Using backpropagation neural networks for flood forecasting in PhraNakhon Si Ayutthaya, Thailand

This research created models and an application for predicting the water levels at the gauging stations C.35 located at the Chao Phraya River, PhraNakhon Si Ayutthaya, Thailand by using backpropagation neural networks. The ranges of forecasting are one day, two days, and three days in advance. Rainfall, the water levels, and the water flowing rates in the Chao Phraya River measured from the gauging stations C.2, C.13, C.35, C.36, and C.37 collected in the year 2008-2010 were used for developing the water-level forecasting models. All created models were validated in term of mean square error (MSE), correlation coefficient, and tested in term of model efficiency index and MSE. The best model produced 90.1218% of accuracy and was employed in our flood warning application.