Monthly Reference Evapotranspiration Forecast Using CFS.v2 And Wavelet Neural Network (WNN)

Potential Evapotranspiration (ETo) is an important hydro-climate variable. It plays a significant role in many areas, for instance, in managing and planning for irrigation systems, rainfall-runoff process, river basin yield and reservoir capacity. In this study, a framework to forecast monthly ETo using the outputs of Climate Forecast System Version 2 (CFS.v2) model with the WNN post-processing approach was proposed. Since the accuracy of temperatureforecasts are usually higher than those of other climatic factors, in the current research, monthly temperature forecasts were utilized to forecast ETo. First, daily ETo was calculated from observed climatic data using the much-taunted FAO-PM56 method for the period of 2010 to 2017.These daily ETo data were then transformed to the mean monthly. In the following step, a onemonth lead time forecasted temperature data at the standard 2m height (minimum, maximum and average) for the same period, was extracted from the outputs of the model. Finally, theforecasted temperature data by the CFS.v 2 model for the next month and the calculated ETo using the observed climatic data were employed as inputs and outputs to the ANN and WNN, respectively. The results showed that both ANN and WNN are able to forecast ETo for the following month with good accuracy. However, it was found that the WNN was more robust. Keywords: CFS.v2, potential evapotranspiration, ANN, WNN, FAO-PM56, Urumia Lake Basin Iran.

Monthly Reference Evapotranspiration Forecast Using CFS.v2 And Wavelet Neural Network (WNN) Keywords:

CFS.v2 , potential evapotranspiration , ANN , WNN , FAO-PM56 , Urumia Lake Basin Iran.