Cross-station assessing the gene expression programming-based approaches for modeling reference evapotranspiration

Daily reference evapotranspiration (ET0), was simulated in six weather stations of Iran, through using the daily meteorological variables of ten years. The data sample consisted of air temperature, sunshine hours, wind speed and relative humidity. The gene expression programming (GEP)-based models were then applied for assessing the ET0 modeling capability through the by station as well as cross-station application scenarios. In the first part, the applied models were trained and tested by using recorded independent variables. The input configurations which were used for feeding GEP were based on similarities with the Hargreaves-Samani, Priestly-Taylor and FAO56-PM models. The results revealed that the GEP shows higher capability in comparison to the original as well as the calibrated ET0 models for simulating the by station and cross-station application data management scenarios. Although the by station approach would lead to higher degree of accuracy (as could be anticipated), the cross-station approach would provide opportunity for assessing ET0 values in the stations with scarce meteorological data

Keywords:

Evapotranspiration , GEP , cross-station application