Performance evaluation of FAO Penman-Monteith and best alternative models for estimating reference evapotranspiration in Bangladesh

Abstract

Proper assessment of reference evapotranspiration (ET0) is necessary for pastoral activity and water management. The Penman-Monteith FAO56 (ETpmf) method has been recommended as the identical ET0 estimation model; nonetheless, it belongs to a vast climatic data requirement. There is an urgent need to discover an ideal alternate model for evaluating ET0 in particular places where all climatic data is insufficient. The performances of 15 empirical models were assessed to get the best alternative model by comparing it with the PMF-56 model. These 15 models were evaluated by employing a daily scatter plot and three well known numerical approaches: relative root-mean-square error, mean absolute error and Nash–Sutcliffe coefficient in this study. Furthermore, a linear regression model was implemented to calibrate and validate the empirical models' performances throughout the 1981–2005 and 2006–2018 time intervals, separately. The outcomes displayed that the ETpmf rose primarily and declined later on a monthly period with the topmost amount in April and the lowermost amount in January. Overall, the Abtew model was the best alternate method showing the highest determination coefficient values more than 0.85 from January to December. In contrast, the Penman, WMO, Trabert, Valiantzas1, Valiantzas2, Valiantzas3 and Jensen-Haise models presented moderate performances with fewer inaccuracies. Afterwards, modification, the version of the above-described models every month has been upgraded deliberately related to actual. The Abtew model had simplicity in the computation process, only used maximum temperature and solar radiation data and linearly well connected to the PMF-56 model.

Keywords

• Reference evapotranspiration
• Penman-Monteith
• Empirical models
• Performance evaluation
• Model modification