Uncertainty propagation in aeolian processes: From threshold shear velocity to sand transport rate

The accurate estimation of aeolian saltation events is a fundamental requirement in the modelling of wind erosion, dust emission, dune movement and aeolian hazard prediction. A large number of semi-empirical sand transport rate models exist, with many relying on a single value for a shear velocity threshold above which saltation is initiated. However, measuring and modelling the sand transport rate suffers from the effects of a number of epistemic and aleatory uncertainties which make the identification of a single threshold value for shear velocity problematic. This paper focuses on the uncertainty propagation evident in calculations that use a threshold shear velocity to estimate sand transport rate. Probability density functions of threshold shear velocity are provided from the authors' previous studies. Grain diameter and shear velocity are considered as deterministically varying parameters. Several sand transport rate statistical metrics are estimated via the Monte Carlo approach adopting four different sand transport models. The sand transport rate estimation in probabilistic terms allows us to assess the amplification/reduction in the uncertainty and to provide a deeper insight into established transport rate models. We find that if the wind speed is close to the erosion threshold, every tested model amplifies the variability of the resulting estimated sand transport rate, especially in the case of coarse sand. If the wind speed is large, the adopted models present substantial differences in uncertainty. An interpretation of these differences is given by conditioning the sand transport rate models to the type of erosion threshold adopted, the fluid or impact threshold.