Estimating potential evapotranspiration using Fuzzy Logic and Neural Network

Potential evapotranspiration (ETp) rates are needed for irrigation scheduling. ETp rates are commonly estimated from weather parameters. The Fao-Penman Monteith, is now accepted for computation of ETp. This study examined the suitability of fuzzy logic for estimating daily potential evapotranspiration with grass reference crop compared with artificial neural networks, Fao-Penman Monteith and evapotranspiration pan. The daily climatic data of the Borojen station in Chaharmahale-Bakhtiari was used. The estimated ETp values from a fuzzy logic model using six input parameters resulted in RMSE=0.74 mm/day, R2=0.88 which has the highest correlation and the lowest error in comparison with other methods.