Please use this identifier to cite or link to this item:
|Title: ||Forecasting water stage at different lead-time by artificial neural network combined with flash flood routing model|
|Authors: ||WEN-CHENG LIU;CHUAN-EN CHANG;Chih-Chieh Young|
|Issue Date: ||2017-10-24T07:28:12Z
|Publisher: ||Taiwan Water Conservancy|
|Abstract: ||Abstract:Accurate forecasts of water stages during typhoon periods are critically important for flood control operations. The present study forecasts water stages for 1, 2, 4, and 6-hour ahead at the Taipei Bridge in the Tamsui River system. To improve the accuracy of forecasting, the water stage is calculated by the flash flood routing model at the first step and then the flash flood routing model is combined with two artificial neural network models (ANN), including the Back-Propagation Neuron Network (BPNN) and the Adaptive Network-based Fuzzy Inference System (ANFIS). Five and three typhoon events were used for model calibration (training) and verification, respectively. Six statistical indicators, including root mean square error (RMSE), mean absolute error (MAE), coefficient of efficiency (CE), index of agreement(IOA), error of peak water level (ELp), and Error of time to peak water level (ETp), are used to evaluate the performance of the combination with flash flood routing model and artificial neural network model. The results indicate that water stage forecasting using the combination of flash flood routing model and ANFIS model is better than that using the combination of flash flood routing model and BPNN model. Besides, both hybrid models reveal that the accuracy of water stage forecasting decreases as the lead-time increases.|
|Appears in Collections:||[海洋環境資訊系] 期刊論文|
Files in This Item:
There are no files associated with this item.
All items in NTOUR are protected by copyright, with all rights reserved.