arX

iv:1

801.

0878

7v1

[as

tro-

ph.E

P] 2

6 Ja

n 20

18

Saltation under Martian Gravity and itsInfluence on the Global Dust DistributionGrzegorz Musiolika, Maximilian Krussa, Tunahan Demircia, Björn Schrinskia, JensTeisera, Frank Daerdenb, Michael D. Smithc, Lori Nearyb, Gerhard Wurma

aUniversity of Duisburg-Essen, Faculty of Physics, Lotharstr. 1-21, 47057 Duisburg, GermanybRoyal Belgian Institute for Space Aeronomy (BIRA-IASB), Ringlaan 3, B-1180 Brussels, BelgiumcNASA Goddard Space Flight Center, Greenbelt, MD 20771, United States

Abstract

Dust and sand motion are a common sight on Mars. Understanding the interaction ofatmosphere and Martian soil is fundamental to describe the planet’s weather, climateand surface morphology.We set up a wind tunnel to study the lift of a mixture between very fine sand and

dust in a Mars simulant soil. The experiments were carried out under Martian gravityin a parabolic flight. The reduced gravity was provided by a centrifuge under externalmicrogravity. The onset of saltation was measured for a fluid threshold shear velocity of0.82±0.04 m/s. This is considerably lower than found under Earth gravity.In addition to a reduction in weight, this low threshold can be attributed to gravity

dependent cohesive forces within the sand bed, which drop by 2/3 underMartian gravity.The new threshold for saltation leads to a simulation of the annual dust cycle with a MarsGCM that is in agreement with observations.

Keywords: Mars, Saltation, Microgravity Experiments, Cohesion, General Circulation Model

1. Introduction

Wind tunnel experiments simulat-ing dust lifting on the Martian

surface date back into the last century(Greeley et al., 1980). These studiesuse different low-density materials tosimulate the reduced gravity on Mars of0.38 g and provide the first thresholdsfor the onset of saltation. Compared to

Email address: gregor.musiolik@uni-due.de(Grzegorz Musiolik)

the available meteorological data whichallows an estimation of the Martianboundary layer winds (Hess et al., 1977;Schofield et al., 1997; Magalhães et al.,1999; Holstein-Rathlou et al., 2010)and to predictions from global circu-lation models (GCMs) (Forget et al.,1999; Haberle et al., 1999, 2003), thisthreshold should be exceeded only rarely(Jerolmack et al., 2006; Kok et al., 2012;Wang and Zheng, 2015; Newman et al.,2017). In contradiction to this, the

Preprint submitted to Icarus January 29, 2018

motion of dust and sand can be ob-served frequently and has a large impacton the Martian climate (Zurek et al.,1992; Smith, 2004; Heavens et al., 2011;Guzewich et al., 2017).

Strong efforts have been made inrecent years to detail the picture of soil-atmosphere interaction (White et al.,1987; Strausberg et al., 2005;Sullivan et al., 2005; Greeley et al., 2006;Merrison et al., 2007; Almeida et al.,2008; Merrison et al., 2008; Sullivan et al.,2008; Kok, 2010b,a; Bridges et al., 2012).Even though, it still remains questionableif dust storms can generally be initiatedby wind drag. For example, a lower shearvelocity would suffice to keep saltationactive but cannot explain the onset ofsaltation. Hence, also supporting effectsare studied. For example, insolation ofthe soil leads to thermal creep and a sub-surface overpressure, capable of reducingthe threshold wind velocity significantly(de Beule et al., 2014; Küpper and Wurm,2015). Also dust devils go along withpressure excursions which can supportgrain lifting (Balme and Hagermann,2006). In any case, numerical modelsoften use an artificially reduced thresh-old which is needed to initiate liftingevents to simulate saltation on Mars(Haberle et al., 2003; Kahre et al., 2006;Daerden et al., 2015).

However, aeolian transport experi-ments at Martian gravity and pressure, ase.g. byWhite et al. (1987), are rare. In thiswork, we investigate the influence of re-duced gravity on saltation and show thatthe threshold velocity for a sand bed pre-pared and subject to gas flow at Martiangravity and pressure is strongly reduced.

