Modelling daily evapotranspiration using artificial neural networks under hyper arid conditions

Abstract

Precisely determined evapotranspiration (ET) is necessary for maximization of water beneficiary use and hydrologic applications, particularly in arid and semiarid regions where water source is so limited, such as Saudi Arabia. Evapotranspiration is a complex, nonlinear process. However, data driven techniques can be used model it without requiring a complete understanding of the physics involved. Therefore, the Artificial Neural Networks (ANN) technique was used to estimate the daily reference evapotranspiration (ETref). Eight combinations of eight climatic parameters and crop height were used as input. The daily climatic variables were collected by 13 meteorological stations from 1980 to 2010. The ANN models were trained on 65% of the climatic data and tested using the remaining 35%. The generalised Penman-Monteith (PMG) model was used as a reference target for evapotranspiration values, with hc varies from 5 to 105 cm with increment of a centimeter. The developed models were spatially validated using climatic data from 1980 to 2010 taken from another six meteorological stations. The results showed that the eight ETref models developed using the ANN technique to estimate ETref varies in significance depending on the climatic variables included. The more input climatic parameters included, the more accurate the ANN model is. The statistical performance criteria values such as determination coefficients (R2) ranged from as low as 67.6% for ANN-MOD1, where air temperature is the only climatic parameter included, to as high as 99.8% for ANN-MOD8 with which all climatic parameters included. Furthermore, an interesting founded result is that the solar radiation has almost no effect on ETref under the hyper arid conditions. In contrast, the wind speed and plant height have a great positive impact in increasing the accuracy of calculating the daily reference evapotranspiration.