Reference evapotranspiration estimates in different cloudiness conditions

The objective of this work was to evaluate the performance of reference evapotranspiration (ETo) estimation methods, in different cloudiness conditions, in Seropedica, RJ, Brazil. The ETo estimates were compared with daily measurements made on weighing lysimeter, between 6/1/2006 and 7/31/2007, by pooling the data or discretizing them according to the variance of daily clearness index (K-T), in four classes: K-T <= 0.35, cloudy sky (Clo); 0.35<K-T <= 0.55, partly cloudy sky (PCDi); 0.55<K-T <= 0.65, partly opened (PCOp); and K-T>0.65, open sky (Op). The mean absolute error (MBE), the square root of the mean square error (RMSE) and indexes of adjustment and performance were used as quality indicators of the different methods. In the pooled data, Penman-Monteith FAO (PMF) and Hargreaves-Samani (HS) methods had 84.05 and 79.52% performance indexes, respectively, while Jensen & Haise, Linacre and Makking methods had performances below 60%. Changes in the cloudiness conditions affected the performance of ETo estimation methods. The best results were obtained with the solar radiation (86.1%) and Camargo (81.8%) methods, under the Op and Clo conditions. The PMF and HS methods can be employed in a complementary way, since their performance were 78.4% (Clo and PCDi) and 77.6% (PCOp and Op), respectively.