A test of granulometric control of desert dune geometry

Abstract

AbstractI. G. Wilson's hypothesis of coarse‐tail grain‐size control of dune spacing was derived from three dunefields in the Sahara and, although it was supplemented by measurements of spacing from other parts of the world, the grain‐size data come only from North Africa. In this paper the hypothesis is tested in the Australian dunefields. Australian dunes do not form separate categories on a P20/s (twentieth percentile/spacing) plot and, when placed on Wilson's diagram, the Australian data form a continuum between dunes and draas. Ripples maintain their identity, suggesting that the average saltation length of sand controls ripple wavelength while dunes and Wilson's draas are formed by secondary flow of some kind. In Australia the spread of data on the P20/s diagram indicates that grain‐size is not the prime control on s. Data from Australia and the Sahara indicate that direct linear relationships between s and h (dune spacing and height means respectively for blocks of dunes) occur but they have different slopes in different areas. These differences possibly reflect variations in vegetation and substrate as well as differences in wind regime. The separation of dunes from draas in the Sahara reflects the greater role of grain‐size in an area where coarse grains frequently occur in dune crests. The relative paucity of coarse grains in Australian dune crests may reflect the fine‐grained alluvium from which the dunes are derived. These differences may be ascribed to differing topography in Australia and the Sahara.