Dune emergence in multidirectional wind regimes

Aeolian dunes result from the erosion and deposition of sand grains transported by the wind. In turn, the flow adapts to the shape of the sedimentary bed. This coupling between topography, hydrodynamics and transport controls the dune size, orientation and morphology, particularly during the early stage of their growth. In this thesis, we highlight the influence of certain environmental parameters, such as wind regime, sand availability, and dune field boundary conditions on dune emergence. For this, we use a theoretical approach, coupled with laboratory experiments, numerical simulations and field studies. First, we validate the description of dune emergence in areas of high sediment availability as a linear instability. These dunes form as surface waves, whose amplitude increases temporally or spatially, depending on whether they form in the middle of the sedimentary bed, or on its upstream edge. Their orientation, wavelength and spatial/temporal growth rate are then mainly controlled by the angular distribution of sand fluxes. On a non-erodible ground, dunes develop from various sources of sediments (other dunes, river and lakes deposits, etc.). Under specific wind regimes, they elongate to form linear ridges. We show that the spatial organization of these dunes is then controlled by the distribution of sedimentary sources, and not by a length scale intrinsic to the elongation mechanism. Thus, they form periodic dune fields at the downstream edge of sediment beds, or remain isolated structures in zones of localized deposition. In the latter case, they can reach a stable equilibrium state, whose morphology is controlled by the wind reorientation period. Finally, we carry out a large-scale analysis of the Namib Sand Sea based on the theoretical framework of sediment transport and dune morphodynamics, coupled to wind data from climate reanalyses. We then compare field observations with our predictions on sand fluxes and dunes (orientations, morphologies, growth rates). These results allow us to discuss the formation and evolution of sand seas on time scales covering several climatic cycles. More generally, this methodology also makes it possible to infer information on winds or sedimentary material when these are not directly measurable, as well as to gain insight into the evolution and resilience of dune systems subject to climate or anthropogenic changes.