This research was carried out to evaluate nine reference evapotranspiration, and find out an alternative models to the standard FAO Penman-Monteith model in Traghen region southwest of Libya. The models applied were standard FAO Penman-Monteith, FAO Radiation, FAO Blaney-Criddle, Hargreaves-Samani, Priestley-Taylor, Makkink, Turc, Thornthwaite, Kharrufa and McCloud. The models were compared with the FAO Penman-Monteith model using root mean square error (RMSE) , mean bias error (MBE), Pearson type goodness of fit index (R2), refined index of agreement (dr), modeling efficiency (Ef) and t-test. Results showed that FAO Radiation and FAO Blaney-Criddle models overestimated ETo by values ranged from 0.64% to 16.06%. However, Kharrufa and McCloud models overestimated ETo in some months and underestimated ETo in some other months. Whereas, the other models underestimated ETo by values ranged from -86.69 % to -6.02%. The FAO radiation model gave the highest dr (0.892) and Ef (0.951) values, and the lowest RMSE (0.534 mm/day), indicating that this model was the best alternative to the FAO Penman-Monteith model in the study region, followed by FAO Blaney-Criddle model with dr value of 0.851, Ef value of 0.878, and RMSE value of 0.845 mm/day. In addition, FAO radiation model showed the second best R2 and MBE values at 0.993 and 0.473, respectively Also, FAO Blaney-Criddle model showed the third best R2 value at 0.989. Therefore, FAO radiation model is ranked the first and FAO Blaney-Criddle model is ranked the second. According to the statistical measures stated above, Turc model ranked the third, Kharrufa model ranked forth. Models of Hargreaves-Samani, Makkink, Priestley-Taylor, McCloud and Thornthwaite ranked fifth, sixth, seventh, eighth and ninth, respectively. T-test analysis at 5% level of significance indicated that there is a significant difference between the FAO Penman-Monteith model and all models except Kharrufa and McCloud models.
Read full abstract