Artificial neural networks for precipitation prediction: A case study on Eğirdir

The aim of this study was to develop an optimum precipitation prediction method, based on artificial neural network (ANN). The methodology was applied to precipitation predicting in Eğirdir in the Lake´s Districts of Turkey. In application, Eğirdir precipitations were predicted from Isparta and Senirkent precipitations. Each precipitation stations is located in same region. For monthly precipitaion predictions, data were taken from Turkish State Meteorological Service. Used data covered 36 years period during 1975-2010 for monthly precipitations. The ANN models had only one output but different numbers of input variables were examined. The comparison of historical records and ANN models showed a better agreement between the ANN models estimations and measurements of monthly precipitations. With the help of ANN model for integrated precipitaton prediction, it was is possible to estimate missing or unmeasured data and it was good at prediction on both low and high precipitations.