Flood Forecasting Using Artificial Neural Network

Abstract The process of assessing the timing, amount, and period flood events based on observed features a river basin is known as forecasting. Floods cause lots damage to properties create risk human life. Flood forecasting critical for developing appropriate management strategies, reducing hazards, evacuating people from flood-prone areas. main objective this study apply artificial neural networks flow in Deo River, located Gujarat. Rainfall discharge are parameters considered model development. developed validated test accuracy model. Trained models evaluated using performance indices. Six alternative prediction have been ANN. These various training algorithms. A single layer feed forward back-propagation network with six different algorithms (Scaled conjugate gradient, Levenberg Marquardt, Resilient back-propagation, Conjugate Cascade back propagation, Bayesian regularization) was developed, 70% data used 30% validation. created models’ assessed statistical parameters. best obtained an ANN algorithm, which had coefficient correlation (r) 0.83, determination (R 2 ) 0.70, root mean squared error (RMSE) 5.58 0.89, 7.27 forecast inflow very close values. This shows that can be successfully predict floods, by control departments across country