Prediction of total monthly rainfall in Jordan using feed forward backpropagation method

Freiwan M., Cigizoglu H.

FRESENIUS ENVIRONMENTAL BULLETIN, vol.14, no.2, pp.142-151, 2005 (SCI-Expanded) identifier

  • Publication Type: Article / Article
  • Volume: 14 Issue: 2
  • Publication Date: 2005
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED)
  • Page Numbers: pp.142-151
  • Istanbul Technical University Affiliated: No


The main purpose of this study is to predict the monthly precipitation amount using Artificial Neural Networks (ANNs). Feed Forward Back Propagation (FFBP) method is used to train the ANNs using the actual monthly precipitation data of Amman Airport meteorological station. The k-fold partitioning method was applied to the monthly total rainfall data. Different ANN configurations were tested in order to obtain the best prediction performance. The ANN models have provided a good fit with the observed actual data. The contribution of the periodic component in the input layer has remarkably improved the performance of the ANNs in estimating total monthly rainfall. A comparison between the predictions of the best-fit ANN models and the estimations by conventional stochastic models resulted in favor of the ANNs in rainfall forecasting.