Soil water stress co‐efficient for estimating actual evapotranspiration of maize in northeastern Brazil

Several of the methods used to estimate actual evapotranspiration ( ET a ) consider the crop evapotranspiration ( ET c ) to be affected by a soil water stress co‐efficient (K s ) such that ET a = K s ET c . The aim of this study was to assess the ET a estimation in maize in Alagoas State (09 ° 38 ′ S and 36 ° 40 ′ W; 260 m msl) in northeastern Brazil using the method of the single crop co‐efficient from the FAO ‐56 paper, combined with three different approaches to determine K s . The K s was expressed as either a linear (K s,lin ) or a logarithmic (K s,log ) function of the soil water depletion. Additionally, the method proposed in the FAO ‐56 (K s, FAO ) was used when the impact was linear, but it was switched on after only one offset. Meteorological and soil water measurements were conducted in a maize crop from July to October 2005. The estimates of ET a were compared with the ET a determined from soil water balance. The ET a estimated with K s,lin had the highest precision ( r 2 = 0.67) and accuracy ( d = 0.89 and root mean square error ( RMSE ) = 0.67 mm d −1 ). However, the ET a determined from K s,log ( r 2 = 0.41, d = 0.57 and RMSE = 1.69 mm d −1 ) or K s, FAO ( r 2 = 0.46, d = 0.62 and RMSE = 1.73 mm d −1 ) was not estimated successfully. When the parameters of K s, FAO were fitted to the local edaphoclimatic conditions, the precision (0.67) and accuracy (0.89 and 0.82 mm d −1 ) were similar to the estimates with K s,lin but with a better performance for ET a > 3.0 mm d −1 .