1.1. Experimental setup

The Martian environment is simulatedin a low pressure wind tunnel designed si-multaneously as a centrifuge to simulateMartian conditions (fig. 1). In detail, theexperiment consists of a vacuum chamberwhich is evacuated to a pressure of 6 mbarand a gas mixture of 95% CO2 and 5% air.It has a radius of 100 mm and can be ro-tated at more than 2 Hz. The wind tunnelis located in the center of the experimentchamber and has a cross section of 100mm × 100 mm. The wind flow is createdby a fan rotating with up to 11.000 rpm atan air flow rate of up to 570 m3/h. Thegas flows through the wind tunnel overthe sand bed and back again on the outerside of the wind tunnel. The total mass ofthe experiment is 161 kg. The Reynoldsnumber for this configuration inside thewind tunnel is on the order of Re ≈ 800.The set up is used in parabolic flights on

the ZERO-G Airbus operated by NOVES-PACE in Bordeaux (Pletser et al., 2016). Aflight consists of 31 parabolas with a du-ration of 22 s per parabola and a resid-ual acceleration on the scale of ±0.05 g(Pletser et al., 2016). The centrifugal forceon the surface of the dust bed is set to0.38 g, while additional experiments onground were carried out at 1 g. The par-ticle sample was ∼ 50 g of a mixture be-tween very fine sand and dust consist-ing of the JSC 1A Martian regolith sim-ulant, which was tempered at 600 K be-fore to remove volatiles and organics. Thissimulant is made out of altered volcanicash from a Hawaiian cinder cone and is arepresentative species for the reflectancespectrum, mineralogy, chemical composi-tion, density, porosity and magnetic prop-erties of the Martian soil (Allen et al.,1997). The size distribution of the used

2

centrifuge

slip ring

connector

wind tunnel

ventilator lifter mechanism

shutter mechanism

dust bed

rotation

Figure 1: A schematics of the experiment. The outer vacuum chamber has a diameter of 200 mm and isdesigned as a centrifuge. The wind tunnel is placed inside this centrifuge and has a cross section of 100mm × 100 mm.

sample is shown in fig. 2.

Figure 2: Grain size distribution of the used sam-ple. The fraction of larger grains dominates themass distribution and therefore the mechanicalproperties of the sample.

Before each parabola, the sample isclosed by a shutter mechanism to protectthe sample against uncontrolled accelera-tions. The experiment runs automatically.With the onset of the microgravity phase,

the chamber starts to rotate. The shutter isremoved once the set rotation frequency isestablished. Due to the momentum of theshutter, the sand sample is first lifted andthen settles back to the ground. This way,the surface of the sand sample is preparedat Martian gravity level before each mea-surement. The erosion is observed opti-cally with a camera installed perpendicu-lar to the wind flow at 457 frames per sec-ond and an exposure time of 200 µs, us-ing backlight illumination (s. fig. 3). Thisprovides a resolution sufficient to tracethe fraction of the larger particles fromfig. 2, but not sufficient to resolve the frac-tion of smaller particles.

2. Results

2.1. Data analysis

We use a Martian simulant JSC Mars 1Aas soil with a particle density of 1.9 g/cm3

3

(or the bulk density of 0.87 g/cm3 in-cluding 54 % porosity) (Allen et al., 1997)and a particle size distribution as de-picted in fig. 2. The shown size dis-tribution represents the volume densityof the particle sizes. We cannot excludethat the smaller dust might have an im-pact on the cohesion properties of thesample. Nonetheless, while the smallerparticles get sustained in the atmospheremore easily, saltation is probably domi-nated by the fraction of the larger par-ticles. The larger particles might haveeither a grain-like or aggregate struc-ture. In general, they fit in size toparticles in Martian dunes, which aregiven to 87µm (Claudin and Andreotti,2006; Kok et al., 2012). Though evenlarger particles, e.g. 40 − 400µm (HighDune Samples) or 50 − 400µm (NamibDune Sample) are discussed in the lit-erature as well (Ehlmann et al., 2017;Tirsch et al., 2012; Sullivan et al., 2008;Edgett and Christensen, 1991) the sampleallows an estimation for the minimumshear velocity needed to lift particles. Anexample for the observation of lifted sandparticles at 0.38 g is shown in fig. 3.

