irrigation management depends on accurate estimation of reference evapotranspiration (ET0) and then selection of the appropriate crop coefficient for each phenological stage. However, the evaluation of water productivity on a large scale can be done by using actual evapotranspiration (ETa), determined by coupling agrometeorological and remote sensing data. This paper describes methodologies used for estimating ETa for 20 centerpivots using three different approaches: the traditional FAO crop coefficient (K-c) method and two remote sensing algorithms, one called SEBAL and other named TEIXEIRA. The methods were applied to one Landsat 5 Thematic Mapper image acquired in July 2010 over the Northwest portion of the Sao Paulo State, Brazil. The corn, bean and sugar cane crops are grown under center pivot sprinkler irrigation. ET0 was calculated by the Penman-Monteith method with data from one automated weather station close to the study site. The results showed that for the crops at effective full cover, SEBAL and TEIXEIRA's methods agreed well comparing with the traditional method. However, both remote sensing methods overestimated ETa according to the degree of exposed soil, with the TEIXEIRA method presenting closer ETa values with those resulted from the traditional FAO K-c method. This study showed that remote sensing algorithms can be useful tools for monitoring and establishing realistic K-c values to further determine ETa on a large scale. However, several images during the growing seasons must be used to establish the necessary adjustments to the traditional FAO crop coefficient method.