Testing models of linear dune formation by provenance analysis with composite sediment fingerprints

Abstract

The formative mechanisms of linear (longitudinal) dunes and dunefields remain uncertain, and multiple hypotheses have been proposed. A central debate is the degree to which dunes act as along-dune sediment transport corridors, implying that dunes grow primarily by extension, or whether they are comprised of locally-derived sands moved from adjacent interdunes (the ‘wind-rift’ model). Sediment fingerprinting studies, with origins in fluvial science, have been shown to offer the possibility to trace the provenance of aeolian sands, and thus elucidate transport pathways.Two models (a Monte Carlo framework and a Generalized Likelihood Uncertainty Estimate framework) are used here to provide quantitative estimates of the sediment sources that have supplied a linear dune in the central Simpson Desert of central Australia. Four possible sources are identified that may have supplied the dune; two adjacent interdunes, one upwind low ridge of sand, and a merging upwind dune. Two sites near the dune's crest are used as the target and provided twenty surface samples for analysis. Following geochemical assay, stepwise discriminant function analysis identified optimum elemental sediment fingerprints for a variety of possible sediment pathway configurations.Results suggest that the sands of the dune are sourced predominantly from upwind dunes and sand sources, and that likely contributions from neighbouring dune swales are typically <20%. As such, wind-rift mechanisms of linear dune formation are not supported by these data. More complex sediment pathway configurations (i.e., other than a binary approach: interdune vs. along-dune), whilst confirming the initial findings, had reduced discriminatory power. Further separation of source pathways (e.g., identifying the relative roles of different upwind sources) was not possible with any confidence.The findings suggest recent sediment accretion of a linear dune dominated by along-dune sand flux, and thus support an extensional component for the development of such dunes. Whilst it is noted that at a point-by-point basis this might not exclude accretion by vertical growth, as some have observed, there is no clear support for a substantive contribution to the dune sands from adjacent interdunes. Moreover, the use of contemporary sediment fingerprinting methods to question hypotheses of aeolian geomorphology suggests that such methods have great potential for addressing other terrestrial geomorphological questions where identifying sediment pathways can provide vital insight.