Experiments were made in a small flume to test the validity of scale-model parameters describing transport of loose sediment. Using seven variables to characterize flow and sediment transport (mean flow depth d, mean flow velocity U, fluid density p and viscosity μ, sediment size D and density p_s, and gravity g), dimensional analysis provides four model parameters: Reynolds number pUd/μ, Froude number U/(gd)^0.5, size ratio d/D, and density ratio p_s/p. Two scaled runs were made in the ripple regime at a scale ratio of about 1.7 to 1 using sand and water with appropriate grain sizes and viscosities. Frequency distributions of ripple height, spacing, and migration rate were determined for 100 ripples; the test for scaling is whether the two curves for each of these variables coincide when the appropriate scaling factor is applied. In each case the curves are closely coincident, indicating that scale modeling of sediment transport by variation of grain size and water temperature is feasible. This technique should be valuable in studying bed forms and transport rates in small to medium-sized rivers by means of scaled hot-water runs in laboratory flumes.

A pilot study was made in an insulated flume 0.9 m wide and 11 m long with water temperatures up to 80°C (providing a scale model ratio of up to 2.3) with sand sizes ranging from effectively 0.11 mm to 0.30 mm and a flow depth of up to effectively 0.5 m. Sand waves are present at flow velocities intermediate between those for ripples and for dunes, with heights comparable to those of ripples but with significantly longer spacings. Dune spacing increases with increasing flow velocity; dune height increases with increasing flow velocity, then decreases as velocities for upper flat bed are approached. The velocity interval for stable sand waves and dunes decreases with decreasing sand size, and disappears at about 0.11 mm.