Estimating bedload transport in a large sand–gravel bed river from direct sampling, dune tracking and empirical formulas

Three methods used to estimate river bedload discharges were compared in a large sand–gravel bed river. The unit and total bedload transport rates estimated by sediment sampling, dune tracking method and the equations of Van Rijn and Meyer-Peter and Müller were compared. The analysis was based on a large data set obtained from field surveys that combined sediment sampling, acoustic doppler profiler (ADP) measurements, and multibeam echosoundings of the middle reach of the River Loire (France) for contrasting flow conditions ranging from low flows to 2-year floods. For transport stages between 2 and 6 (i.e. between the annual mean discharge and the discharge for a 2-year flood), the tested equations predicted fairly well the unit and total bedload discharges and roughly estimated the temporal variability of the bedload during floods. For these transport conditions, the best agreement with the sediment sampling measurements was observed with the Van Rijn formula. For lower flow conditions, the tested formulas provide lower estimates than those sampled. To improve the accuracy of the equations for low flows, potential solutions include integrating near-bed turbulences and non uniformity of bed particles into the models. The ability of the formulas to predict the temporal variability of the bedload can be improved by using the bedform roughness to adjust the estimates. The bedload transport rates calculated with dune tracking method are lower than the fluxes sampled because a substantial fraction of the bedload does not participate to the dune migration. Finally, a first rating curve that was implemented for the River Loire estimates the annual bedload discharge for 2010 to be 480,000 t.