The prototypical and most commonly cited threshold of motion model used in sand transport predictions was derived by R.A. Bagnold in 1936. His wind-tunnel experiments resulted in two thresholds: (1) a fluid threshold, where particles are moved by the action of the wind alone, and (2) a dynamic threshold, where particles are moved by the fluid and saltating grains. However, Nickling (1988) advocated that the number of particles moving from the dynamic threshold greatly exceeded the number of particles moving by the fluid threshold alone. Since his publication, it is common practice to use only the impact threshold in sand transport models. Recently, Sherman et al. (in press) recalibrated six of the most commonly used transport models under ideal conditions (Bagnold 1937; Kawamura 1951; Zingg 1953; Owen 1964; Hsu 1971; Lettau and Lettau 1977). Recalibrating empirical constants, root-mean squared errors from model predictions decreased by as much as one order of magnitude, increasing the predictive power of models to determine observed transport rates. Yet, they found the largest deviations from recalibrated models occur at smaller transport rates (Sherman et al. in press), and these errors in the recalibrated transport models are attributed to errors in the threshold of motion. This calls for a re-examination of threshold of motion demarcation at small transport rates.

This study presents results of a field experiment designed to separate particle initiation from the fluid and dynamic thresholds at small transport rates. To do this, a bedload trap was designed capable of separating bedload from saltation transport.