Flood forecasting using artificial neural network (ANN) in Maran, Pahang
Maran is located at district of the same name between Temerloh and Kuantan, Pahang which is surrounded by remote forest and palm oil plantations. In December 2013, a worst flood occurred in Maran and had cause loss of lives and massive damages to the area. In this study, the data from Lubok Paku Sta...
| Main Author: | |
|---|---|
| Format: | Undergraduates Project Papers |
| Language: | English |
| Published: |
2014
|
| Subjects: | |
| Online Access: | http://umpir.ump.edu.my/id/eprint/11624/ http://umpir.ump.edu.my/id/eprint/11624/ http://umpir.ump.edu.my/id/eprint/11624/1/NUR%20ATIYAH%20DINIE%20BINTI%20MAT%20ARIFIN.PDF |
| Summary: | Maran is located at district of the same name between Temerloh and Kuantan, Pahang which is surrounded by remote forest and palm oil plantations. In December 2013, a worst flood occurred in Maran and had cause loss of lives and massive damages to the area. In this study, the data from Lubok Paku Station was used for analyzing data for flood forecasting. Lubok Paku Station in district Maran was one of the stations that are located along Sungai Pahang. In order to reduce the possibility for flood events to occur again in Maran, a reliable water level forecasting models is extremely important. Developing water level forecasting models is essential in water resources management and flood prediction. Accurate water level forecasting aids in achieve proficient and ideal utilization of water resources and minimize flooding damages. Conventional linear modelling forecasting model approach such as regression mostly provided relatively poor accuracy for forecasting peak inflow events of floods or drought. An artificial neural network or known as ANN is a computing model with a nonlinear mathematical approach that has been proven in many forecasting studies. Improving the ANN computational approach could help produce accurate forecasting results. ANN also has the ability to map inputs and outputs pattern and copy different elements experienced in the data. In this study, ANN Modelling is used to analyze the water level data. Different model architectures are examined based on the length of the forecasting period, epochs and training data. The data training was conducted using six ANN architectures. The performance of data training and data validation were evaluated using the coefficient of efficiency (NSC) and root mean square (RMSE). The results showed a strong performance of forecasting accuracy which can provides a significant tool for the authorities to take proper actions to minimize the damage of flooding. |
|---|