Application of Multivariate Regression and Gene Expression Programming in Modeling Reference Evapotranspiration (Case Study: Khorramabad Station)

Accurate estimation of water requirements of plants is a key factor in controlling several hydrological processes including: planning and management of water resources, especially in arid and semi-arid regions (Laaboudi et al., 2012; Wen et al., 2015) water pricing and water requirement for Irrigation (Yassin et al., 2016). In this study, multivariate regression methods and gene expression planning were evaluated to estimate reference evapotranspiration. For model input data, the Khorramabad Synoptic Station information including: maximum and minimum temperatures, maximum and minimum relative humidity, sunny hours and monthly wind speeds in the range of 1983-2017(420 months) were used. Based on the relationship between input and output parameters, six input patterns were determined for modeling.70% of the data were used for training and 30% were used for model validation.The results of multivariate regression showed that the proposed model had acceptable accuracy with R2 = 0.952.The analysis of model coefficients showed the greatest effect of maximum temperature with a coefficient of 0.604 on reference evapotranspiration. Gene expression planning results showed that the fifth pattern with four main operators was R2 = 0.958 and RMSE = 0.704 in the training phase and R2=0.977 and RMSE = 0.615 in the test phase had better performance.