Reliability of flood discharge estimates

The high accuracy of normal river discharge measurements is contrasted with the approximate methods used to estimate large floods. The most commonly used method, the slope–area method, is examined and it is argued that there is likely to be a strong upward bias in these flood estimates. Sources of error are subdivided into four main types: (i) error in the Manning friction coefficient, (ii) error in after-flood measurements of channel cross section, (iii) estimation of surface slope and energy grade line from high-water marks, and (iv) the effects of erosion and sediment transport. These various factors are seen to be influenced by channel curvature and meandering, floodplain flow and interaction with the main channel, and changes to channel geometry during a flood. From theoretical work on high sediment transport, it is argued that up to one third of the channel slope is required for moving the sediment, so that only two thirds of the slope is available for transporting the water. Key words: floods, rive...