Evaluation of sand transport models by in situ observations.

Abstract

Sand transport rates measured from nine deployments of an annular benthic flume off the western Newfoundland coast and Sable Island Bank (Canada), are compared with predictions using: (1) Engelund and Hansen's total load equation; (2) Einstein-Brown's bedload equation; (3) Bagnold's total load equation, modified by Gadd et al.; and (4) Yalin's bedload equation. TheD grain diameters examined were: 0.19mm(15 trials); 0.33mm(37 trials); and 0.50 mm (7 trials). Results showed that equation (1) under-predicted in about a 40% for all the grain sizes. Equations (2) and (4) showed almost the same range of under-prediction: approximately 30% for D = 0.19 mm, 40% for D = 0.33 mm and 50% for D = 0.5 mm. The maximum variations on the observed-predicted ratios were obtained by equation (3), over-predicting by a factor of 2.45 and 1.9 for D = 0.19 mm and D = 0.33 mm, respectively, and under-predicting by a factor of 0.42 for D = 0.5 mm, but showing the highest correlations among all the tested models (R = 0.96 and 0.8 forD = 0.19mmand 0.33 mm, respectively). These variations are related to the inclusion of outliers in the original computation of the grain size dependent -parameter employed by equation (3). Based on the obtained results and through the analysis of published data, a significant improvement on the formulation (3) is suggested.