Search

Bangladesh is land of rivers with about 700 km shoreline at the Bay of Bengal. More than 300 rivers flow over the land of Bangladesh having an area of about 1,47,000 km. A wide part of this country frequently face flood hazard caused by the excessive flow from three major rivers named the Ganges, the Brahmaputra and the Meghna due to heavy monsoon rainfalls on the upstream catchment area. The control of water flow is an international issue shared among Bangladesh, Nepal and India. Among 57 trans-boundary rivers, 54 enter into Bangladesh from India. Seasonal flood management is a regular exercise and costly event in Bangladesh. Among the several tools of flood management, flood forecasting is a major non-structural measure to protect the people and property from the damage of flood. Hence, the water experts of Bangladesh use several water models to assess the flood. All those are hydrodynamic models based on the mathematical framework of the elliptic differential equations. The hydrodynamic models require huge and reliable data to run the model and predict the flood elevations at river cross sections ahead of time. The models are complex and require calibration and validations for geophysical and morphological changes of river cross sections. Flood Forecast Warning Cell forecasts the flood levels during the monsoon period in more than 80 locations spread over the Bangladesh in major rivers which are generated based on boundary water levels at the Ganges, Brahmaputra and estimated rainfalls using rainfall runoff and river models. In order to circumvent the scarcity of reliable and accurate hydrological data, reduce run time and to make faster and simpler flood level prediction, this research investigated a method of using Artificial Neural Network (ANN) to generate the water levels along the rivers at selected locations. The proposed ANN can predict the incoming flood properly using boundary water levels at Bahadurabad in Brahmaputra and at Harding Bridge in the Ganges. The developed ANN has been given the name as 'Bangladesh River Artificial Neural Network System' and used to provide an early flood warning system. [ABSTRACT FROM AUTHOR]