Near-bed fluid-particle interactions during aeolian saltation

Bagnold's (1936) model is still the basis for most commonly used aeolian sand transport models that predict transport as a function of shear velocity. These time-averaged models ignore turbulence characteristics and tend to over predict sand transport rates compared to direct observations of sand flux. To improve our understanding and modeling of aeolian sand transport, we examine unexplored questions on the coupling between fluid and particle motion at turbulence time scales in a controlled wind tunnel simulation.

Analysis of high-speed imagery of sediment transport has allowed researchers to directly measure fluid-particle interactions with increasing spatial and temporal resolution. Using particle imaging/tracking velocimetry (PIV/PTV), we explore whether the variability in saltation is random or correlated with turbulence. A high-speed video camera was used to capture the motions of sand grains and flow seeding material in streamwise and vertical vectors. Flow behavior was visualized using seeding particles that were released 1.5 m upwind from a 2 mm laser sheet. A 3D sonic anemometer was used to validate flow velocities derived from PIV analysis at 15 cm, and a vertical array of sand traps was used to validate estimates of sand flux. Ripple dimensions were measured using a Riegl VZ-400i terrestrial laser scanner.

High-resolution observations of the fluid velocity field and resulting sand transport across varying wind speeds and ripple formations were used to examine the relationships between spatial and temporal fluctuations in the flow field (turbulence) and grain transport. We then compare sand and seed particle motions near the bed using spectral analysis and correlations with turbulence. Results over flat and rippled beds include the direct observations of (1) vertical saltation concentration profiles, (2) vertical velocity profiles, and (3) turbulence statistics (i.e., turbulence intensity, Reynolds shear stress, TKE).