Evolving dunes under flow reversals: from an initial heap toward an inverted dune

TL;DR

This study investigates how 3D barchan dunes and 2D dunes evolve and reverse flow conditions, revealing key dynamics and time-scales involved in dune inversion processes through experiments and simulations.

Contribution

It compares 2D and 3D dune morphologies and characterizes the reversal dynamics, providing insights into the time-scales and static grain behavior during flow reversals.

Findings

Reversal time is twice the initial dune development time.

Grains on the lee side climb back onto the dune during reversal.

Numerical 2D slice models predict realistic reversal dynamics.

Abstract

Sand dunes are ubiquitous in nature, and are found in abundance on Earth and other planetary environments. One of the most common types are crescent-shaped dunes known as barchans, whose mid-line could be assumed to behave as 2D dunes. In this work, we (i) compare the morphology of the mid-line of 3D barchans with 2D dunes; and (ii) track the evolution of 3D barchans and 2D dunes while reversing flow conditions. We performed experiments on 2D dunes in a 2D flume and Euler-Lagrange simulations of 3D bedforms. In all reversal experiments and simulations, the initial condition start with a conical heap deforming into a steady-state dune, which is then perturbed by reversing the flow, resulting in an inverted dune. We show that during the reversal the grains on the lee side immediately climb back onto the dune while its internal part and toe remain static, forming a new lee face of varying angle on the previous stoss slope. We show that (i) the characteristic time for the development of 2D dunes scales with that for 3D barchans, (ii) that the time for dune reversal is twice the time necessary to develop an initial triangular or conical heap to steady-state, and (iii) that a considerable part of grains remain static during the entire process. Our findings reveal the dynamics for dune reversal, and highlight that numerical computations of barchans based on 2D slices, which are more feasible in geophysical scales, predict realistic outcomes for the relevant time-scales.