The roughness of the surface is con-sistent with the roughness map ofHébrard et al. (2012) derived from MOLAdata, in which the mean surface rough-ness on Mars is 4.435 mm and the mediansurface roughness is 11.05 mm, with36 % of the Martian surface having aroughness value higher than 5 mm. Thegas flow is just set high enough for liftingevents to occur and the fluid thresholdshear velocity u∗ is determined. Saltationtakes place as well as suspension. Onceinitiated, a lower wind velocity at theimpact threshold is needed to sustainthe particle flow, but this is not furtherinvestigated in this work.For Martian gravity of 0.38g, 51 trajec-

tories of lifted sand particles are analyzed,while 53 trajectories are analyzed for 1g.From these trajectories, the horizontalgas velocity and its dependency on theheight above the sand are calculated.The eroded sand particles couple tothe motion of the gas inside the windtunnel and are used to trace the gasvelocity close to the sand bed. For agiven height z, the trajectory of the sandaggregates along the (horizontal) x-axiscan be described by (Wurm et al., 2001)

x(t,z) =(

vg (z)− v0)

tC exp

(

−t

tC

)

+ vg(z)t + c. (1)

This equation is valid for spherical parti-cles with a constant coupling time but canbe used as an approximation for bumpyparticles as shown in fig. 4. The follow-ing fit parameters are obtained from fit-ting the trajectories of the sand particlesaccording to eq. (1): The initial veloc-ity v0 of the grain at a certain height z,the gas-aggregate coupling time tC , a con-

stant c and finally the gas velocity vg(z)for a given height z above the dust sam-ple. Furthermore, t is the time after thelifting event. Note, that the Coriolis forceis negligible for the lifting process of theparticles (as they are at rest) as well as forthe grain motion at a constant height z atwhich the particles are tracked.

For the 0.38g trajectories, the values for

4

5mm

Figure 3: Snapshot of particles lifted at 0.38 g close to threshold wind velocity. The wind is flowing fromleft to right. The shown surface roughness is typical.

Figure 4: Sample trajectories in wind direction for1 g (top) and 0.38 g (bottom). The motion of theparticles was fitted according to eq. (1). The fitsare overplotted in black. The particles are acceler-ated by the gas motion until they finally couple toit.

the gas velocities are binned in 0.5 mmsteps. For each bin, the values for the me-dian gas velocity are calculated with 7-8

individual values for the gas velocity fromthe fitted trajectories according to eq. (1).For the 1g trajectories, the bin size is setto 0.6 mm. The binned data is given infig. 5. Both profiles indicate a linear cor-

Figure 5: Gas velocity profile over height at thethreshold of particle lifting for 1 g (green) and 0.38g (blue). The data are binned, including 53 indi-vidual values for 1 g and 51 values for 0.38 g. Theslopes dvg (z)/dz resulting from the linear fits are

680 s−1 for 1 g and 453 s−1 for 0.38 g. The thresh-old u∗ was calculated using eq. (2).

relation between the horizontal gas veloc-ity and the height above the sand surface.A linear profile close to the ground is alsoin agreement with former experiments inwind tunnels (Merrison et al., 2008). Theerror bars show the mean error calculatedfrom the fits of all trajectories for 0.38 g

5

and 1 g, respectively.For the microgravity measurements,

additional uncertainties resulting fromthe residual acceleration and vibrationsinside the aircraft have to be considered.In order to avoid errors due to the resid-ual acceleration of ±0.05 g, the gas veloc-ities deduced from the trajectories are av-eraged as described. The vibrations in-side the cabin have a higher frequency ofω ≈10 Hz, which is an estimation fromthe acceleration data. If the amplitudeof the vibrations was A ≈ 100µm, whichequals the grain size and hence is a max-imum estimation, the error in accelera-tion would be on the order of gv = Aω2

≈

0.01m/s2. Compared to the Martian grav-itation of 0.38g, this is a relative error of2.7%, which we consider as negligible.

2.2. Threshold shear velocity and cohesionreduction for Mars

Interaction of a turbulent wind flowwith a surface can be characterized bythe shear velocity u∗ =

√

τ/ρ with theshear stress τ and the fluid density ρ(Schlichting and Gersten, 2016). Thisquantity can be interpreted as the windvelocity acting directly at the soil. Theshear stress can be expressed by Newton’slaw of viscosity to τ = ηdvg(z)/dz with thedynamic viscosity η and the flow heightprofile dvg(z)/dz depending on the gas ve-locity vg(z) and the height z. Thus, u∗ canalso be expressed as

u∗ =

√

η

ρ

dvg(z)

dz. (2)

The gas velocity vg(z) is logarithmic inz within a turbulent sublayer and lin-ear in z within a viscous sublayer closeto the ground as measured in this work.

