A lower-than-expected saltation threshold at Martian pressure and below

Abstract

Aeolian sediment transport is observed to occur on Mars as well as other extraterrestrial environments, generating ripples and dunes as on Earth. The search for terrestrial analogs of planetary bedforms, as well as environmental simulation experiments able to reproduce their formation in planetary conditions, are powerful ways to question our understanding of geomorphological processes toward unusual environmental conditions. Here, we perform sediment transport laboratory experiments in a closed-circuit wind tunnel placed in a vacuum chamber and operated at extremely low pressures to show that Martian conditions belong to a previously unexplored saltation regime. The threshold wind speed required to initiate saltation is only quantitatively predicted by state-of-the art models up to a density ratio between grain and air of [Formula: see text] but unexpectedly falls to much lower values for higher density ratios. In contrast, impact ripples, whose emergence is continuously observed on the granular bed over the whole pressure range investigated, display a characteristic wavelength and propagation velocity essentially independent of pressure. A comparison of these findings with existing models suggests that sediment transport at low Reynolds number but high grain-to-fluid density ratio may be dominated by collective effects associated with grain inertia in the granular collisional layer.

Keywords: Mars; impact ripples; saltation at low pressure; sediment transport threshold.

Fig. 1.

Impact ripple characteristics vs. pressure $P$. (A) Emergent wavelength $\lambda$ (black circles). (B) Photograph of the ripples in the tunnel at $u_{*}/u_t=1.1$. (C) Propagation speed $c$ (black squares) and maximum net bed erosion rate ${\phi }_m$ (red circles). (D) Space–time diagram of the ripple elevation profiles showing their coarsening dynamics when the wind speed is suddenly increased from $u_{*}/u_t=1.1$ to 1.5. Time goes from bottom to top. Wind is from left to right. These are data in ${\mathrm{C}\mathrm{O}}_2$ Martian conditions. Solid lines in A and C: adjustment of Eqs. 4 and 5 with factors calibrated on independent data (32). Statistical error bars corresponding to data dispersion on independent measurements are displayed.

Fig. 2.

Three-layer picture of aeolian sediment transport. The main, central layer is the feedback layer, where the saltating grains slow down the wind, and their feedback ensures a unit replacement capacity during rebounds (22). Above Bagnold’s focal point, defined as the top of that central layer, this feedback is negligible and the wind is unperturbed (“free wind” layer). At the interface with the static bed, momentum is transferred by collisions between the grains (39) (“collisional” layer). Note that, in this schematic, the vertical axis is not to scale: The altitude of Bagnold’s focal point is on the order of $50d$, or a centimeter (28, 40, 41).

Fig. 3.

Sediment transport threshold. (A) Averaged number of grains passing through the control window during a 30-s time interval vs. wind speed. The threshold $u_t$ is the cross-over between a regime of intermittent individual grains (arrows in B) and a fluctuating but steady transport (C). The precision on the number of grains is of the size of symbols; dispersion of data indicates the repeatability. (D) Threshold shear stress ${\rho }_f{u}_t^2$ vs. pressure $P$ (filled circles). Square: experiment in $C{O}_2$, in Martian conditions. All these data are for quartz grains of size $d=125\hspace{0.17em}\mathrm{\mu }$m. Solid line: prediction of the model (Materials and Methods and SI Appendix) adjusted on subaqueous and aeolian data, for different $d$ at ambient pressure. Green background: zone corresponding to the new low-pressure regime (large density ratio ${\rho }_p/{\rho }_f$). The 7% error bars reflect both measurement statistical errors on the threshold and uncertainties on the relation between shear velocity and rotation speed.

Fig. 4.

Transport threshold in the dimensionless plane Shields $\mathrm{\Theta }$ vs. Galileo $\mathcal{G}$ numbers. Gathering of subaqueous (violet diamonds) and aeolian (green square) data of the literature (33, 36, 50, 51) for sand grains with variable $d$ in ambient conditions and our aeolian data (circles) with variable $P$ and fixed grain size $d=125\hspace{0.17em}\mathrm{\mu }$m. The symbols’ color codes for the density ratio ${\rho }_p/{\rho }_f$: from 1 (violet) to $10^6$ (red), in log scale; green is in the range $10^3$ to $10^4$. Previously published data from the NASA wind tunnel using walnut shells (26) of size 212 $\mathrm{\mu }$m in ${\mathrm{C}\mathrm{O}}_2$ are displayed (triangles) too; see also SI Appendix, Fig. S15 for data corresponding to other grain diameters used in that paper. Solid lines: threshold predictions for subaqueous (purple), aeolian at ambient (green), and variable (orange) pressure conditions. Green background: new regime range.