Long-term dune geometry bounded by geology and climate

Dunes coarsen after forming at their incipient wavelength. Does this process of coarsening goe on indefinitely? What are the limits to the size of aeolian dunes? Andreotti et al. (2009, Nature) proposed a general explanation based on a mechanistic theory: aeolian dunes may grow until there is a resonance with the capping inversion on the atmospheric boundary layer (ABL),which essentially acts as a rigid lid. On the other hand, each dune field has a unique geologic and climatic trajectory that gives rise to the particular patterns we see today; many case-specific hypotheses exist for the size of dunes. In this study we test these end-member hypotheses exhaustively for many dune fields; is giant dune geometry set universally or uniquely? To that end, we find the geometry of all Earth dunes with wavelengths >100 m using ASTER, and pair this with data on: dune type, dune-field size and age, grain sizes, and contemporary winds from ERA-5. Importantly, we determine ABL heights from measurements by the space-borne lidar on board the CALIPSO satellite, rather than inferring this height from ground measurements as was done in previous work. This study is the most exhaustive and self-consistent dataset of aeolian dune conditions to date. We find that there is no correlation between dune size and observed ABL height. This does not mean, however, that there are no general trends in the data. The geometry of unimodal (barchan and transverse) and bimodal (linear dunes) flux-direction dunes all follow scaling laws seen in other studies. Interestingly, the class of multi-modal flux direction dunes (e.g., star dunes) follows a separate trend; the largest dunes on Earth belong to this group. These general trends are borne out in simulations of dune-field growth under a wide variety of boundary conditions, using the cellular automaton model ReSCAL. Our results indicate that the size of giant dunes is ultimately limited by the availability of sediment and the duration of climatic conditions favorable to aeolian sand transport. We observe a correlation between dune size and dune-field age that supports this hypothesis.