Considering η ≈ 15 µPa·s and ρ ≈ 0.01kg/m3 (CO2 at 6 mbar and 300 K) aswell as dvg(z)/dz from fig. 5, the thresh-old shear velocity can be derived directlyfrom eq.(2) and yields 0.82 ± 0.04 m/s for0.38 g and 1.01± 0.04 m/s for 1 g.The threshold shear velocity at 0.38 g is

lower than values determined in prior ex-periments on ground (Greeley et al., 1980;Merrison et al., 2008) which are generallysomewhat larger with ∼ 1.5−2 m/s. How-ever, u∗ was measured in a different gravi-tational environment in this work and de-pends also on the grain species. Thus,it cannot be compared directly to theseother works. This might also be an indica-tion that prior experiments perhaps over-estimated this value for the Martian soil.Using the models from Shao and Lu

(2000) and Merrison et al. (2008) with thethreshold shear velocities for 0.38 g and1 g, the particle density of 1.9 g/cm3

and a mean particle diameter of approx-imately 85 µm we get a surface energy ofγSL ≈ 1.1 · 10−7 J/m2. This is an unrea-sonably low value as the used JSC speciesmostly consists of SiO2, Al2O3, Fe2O3 andCaO which all exceed values of 10−2 J/m2

for the surface energy (Heim et al., 1999;Miller, 2011).In consequence of the low value for

γSL we consider a lower cohesive forceat lower gravity influencing the ratio ofthe determined threshold shear velocities.The cohesion force at the threshold can beestimated from the force balance

CLπ

2ρr2u∗2 =

∑

j

FC,j +Mg. (3)

The lifting force is given on the left side(Küpper and Wurm, 2015). CL is thelifting coefficient which depends on theboundary conditions of the wind tunnel

6

and the shape of the particles, r is theaverage radius of the particles and ρ isthe fluid density. Counteracting are thegravitational forceM ·g with the particle’smassM and the gravitational accelerationg and the sum over all cohesive contactsFC,j of a grain. Grains of ∼100 µm are usu-ally easiest to move as cohesive forces andgravity are similar (Greeley et al., 1980).Thus, none of the addends can be ne-glected. We assume that all individualcontacts are sharing the same contact areaand can be described by the JKR model(Johnson et al., 1971; Tomas, 2006) whichgives

∑

j FC,j ≈ NFJ = N 32πγr with the

amount of contacts per grain N and thesurface energy γ . This approach finallyresults in the threshold

u∗ =

√

43CL

(

9N2

γ

ρd+ρp

ρdg

)

, (4)

with the particle density ρp and diam-eter d. Except the dependency in N ,this expression is similar to the equationprovided by Shao and Lu (2000). If thecontact number N depends on the grav-itational acceleration, u∗ might be lowerfor reduced gravity. With two values forthe u∗, this dependency N (g) can be esti-mated.The ratio between both threshold shear

velocities at different gravitational envi-ronments can be written as

u∗21u∗22

=N1FJ +Mg1N2FJ +Mg2

≡FN +Mg1χFN +Mg2

, (5)

with the sum of all contact forces FN ≡N1FJ and the contact number ratio χ ≡N2/N1. We can derive χ from eq. (5) to

χ =u∗22u∗21

(

1+Mg1FN

)

−Mg2FN

. (6)

Applying the values for the fluid thresh-old shear velocity in this work with the av-erage number of contacts N1 in 0.38g andN2 in 1g gives

χ ≈32∀ FN ≫ 10−8N. (7)

This result shows, that the average num-ber of contacts and thus also the total con-tact forces are only 2/3 as large in 0.38 gas in 1 g, if FN exceeds 10−8 N by an or-der of magnitude. If we consider N = 1,γ ≈ 0.01 J/m2 which is a typical value forsilicate spheres (Heim et al., 1999) and r =10−5 m (as minimum estimation) the ad-ditional condition is easily fulfilled withF1 =

32πγr ≈ 5·10−7 N. Experimental work

on contact forces confirms this likewise(Heim et al., 1999). This is the first timethat it is considered that cohesion is notconstant in soils of different planets asgravity does compress the soil differently.A reduction in contact number in the lowgravity environment of Mars can explaina reduction in the threshold wind velocitynecessary to lift particles. Absolute val-ues of the fluid threshold shear velocityderived from our experiment under 0.38g indicate that saltation and suspensionare possible under the conditions given onMars and in agreement to particles beingobserved in motion.

