Challenge in Flood Forecasting Using Rating Curve of Unstable River Systems: The Brahmaputra

A reliable and comprehensive flood warning system is one of the key factors to minimize damage and disruption create by a flood. Ganges and Brahmaputra is the major river system of Bangladesh and authorities of Bangladesh use gauging data of this two rivers for their flood forecasting. This study only focuses on Brahmaputra River, which is still a morphologically very active river. This paper discusses how the prediction can be biased using conventional method for flood prediction, i.e. using only stage-discharge relationship. In this study, we used 1,687 observations (discharge, water level, river width and cross-sectional area) at Bahadurabad station of Brahmaputra River selected from 1976 to 2009. We examined the stationarity of river cross-section and water level-discharge relationship. The study shows that the relationship between water level and river flow changes every year significantly and without any indication which way it may change in the next year. The relationship of stage discharge is so wide that the river may experience approximately double of the calculated flow. Therefore, river water management and flood forecasting would require comprehensive method rather than relying only on stage discharge correlation. Moreover, there is also trend in river geometry which may create more uncertainty in flood forecasting