Modelling evapotranspiration using discrete wavelet transform and neural networks.

摘要: This study combines wavelet transforms and feed-forward neural network methods for reference evapotranspiration estimation. The climatic data (air temperature, solar radiation, wind speed, relative humidity) from two stations in the United States was evaluated estimating models. For (WNN) model, input decomposed into sub-time series by transformation. Later, new (reconstructed series) are produced adding available components these reconstructed used as of WNN model. phase is pre-processing raw main different performance model compared with classical networks approach [artificial (ANN)], multi-linear regression Hargreaves empirical method. shows that could be applied successfully modelling data. Copyright © 2009 John Wiley & Sons, Ltd.