Evaluation of reference evapotranspiration models in coastal Mediterranean and mountainous regions of Libya

Abstract

This research was carried out to evaluate performance of nine reference evapotranspiration (ET₀) models against the standard FAO Penman-Monteith model  and identify suitable alternative models to FAO Penman-Monteith at mountainous and  coastal regions of Libya.   Shahat and Nalut sites  represents mountainous region and Benghazi, Sirt and Tripoli sites representing coastal region of Libya.  The models tested against the standard FAO Penman-Monteith model were the FAO Radiation, FAO Blaney Criddle, Hargreaves-Samani, Priestley-Taylor, Makkink, Turc, Thornthwaite, Kharrufa, and Khosla.  The results illustrate that FAO Radiation model overestimated standard ET_o values.  In contrast, the models of Hargreaves-Samani, Makkink, Turc and Thornthwaite underestimated standard ET_o values.  However, models of FAO Blaney-Criddle, Priestley-Taylor and Kharrufa overestimated standard ET_o values in some months and underestimated standard ET_o values in other months.  The statistical criterions used are index of agreement, modeling efficiency, root mean square error, mean bias error, % mean bias error, coefficient of determination and t-test.  Statistical results reveal that the FAO Blaney-Criddle model had superior performance among the other models.  FAO Blaney-Criddle model ranked the first in Mountainous and coastal regions except in Tripoli site.  Thornthwaite and Khosla models performed poorly in the all study sites.  All other models performed irregularly different.  T-test analysis at 5 % level of significance indicated that there are significant differences between the standard FAO Penman-Monteith and all models in both regions, except FAO Blaney-Criddle and Kharrufa models in Shahat and Nalut sites, Priestley-Taylor model in Sirt site and FAO Radiation model in Tripoli site.