Evaluation of bed load transport formulae in a large regulated gravel bed river: The lower Ebro (NE Iberian Peninsula)

Abstract

This paper tests the predictive power of 10 bed load formulae against bed load rates obtained for a large regulated river (River Ebro) the armor layer of which is subject to repeated cycles of break-up and reestablishment. The theoretical principles of two of the 10 formulae explicitly include the effects of river bed armoring. The results obtained showed substantial differences in equation performance but no evident relationship between predictive power and theoretical approach (e.g., discharge, stream power and probability) was found. Overall, the predictive power of the tested formulae was relatively low. The average percentages of predicted bed load discharge that did not exceed factors of 2 (0.5 < r < 2) and 10 (0.1 < r < 10) in relation to the observed discharge were 19% and 57%, respectively (where r is the discrepancy ratio between the predicted and observed values). In particular, the formulae of Yang (1984) and Parker et al. (1982) presented the better levels of agreement with the observed bed load discharges. The bed load rating curve for the lower Ebro showed a similar degree of agreement to the best-performing formulae. However, its predictive power was limited because only flow discharge acts as an independent variable and river bed dynamics, such as armoring cycles, are not contemplated.