Reference evapotranspiration estimation in a highly advective semiarid environment

Hourly evapotranspiration (ETo) rates were measured during the irrigation season in a highly advective area in Southern Spain, using a precision weighing lysimeter. Close to the lysimeter, an automatic weather station was located to register hourly values of the most relevant climatic variables. Several methods to estimate ET, were evaluated for hourly and daily estimates. Accuracy was assessed from ordinary regression and from error analysis of the comparisons against measured values. FAO-24 methods showed a strong tendency to overestimate throughout the whole range of evaporation. Ritchie-type and Hargreaves equations had a similar behavior, underpredicting for ETo > 5.3 mm day(-1) and vice versa. Penman locally adjusted and ASCE-PM performed the best, followed by FAO-PM. Penman-63 also gave excellent daily estimates. The poor behavior of the Priestley-Taylor equation improved after applying the Jury and Tanner correction, but still remains at a very high noise level. With the exception of ASCE-PM and FAO-PM, the rest of the methods showed a tendency to underpredict more with increasing advective intensities. However, that tendency is very mild in the case of Penman locally adjusted.