Numerical Models - Dune Worlds: How Windblown Sand Shapes Planetary Landscapes - page 232

Fig. 19.3 CFD simulation of the surface wind flow over a coastal 'blowout' dune as modeled by Smyth et al. (2012). Notice the circulating flow
at the lee edges of ridges. Image courtesy Derek Jackson

Fig. 19.4 Flow streamlines over a barchan dune, computed with a lattice gas model. The streamline compression near the crest, and the flow
separation and recirculation bubble are clearly seen. Image courtesy of Clement Narteau

domains of cells can actually be simulated quite well (i.e.,
when there are thousands of little particles, the sum of the
little vectors can have a wide range of values, as if it were a
continuous variable). LB models can be parallelized effec-
tively (i.e., configured to run efficiently on multiprocessor
computer clusters) and, especially important for aeolian
studies, can handle multiphase flows and complex and
changing boundaries easily. LB models (one example, for
snow transport, is by Masselot and Chopard 1998) are likely
to see significant application to aeolian problems; recently
Narteau et al. (2009) explored the coupling of a cellular
automaton (see later) sand transport model to a lattice gas
flow model (e.g., Fig. 19.4 ) and found some fascinating

results, e.g., in the formation of star dunes (that arms grow
upwind, but only when an odd number of wind directions is
imposed—Zhang et al. 2012).

19.2 Modeling the Sand

At the level of individual sand grains, trajectories can be
modeled as we discuss in Chap. 4 , where the aerodynamic,
gravitational (and, if applicable, electrostatic) forces act on
a particle. Collisions between grains, the effect of particle
rotation (e.g., on lift via the Robins-Magnus effect), and the
variation in wind speed across the boundary layer, can all be