Anshul Yadav, Sumit Sen, Luca Mao, Wolfgang Schwanghart

• A dataset of 2184 field measurements reported in the literature was used to evaluate the predictive capability of eight conventional flow resistance equations to predict the mean flow velocity in gravel-bed rivers. The results reveal considerable disagreement with the observed flow velocities for relative submergence less than 4 and for the non-uniformity of the bed material greater than 7.5 for all the equations. However, the predictions made using the Smart and Jaggi (1983), Ferguson (2007), and Rickenmann and Recking (2011) equations were closer to the observed values. Furthermore, bedload sediment transport also reduces the predictive capability of the equations considered in this study except for the Recking et al. (2008) equation, which was developed consid- ering active bedload transport. The performance of flow resistance equations improves when corrected by considering the geometric standard deviation of the bed material. Here we present an empirical approach using the whole dataset and its subsets for accounting for theA dataset of 2184 field measurements reported in the literature was used to evaluate the predictive capability of eight conventional flow resistance equations to predict the mean flow velocity in gravel-bed rivers. The results reveal considerable disagreement with the observed flow velocities for relative submergence less than 4 and for the non-uniformity of the bed material greater than 7.5 for all the equations. However, the predictions made using the Smart and Jaggi (1983), Ferguson (2007), and Rickenmann and Recking (2011) equations were closer to the observed values. Furthermore, bedload sediment transport also reduces the predictive capability of the equations considered in this study except for the Recking et al. (2008) equation, which was developed consid- ering active bedload transport. The performance of flow resistance equations improves when corrected by considering the geometric standard deviation of the bed material. Here we present an empirical approach using the whole dataset and its subsets for accounting for the additional energy losses occurring due to the wake vortices, spill losses, and free surface instabilities occurring due to the protrusions from the bed. The results obtained using the validation dataset shows the importance and usefulness of this approach to account for the additional energy losses, especially for the Strickler (1923) and Keulegan (1938) equations.…