Relatively little attention has been given to dust emission from southern Africa, despite the identification of major source areas across the arid southwest of the subcontinent (Prospero et al., 2002;Vickery

For these reasons, dust constitutes an important component of the earth system, and so it is essential that it is accurately represented in state-of-the-art climate models (Boucher et al., 2013). In most coarse-resolution models, dust emission is parameterized, and is given as a function of the cube of surface wind speed u above a specified threshold u t , so that;where q is the vertical sand flux (Knippertz & Todd, 2012;Kok et al., 2012). Realistic simulation of the emission process is therefore dependent on capturing the surface wind speed distribution in dust source areas, however evaluation studies show that models tend to miss the tail of low-frequency, high wind speed events that drive uplift (Evan, 2018;Roberts et al., 2017), an inevitable consequence of their inability to resolve some key atmospheric processes (Allen & Washington, 2014; Garcia-Carreras et al., 2013).