A new apparent shear stress-based approach for predicting discharge in uniformly roughened straight compound channels

Accurately predicting flow discharge in a compound river channel becomes increasingly important when in river eco-environment design and flood risk management, in which the conventional methods, such as divided channel method (DCM) and single channel method, do not meet precision requirement. In the present paper, a new approach is proposed to improve the prediction of flow based on the concept of the apparent shear stress at the diagonal interface between main channel and floodplains. The new approach is compared with a wide range of the author's experimental data and the data available in the literature. The 27 datasets used in this study cover homogenous symmetric channels (22 datasets) and asymmetric channels (5 datasets), which have various width ratios [channel total width (B) at bankfull / main channel bottom (b) = 1.5 ~ 15.8] and bed slopes (S _o = 2.6 × 10 ⁻⁴ ~2.5 × 10 ⁻³ ). It was found that the new approach can significantly improve the accuracy of flow prediction compared with the conventional DCM for all the datasets, particularly for relatively low flow depths of floodplain (e.g. Dr < 0.15) where the flow discharge was very difficult to predict correctly. The new approach predicts the total discharge well for straight homogeneously symmetric and asymmetric channels, within a mean absolute percentage error (MAPE) of 5%. Furthermore, compared with the DCM, the new approach can also have a better prediction of zonal discharge percentage and it has less variation of discharge prediction error over the flow depth.