3. Simulation with the Global Circula-tion Model (GCM)

3.1. Mars GCM

A General Circulation Model (GCM)for the atmosphere of Mars is ap-plied to calculate surface shearvelocities (Daerden et al., 2015;Neary and Daerden, 2018). It is oper-ated on a grid with a horizontal resolution

7

of 4◦x4◦ and with 103 vertical levelsreaching from the surface to ∼150 km.The model calculates heating and coolingof atmospheric CO2 and dust and ice par-ticles by solar and IR radiation and solvesthe primitive equations of atmosphericdynamics. The geophysical boundaryconditions are taken from observationsand include a detailed surface roughnesslength map. Physical parameterizationsin the model include an interactive CO2condensation and surface pressure cycle,a thermal soil model, turbulent transportin the atmospheric surface layer andconvective transport inside the planetaryboundary layer. The effects of the extremeMartian topography are considered witha low level blocking scheme. The shearvelocity is derived from the computedwind field in the second lowest verticalmodel level (at height ∼15 m), followingthe expressions derived from similaritytheory (Jacobson, 2005). In the model,dust is lifted by saltation whenever theshear stress exceeds a critical value that iscalculated from the threshold shear veloc-ity given by eq. (2), where dvg /dz is takenfrom fig. 5, and the dynamic viscosity forCO2 is calculated from Sutherlands for-mula after Crane (1988) using the GCMpredicted surface temperature. Dust islifted in a lognormal distribution withmean radius 1.5 µm, which contains 3size bins: 0.1, 1.5 and 10 µm. The idea ofsaltation is that the larger sand particlesare lifted, if the shear velocity is exceededand fall back to the surface as they are toolarge to stay aloft. From the collision withthe surface, smaller (dust) particles arelifted, which are able to go in suspension.In the GCMmodel this process is shortcutby lifting µm-size particles directly whenthe threshold shear velocity is exceeded.

Dust is lifted in the GCM following theKahre-Murphy-Haberle (KMH) method(Kahre et al., 2006), in which the dustmass flux from the surface is calculated as

F =(

2.3 · 10−3)

ατ2(

τ − τ∗

τ∗

)

(8)

with τ the actual and τ∗ the thresh-old surface wind stress. α is a pro-portionality factor that has to be set foran optimal match with observations. Itdoes not control where and when dust islifted, but only how much dust is actu-ally lifted. Dust particles are sedimentedin the model using the size-dependentStokes settling velocity with Cunning-ham slip-flow correction (Jacobson, 2005).Dust is radiatively active in the GCM, by a2-stream approximation applying the lat-est optical properties (Wolff et al., 2006,2009). It undergoes all the transport pro-cesses in the model such as diffusive mix-ing and advection. Dust is the main ther-modynamic agent in the middle and loweratmosphere of Mars and drives the globalcirculation under differential solar heat-ing in combination with local processessuch as saltation. In this way, saltation issimulated in the GCM as fully interactive.One assumption that is made is that of alimitless surface reservoir of dust.Until now, in GCMs a threshold of typ-

ically 0.0225 Pa is used (Haberle et al.,2003; Kahre et al., 2006; Daerden et al.,2015; Neary and Daerden, 2018), a valuecorresponding to a ∼40% reduction of thecritical shear stress derived from the orig-inal lab measurement for static conditions(Greeley et al., 1980), to have any dust lift-ing at all. It is found in our simulationsthat the new threshold shear stress can betypically 5 times lower than the one de-rived from previous laboratory data, and

8

so that the new threshold for saltation onMars does no longer require GCMs to ap-ply a strongly reduced value to simulatedust lifting.

3.2. Simulations with the new threshold val-ues

Using the new values for the thresh-old shear velocity with the GCM allowsa prediction of locations where dust andsand movement is preferred on the Mar-tian surface. Fig. 6a shows the zonally av-eraged dust optical depth measurementsfrom the Thermal Emission Spectrome-ter (TES) instrument on the NASA MarsGlobal Surveyor (MGS) orbiter (Smith,2004) for 3 consecutive Martian years.The seasonal behavior with a less activedust season during northern spring andsummer and a highly active dust (or duststorm) season during southern spring andsummer is clearly visible.The dust data from TES shown in

