Efficiency of methods for calculating class A pan coefficients to estimate evapotranspiration reference

The class A pan coefficients (Kp) can be determined with several methods, most based on wind speed, relative humidity and conditions and extension of the border which circles the pan. This study aimed at evaluating the performance of methods used for estimating the Kp, as well as an average and constant Kp (0.70), to evaluate the reference evapotranspiration (ET0) under Brazilian Savannah conditions, in Santo Antônio de Goiás, Goiás State, Brazil. The methods were assessed on an annual basis and during the dry and rainy seasons. The ET0 resulting from the class A pan evaporation, multiplied by the Kp values estimated with the aid of the Doorenbos & Pruitt (1977), Cuenca (1989), Snyder (1992), Pereira et al. (1995) and Allen et al. (1998) methods, were compared with the ET0 estimated by using the Penman-Montheit equation. The coefficients of determination (r2), correlation (r), Willmott index of agreement (d) and performance (c), as well as the mean and maximum absolute errors, were also estimated. Considering the weather conditions during the experiment, the best Kp method to estimate the annual ET0 was the one proposed by Pereira et al. (1995). For the dry season, the method proposed by Cuenca (1989) performed better, while, for the rainy season, all methods showed a low performance, with the Pereira et al. (1995) method being the most efficient one. The use of a constant Kp of 0.70 resulted in a good performance and showed to be a practical option, however, it is necessary to determine it where it will be applied