PRECIPITATION PREDICTION MODEL WITH GENETIC EVALUATIONARY PROGRAMMING

35851

The aim of this study was to develop an optimum precipitation prediction model, based on genetic evaluationary programming (GEP) and artificial neural network (ANN). The methodologies were applied to predictprecipitation in Eğirdir located in the Lakes District of Turkey. The precipitation values of Eğirdir station were predicted using precipitation values of Isparta and Senirkent stations located in same region. For monthly precipitaion predictions, the data were taken from Turkish State Meteorological Service. The used data covered 36 years period during 1975-2010 for monthly precipitations. The GEP and ANN models were developed using different combinations of input variables. The comparison of historical records and models showed a better agreement in the GEP models than ANN models. With the help of GEP model for integrated precipitaton prediction, it is possible to estimate missing or unmeasured data and it wasgood at prediction of min and max precipitations.

8

Monthly precipitation , Genetic Evaluationary Programming , Artificial Neural Networks , Eğirdir

Sdu International Technologic Science

21

Sdu International Technologic Science