fig. 6a are extinction optical depths ob-tained from the measured absorption op-tical depths by multiplying by 1.3 (Smith,2004). The optical depths are scaledto visible wavelengths from the originalmeasurement at 1075 cm1 (9.3 µm) bymultiplying by 1.8 (Clancy et al., 2003).The TES measurements are mostly takenaround 2 p.m. local time (Smith, 2004).The values are scaled to a surface pressureof 610 Pa and are averaged over all longi-tudes and over bins of 2◦ in latitude and2◦ in solar longitude. Solar longitude (LS )is the angle from the sun between Marsand its orbital vernal equinox and oftenused to indicate Mars time of year.6b shows the result of a simulation of

the dust cycle with the GCM applyingthe old threshold for dust lifting that was

strongly reduced from previous experi-mental work (Haberle et al., 2003). Thevalue of the efficiency factor α was set to0.015. 6c shows the result of a simula-tion of the dust cycle with the GCM ap-plying the new threshold for dust liftingderived from our experiment, without anyfurther reduction. The efficiency factor αwas set to 0.0026. The model results pre-sented in the figure are obtained as fol-lows from the GCM output. The dust op-tical depth is calculated in the model at0.67 µm. Themodel output is sampled ev-ery 30 minutes, and averaged over all lon-gitudes with local time between 1 and 3p.m. The resulting dataset is binned likethe TES data over 2◦ in LS . A mask wasapplied to the resulting time series to re-move the times and latitudes for whichthere is no TES data available.

The figure shows that the GCM is ableto predict the times and latitudes wheredust lifting occurs and provides a dust cy-cle that is qualitatively comparable to thedata. The interannual variability of thepeaks in the dust storm season is a topicof ongoing research (Mulholland et al.,2013; Shirley and Mischna, 2017) and be-yond the scope of the present work. Thefigure also shows that applying the newthreshold obtained from our measure-ments in the model simulation is equiv-alent to applying the threshold that wasstrongly reduced from values obtainedin previous measurements, i.e. the newthreshold does not lead to unforeseencomplications, and allows for a dust cyclesimulation using an experimentally foundthreshold for saltation.

9

Figure 6: (a) Latitude versus time distribution of the dust optical depth measured on Mars by the TESinstrument on MGS for 3 consecutive Mars years (1999-2004). The horizontal coordinate is the solar longi-tude (LS ). The data was scaled to visible wavelengths and to a surface pressure value of 610 Pa, and binnedover 2◦ latitude and 2◦ LS . (b) Simulation of the same quantity in the GCM by applying a threshold on sur-face shear velocity that was strongly reduced from previous experimental work. (c) Simulation of the samequantity in the GCM by applying the threshold on surface shear velocity derived from the experiment inthis work, without reduction. Model output is averaged in the same way as the data and removed whereno data is available.

χ =u∗2u∗1

(

1+Mg1FN

)

−Mg2FN≈ 1.22 ∀ FN ≫ 10−8N. (9)

4. Discussion and Conclusion

We measured the fluid threshold shearvelocity for a Martian simulant JSC 1Awith a dominating grain size on the orderof ∼ 100 µm. For Martian gravity of 0.38

g, this value yields 0.82 ± 0.04m/s andincreases to 1.01 ± 0.04 m/s for 1g. Weattribute the difference between boththreshold shear velocities to a reducednumber of contacts between particles.

10

The low absolute value for the thresholdshear velocities also shows the importanceof the chosen sand and the conditionsfor the sample preparation. Prior exper-iments perhaps overestimated the fluidthreshold shear velocity on Mars, as thesoil sample was always prepared underEarth gravity and therefore with the cor-responding number of contacts betweenthe particles. Our new findings bringnumerical simulations of dust transporton Mars by general circulation models inagreement with observations, without theneed for reduction of the threshold.

With this work, we perform the firstwind tunnel studies on saltation directlyunder Martian gravitational and atmo-spherical conditions. Nonetheless, the ex-periments are performed on a small timescale of ∼20 s. It is an important ques-tion whether the results would be appli-cable on longer timescales which cannotbe answered by this work. In future, fur-ther quantitative studies comprising ex-periments for several g-levels might con-firm the tendencies and give a clearer pic-ture of the relation between gravity andcohesion.

Acknowledgements The experimentswere carried out on the 65th ESAparabolic flight campaign as part of theFly Your Thesis! 2016 programme. Thework was supported by ESA Education,the DFG under grant numberWU 321/12-1 and DLR Space Administration withfunds provided by the Federal Ministryfor Economic Affairs and Energy (BMWi)under grant number DLR 50 WM 1760.We thank Jan Raack and an anonymousreviewer for a constructive review.

