Field measurements of mean and turbulent airflow over a barchan sand dune

Advances in our knowledge of the aeolian processes governing sand dune dynamics have been restricted by a reliance on measures of time-averaged airflow, such as shear velocity (u⁎). It has become clear that such measures are incapable of explaining the complete dynamics of sediment transport across dune surfaces. Past evidence from wind tunnel and modelling studies has suggested that in some regions on a dune's surface the sediment transport might be better explained through investigations of the turbulent nature of the airflow. However, to date there have been no field studies providing data on the turbulent characteristics of the airflow around dunes with which to support or refute such hypotheses. The field investigation presented here provides mean and turbulent airflow measurements across the centre-line of a barchan sand dune in Namibia. Data were collected using arrays of sonic anemometers and were compared with sand flux data measured using wedge-shaped traps.Results support previously published data derived from wind tunnels and numerical models. The decline in mean wind velocity at the upwind toe of the dune is shown to coincide with a rise in turbulence, whilst mean velocity acceleration on the upper slope corresponds with a general decline in measured turbulence. Analysis of the components of Reynold shear stress (−u'¯w'¯) and normal stresses (u′2¯ and w′2¯) supports the notion that the development of flow turbulence along the dune centre-line is likely to be associated with the interplay between streamline curvature and mean flow deceleration/acceleration. It is suggested that, due to the nature of its calculation, turbulence intensity is a measure of less practical use than direct assessments of the individual components of Reynolds stress, particularly the instantaneous horizontal streamwise component (u′2¯) and shear stress (−u′w′¯). Whilst, increases in Reynolds shear stress and the horizontal streamwise component of stress in the toe region of the dune may effectively explain the maintenance of sand flux in a region of declining mean velocity, they have much less explanatory power for sand flux on the upper windward slope and in the crestal region of the dune. Here, it is suggested that mean flow acceleration is likely to provide the most significant driving force on sand flux, possibly augmented by a rise in the horizontal streamwise component of Reynolds stress (u′2¯) in the crest/brink region. Therefore, although wind turbulence is considered to be of fundamental importance in explaining the sediment transport dynamics across the dune's surface it is recognised that the interaction between mean flow deceleration/acceleration, streamline curvature and individual components of Reynolds stress is complex and the identification of a single element of flow that offers a panacea for accounting for sand flux and dune dynamics is difficult to find.