Evaluation of the Penman–Monteith reference evapotranspiration under limited data and its sensitivity to key climatic variables under humid and semiarid conditions

Abstract

The objectives of this study were to assess the accuracy of FAO Penman Monteith equation (FAO PM) under limited data conditions and to perform sensitivity analysis to determine approximately the change in reference evapotranspiration (ETref) expected for a known change in one of the independent variables and derive the sensitivity coefficient. Meteorological data were collected from 8 weather stations under humid and semiarid condition in Cote d’Ivoire. The results showed good performance of FAO PM equation under missing solar radiation (Rs) in semiarid condition and under missing wind speed data (U2) and relative humidity (RH) in all locations with coefficient of determination (R2) ranging from 0.70 to 0.99 and regression slope from 0.99 to 1.05. Under missing Rs, RMSE varied from 0.45 to 0.48 mm/day and AME from 0.30 to 0.35 mm/day. The RMSE and AME vary respectively from 0.12 to 0.51 mm/day and from 0.09 to 0.30 mm/day under missing RH data, and respectively from 0.11 to 0.60 mm/day and 0.04 to 0.34 mm/day under missing wind speed data. The poor performance of FAO PM method to estimate ETo was observed when three climatic variables were missing with regression slope from 0.93 to 1.06 and R2 from − 0.06 to 0.26. The response of ETo to changes in all climate variables was linear, with high R2 values (≥ 0.99) in most cases. Any error in Rs, maximum temperature (Tmax) data would have contributed to significant change in ETo estimate. The effect of Rs on change in ETo estimates had the greatest slope (≥ 2.80) in Bouake, Daloa, Korhogo, Man, Seguela whereas it had the lowest slope in Ferkessedougou (slope = 2.74), Odienne (slope = 2.73), Yamoussoukro (slope = 2.77). The effect of Tmax in change in ETo was also important in all location except Daloa and Man with low regression slope values of 1.63, 1.74, respectively. All sensitivity coefficients showed a large degree of daily and seasonal fluctuations and revealed significant differences in northern and central study locations. The sensitivity coefficients of U2 and Tmax were greatest under semiarid condition while the one of Rs were very high in humid condition. Accurate measuring of U2, Tmax and Rs in estimating reference evapotranspiration using Penman–Monteith equation is required.