References

Allen, C. C., Morris, R. V., Lindstrom, D. J., Lind-strom, M. M., and Lockwood, J. P. (1997). JSCMars-1 - Martian Regolith Simulant. In Lunarand Planetary Science Conference, volume 28 ofLunar and Planetary Inst. Technical Report.

Almeida, M. P., Parteli, E. J., Andrade, J. S., andHerrmann, H. J. (2008). Giant Saltation onMars. Proceedings of the National Academy of Sci-ences, 105(17):6222–6226.

Balme, M. and Hagermann, A. (2006). ParticleLifting at the Soil-Air Interface by AtmosphericPressure Excursions in Dust Devils. GeophysicalResearch Letters, 33(19).

Bridges, N. T., Bourke, M. C., Geissler, P. E., Banks,M. E., Colon, C., Diniega, S., Golombek, M. P.,Hansen, C. J., Mattson, S., McEwen, A. S., et al.(2012). Planet-Wide Sand Motion on Mars. Ge-ology, 40(1):31–34.

Clancy, R. T., Wolff, M. J., and Christensen, P. R.(2003). Mars aerosol studies with the mgs tesemission phase function observations: Opticaldepths, particle sizes, and ice cloud types ver-sus latitude and solar longitude. Journal of Geo-physical Research: Planets, 108(E9).

Claudin, P. and Andreotti, B. (2006). A ScalingLaw for Aeolian Dunes on Mars, Venus, Earth,and for Subaqueous Ripples. Earth and Plane-tary Science Letters, 252(1):30–44.

Crane (1988). Flow of Fluids Through Valves, Fit-tings, and Pipe. Number 410. Crane.

Daerden, F., Whiteway, J., Neary, L., Komguem, L.,Lemmon, M., Heavens, N., Cantor, B., Hébrard,E., and Smith, M. (2015). A Solar Escalatoron Mars: Self-Lifting of Dust Layers by Ra-diative Heating. Geophysical Research Letters,42(18):7319–7326.

de Beule, C., Wurm, G., Kelling, T., Küpper, M.,Jankowski, T., and Teiser, J. (2014). TheMartianSoil as a Planetary Gas Pump. Nature Physics,10:17–20.

Edgett, K. S. and Christensen, P. R. (1991). TheParticle Size of Martian Aeolian Dunes. Journalof Geophysical Research: Planets, 96(E5):22765–22776.

Ehlmann, B. L., Edgett, K. S., Sutter, B., Achilles,C. N., Litvak, M. L., Lapotre, M. G. A., Sulli-van, R., Fraeman, A. A., Arvidson, R. E., Blake,D. F., Bridges, N. T., Conrad, P. G., Cousin, A.,Downs, R. T., Gabriel, T. S. J., Gellert, R., Hamil-ton, V. E., Hardgrove, C., Johnson, J. R., Kuhn,

11

S., Mahaffy, P. R., Maurice, S., McHenry, M.,Meslin, P.-Y., Ming, D. W., Minitti, M. E., Mo-rookian, J. M., Morris, R. V., O’Connell-Cooper,C. D., Pinet, P. C., Rowland, S. K., SchrÃűder,S., Siebach, K. L., Stein, N. T., Thompson, L. M.,Vaniman, D. T., Vasavada, A. R., Wellington,D. F., Wiens, R. C., and Yen, A. S. (2017). Chem-istry, mineralogy, and grain properties at namiband high dunes, bagnold dune field, gale crater,mars: A synthesis of curiosity rover observa-tions. Journal of Geophysical Research: Planets,pages n/a–n/a. 2017JE005267.

Forget, F., Hourdin, F., Fournier, R., Hourdin, C.,Talagrand, O., Collins, M., Lewis, S. R., Read,P. L., and Huot, J.-P. (1999). Improved GeneralCirculation Models of the Martian Atmospherefrom the Surface to Above 80 km. Journal ofGeophysical Research: Planets, 104(E10):24155–24175.

Greeley, R., Leach, R., White, B., Iversen, J., andPollack, J. B. (1980). Threshold Windspeeds forSand on Mars - Wind Tunnel Simulations. Geo-physical Research Letters, 7:121–124.

Greeley, R., Whelley, P. L., Arvidson, R. E., Cabrol,N. A., Foley, D. J., Franklin, B. J., Geissler,P. G., Golombek, M. P., Kuzmin, R. O., Landis,G. A., et al. (2006). Active Dust Devils in Gu-sev Crater, Mars: Observations from the MarsExploration Rover Spirit. Journal of GeophysicalResearch: Planets, 111(E12).

