Reference evapotranspiration for Londrina, Paraná, Brazil: performance of different estimation methods

Abstract

Aiming at assessing the performance of alternative methods to Penman-Monteith FAO56 for estimating the reference evapotranspiration (ETo) for Londrina, Paraná, Brazil, the methods temperature radiation, Hicks-Hess, Hargreaves-Samani (1982), Turc, Priestley-Taylor, Tanner-Pelton, Jensen-Haise, Makkink, modified Hargreaves, Stephens-Stewart, Abtew, global radiation, Ivanov, Lungeon, Hargreaves-Samani (1985), Benavides-Lopez, original Penman, Linacre, Blaney-Morin, Romanenko, Hargreaves (1974), McCloud, Camargo, Hamon, Kharrufa, McGuiness-Bordne, and Blaney-Criddle were compared to that standard method recommended by FAO. The estimations were correlated by linear regression and assessed by using the Person's correlation coefficient (r), concordance index (d), and performance index (c) using a set of meteorological data of approximately 40 years. The methods modified Hargreaves, Stephens-Stewart, Abtew, global radiation, Ivanov, Lungeon, Hargreaves-Samani (1985), Benavides-Lopez, original Penman, and Linacre should be avoided, as they did not present excellent results. The methods McCloud, Camargo, Hamon, Kharrufa, McGuinness-Bordne, Blaney-Criddle, Hargreaves (1974), Romanenko, and Blaney-Morin were classified as very bad, not being recommended. In contrast, the methods temperature radiation, Hicks-Hess, Hargreaves-Samani (1982), Turc, Priestley-Taylor, Tenner-Pelton, Jensen-Haise, and Makkink presented excellent performance indices and can be applied in the study region.