Reference Evapotranspiration Estimation from Class «A» Pan in the Northwest of Uruguay

Efficient water management in agriculture requires accurate estimation of the evapotranspiration. The difficulty in obtaining records of all the variables needed to estimate reference evapotranspiration (ETo) by the model proposed by PenmanMonteith-FAO56 (Allen et al., 1998), leads to the use of alternative methods, such as pan evaporation class «A» (Eo), which requires regional calibrations for successful implementation. This paper compares four methods for estimating the adjustment coefficient (Kp) of Eo over ETo: i) the procedure proposed by Snyder (1992); ii) by Allen et al. (1998); iii) by Puppo and García Petillo (2009); and iv) the Kp-combined obtained through the regression coefficient of Eo to ETo for the region under study. The regression coefficient values of Eo over ETo change slightly according to the average interval used, being 0.77, 0.75 and 0.73 for the monthly interval, decadic (10 days) and weekly. The estimation error decreases as we average at longer intervals. The procedure i) significantly overestimated ETo, while ii) underestimated both, with a mean absolute error of 0.49 and 0.86 mm respectively, while in iii) is 0.32 and the iv ) is 0.37 mm. Procedures iii) and iv) with constant values of Kp had the best performance for the northwestern region of Uruguay.