LAUSR

An accurate threshold velocity ut for sand movement is essential for predicting wind erosion events and calculating the magnitude of horizontal sand flux. We compared various approaches for determining ut using data observed during 4–31 July 2009 in the hinterland of Taklimakan Desert. We used the Gaussian time fraction equivalence method developed by Stout with an interval of 1 day to obtain ut values of 3.03–5.62 m/s (the approach of observations through field experiments). The Kurosak (ut50%) method yielded values of 3.71–5.74 m/s (the approach of statistical calculations). The Marticorena and Shao models gave ut values of 4.87–4.90 and 5.82–6.78 m/s, respectively (the approach of model parameterizations). To test the accuracy of these threshold velocities above, we estimated the total horizontal sand flux and the duration of sand saltation. Estimates of the total horizontal sand flux were 1311.9, 1166.4, 1279.9, and 661.6 kg/m for the Stout, Kurosak, Marticorena, and Shao methods, respectively, while the observed value was 732.9 kg/m. The correlation coefficient between observed and estimated values based on the Stout, Kurosak, Marticorena, and Shao methods was 0.75, 0.79, 0.77, and 0.83, respectively. The estimated duration of sand saltation using the ut been as Stout, Kurosak, Marticorena, and Shao was 8211, 6575, 7567, and 3463 min, of which 6208, 5646, 5986, and 3346 min were correct reports, respectively, and the observed value was 7663 min. We discuss the advantages and disadvantages of these methods combining those results above, which can be used to advise researchers to improve the threshold velocity used in future studies.