Explaining the surprisingly poor correlation between turbulent surface wind and aeolian sand flux

Existing models of aeolian sand transport, derived theoretically and from wind-tunnel experiments, often disagree substantially with field observations. Despite advancements in anemometry and sediment flux detection technologies, even very high-resolution observations of aeolian sand transport show only weak correlation with concurrent surface wind speeds and model predictions. Unlike in experiments and numerical models, winds in natural environments exhibit turbulent fluctuations over a broad range of length scales extending from individual grains to the top of the atmospheric boundary layer and over a similarly large range of time scales. Here, we present simultaneous high-resolution (10 Hz) measurements of surface wind and saltation sand transport over a ~5 m tall barchan dune (median grain diameter = 0.35 mm) collected at White Sands Dune Field, New Mexico, USA. Studying aeolian transport in the field offered a natural experiment for understanding how the rate of aeolian saltation responds to turbulent changes in wind and frequent crossings of the threshold for particle motion. In agreement with past observations, our data indicate that: (1) saltation flux lags wind fluctuations by about 1 second, (2) the threshold for initiation of particle motion ("entrainment") exceeds the threshold for cessation ("distrainment") by about 20%, (3) concurrent instantaneous wind and sediment flux measurements are poorly correlated. Based on our data, we show how lagged transport and threshold hysteresis are related to inertia in the transport system arising from ballistic particle trajectories and non-instantaneous momentum transfers among grains and wind. We argue that this nonlinear and lagged response of saltation to turbulent wind fluctuations accounts for the poor correlation between wind and transport as well as the poor performance of existing saltation models.