|ZHENG, XIAO JING|
|Van Pelt, Robert - Scott|
Submitted to: Journal of Geophysical Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/30/2006
Publication Date: 10/17/2006
Citation: Huang, N., Zheng, X., Zhou, Y., Van Pelt, R.S. 2006. The simulation of wind-blown sand movement and probability density function of lift-off velocities of sand grains. Journal of Geophysical Research. 111, D203201, doi: 10.1029/2005JD006559.
Interpretive Summary: Mechanistic computer models of wind erosion must have accurate input parameters and subroutines if we are to rely on their output values as decision making aids. Saltation is the primary mode of soil movement during an erosion event and yet little is known about the mechanisms of ejection and trajectories of the individual particles. The height that an individual particle will attain in a saltation cloud is dependent on its initial or lift-off velocity. Like most natural phenomena, the lift-off velocity is not constant and equal for all particles and the range of variation is sometimes characterized by a probability density function that is used as an input parameter. It was assumed that the heights and trajectories of particles in a saltation cloud were a function of their lift-off velocity and that random placement of lift-off would average the vertical gradient of sand flux at any point in the active saltation cloud. By measuring the vertical gradient of sand flux by use of a common sediment sampler, the vertical gradients could easily be measured. An accepted model of lift-off velocity was fit by use of a multi-objective optimization routine to the measured vertical gradients of sand flux obtained in wind tunnel experiments in order to obtain a representative probability density function for a specific sand material. It was found that this method is a robust and efficient technique for developing the probability density function of particle lift-off velocities.
Technical Abstract: Accurately describing the probability density function (PDF) of lift-off or initial velocities of wind-blown sand ejecting from a sand bed is fundamental to understanding the mechanisms of wind-blown sand movement. Our objective was to investigate the efficacy of developing the PDF of lift-off velocities based on wind tunnel measurements of sand-flux and wind speed profiles. After measurements of sand-flux and wind speed profiles in the saltation layer were conducted in a field wind tunnel, an optimization algorithm to numerically search for the PDF was developed by multi-objective optimization of a simulation model for wind blown sand movement proposed by Sorensen. Then an empirical formula of the PDF was inversely fit to the measured data by the least squares method. To show the efficacy and accuracy of the numerical results obtained using this approach, simulations of the sand-flux and wind speed profiles using selected PDFs of lift-off velocities were performed with the function we propose and functions previously proposed by other researchers. By comparing the suite of predicted PDFs with the experimental measurements, we found that although all the predictions agree reasonably well with the measured data, the method we propose for obtaining the PDF results in a more accurate PDF for predicting sand-flux and wind speed profile than functions previously proposed. These results validate the approach of obtaining the PDF of lift-off velocities based on wind tunnel measurements of sand flux profiles.