Using gene expression programming in monthly reference evapotranspiration modeling: A case study in Egypt

Abstract

The Penman-Monteith FAO-56 equation requires the complete climatic records for estimating reference evapotranspiration (ETo). The present study is aimed at developing and evaluating a gene expression programming (GEP) model for estimating mean monthly ETo by using minimal amount of climatic data. The data used in the analysis are collected from 32 weather stations in Egypt through the CLIMWAT database. The results showed that the accuracy of the GEP model significantly improved when either mean relative humidity (RH) or wind speed at 2-m height (u2) was used as additional input variables. The GEP model with the inputs as maximum and minimum air temperature, RH, and u2 showed the lowest root mean square error (0.426 mm d−1 and 0.430 mm d−1) and, the highest coefficient of determination, (0.963 and 0.962) overall index of model performance (0.960 and 0.960), and index of agreement (0.991 and 0.990) for training and testing sets, respectively. Comparing the results of GEP models with other empirical models showed that the GEP technique are more accurate and can be employed successfully in modelling ETo.