spatial and temporal evaluation of different methods for prediction of reference evapotranspiration (Case study: Khuzestan province)

The correct estimation of reference evapotranspiration (ET0) is considered in order to accurately monitor the water requirement of crops and water balance in large scale irrigation and drainage plans.The present study aim to evaluation of spatial and temporal of twenty-four different methods to predict ET0 based on air temperature, solar radiation and mass transfer models, in 11 stations of Khuzestan province in Iran (1996-2005). The results were compared with the result of the FAO Penman – Monteith method (PMF-56) with use of statistical indices: R2, RMSE and PBIAS. Also the best map of each category was provided in the ArcGIS software. The results showed that, the methods have different performance in different seasons, but the performance of the models at most in the methods is satisfactory, this can be due to the same climate type and geographic location. Also according to the annual results, among the methods based on air temperature, Blaney and Criddle methods with R2=0.92 and RMSE=2.9 and Hargreaves and Samani methods with R2=0.91 and RMSE=2.2, among the methods based on solar radiation, Doorenbos and Pruitt methods with R2=0.93 and RMSE=3.0 and among the methods based on mass transfer, WMO method with R2=0.97 and RMSE=5.0, had the best performance to predict reference evapotranspiration comparing with PMF-56 in Khuzestan province. Finally, according to the annual, results the based on air temperature methods were the most accurate in Khuzestan province. It is notably, temporarily, summer season was the least accuracy in all methods of temperature, radiation and mass transfer.