Guzewich, S. D., Toigo, A. D., andWang, H. (2017).An Investigation of Dust Storms Observed withthe Mars Color Imager. Icarus, 289:199–213.

Haberle, R. M., Joshi, M. M., Murphy, J. R.,Barnes, J. R., Schofield, J. T., Wilson, G., Lopez-Valverde, M., Hollingsworth, J. L., Bridger, A. F.,and Schaeffer, J. (1999). General CirculationModel Simulations of the Mars Pathfinder At-mospheric Structure Investigation/MeteorologyData.

Haberle, R. M., Murphy, J. R., and Schaeffer,J. (2003). Orbital Change Experiments witha Mars General Circulation Model. Icarus,161(1):66–89.

Heavens, N., Richardson, M., Kleinböhl, A., Kass,D., McCleese, D., Abdou, W., Benson, J.,Schofield, J., Shirley, J., and Wolkenberg, P.(2011). Vertical Distribution of Dust in the Mar-tian Atmosphere during Northern Spring andSummer: High-Altitude Tropical Dust Maxi-mum at Northern Summer Solstice. Journal of

Geophysical Research: Planets, 116(E1).Hébrard, E., Listowski, C., Coll, P., Marticorena,

B., Bergametti, G., Määttänen, A., Montmessin,F., and Forget, F. (2012). An aerodynamicroughness length map derived from extendedmartian rock abundance data. Journal of Geo-physical Research: Planets, 117(E4).

Heim, L.-O., Blum, J., Preuss, M., and Butt, H.-J.(1999). Adhesion and Friction Forces BetweenSpherical Micrometer-Sized Particles. PhysicalReview Letters, 83:3328–3331.

Hess, S. L., Henry, R. M., Leovy, C. B., Ryan, J. A.,and Tillman, J. E. (1977). Meteorological Re-sults from the Surface of Mars: Viking 1 and2. Journal of Geophysical Research, 82(28):4559–4574.

Holstein-Rathlou, C., Gunnlaugsson, H. P., Merri-son, J. P., Bean, K., Cantor, B., Davis, J., Davy, R.,Drake, N., Ellehoj, M., Goetz, W., et al. (2010).Winds at the Phoenix Landing Site. Journal ofGeophysical Research: Planets, 115(E5).

Jacobson, M. Z. (2005). Fundamentals of Atmo-spheric Modeling. Cambridge university press.

Jerolmack, D. J., Mohrig, D., Grotzinger, J. P., Fike,D. A., and Watters, W. A. (2006). Spatial GrainSize Sorting in Eolian Ripples and Estimationof Wind Conditions on Planetary Surfaces: Ap-plication to Meridiani Planum, Mars. Journalof Geophysical Research: Planets, 111(E12):n/a–n/a. E12S02.

Johnson, K., Kendall, K., and Roberts, A. (1971).Surface Energy and the Contact of ElasticSolids. In Proceedings of the Royal Society of Lon-don A: Mathematical, Physical and EngineeringSciences, volume 324, pages 301–313. The RoyalSociety.

Kahre, M. A., Murphy, J. R., and Haberle, R. M.(2006). Modeling the Martian Dust Cycle andSurface Dust Reservoirs with the NASA AmesGeneral Circulation Model. Journal of Geophysi-cal Research: Planets, 111(E6):n/a–n/a. E06008.

Kok, J. F. (2010a). An Improved ParameterizationofWind-Blown Sand Flux onMars that Includesthe Effect of Hysteresis. Geophysical ResearchLetters, 37(12).

Kok, J. F. (2010b). Difference in the WindSpeeds Required for Initiation Versus Continu-ation of Sand Transport on Mars: Implicationsfor Dunes and Dust Storms. Physical Review Let-ters, 104(7):074502.

Kok, J. F., Parteli, E. J., Michaels, T. I., and Karam,

12

D. B. (2012). The Physics of Wind-BlownSand and Dust. Reports on Progress in Physics,75(10):106901.

Küpper, M. and Wurm, G. (2015). Thermal CreepAssisted Dust Lifting onMars: Wind Tunnel Ex-periments for the Entrainment Threshold Veloc-ity. Journal of Geophysical Research: Planets, 120.

Magalhães, J. A., Schofield, J. T., and Seiff, A.(1999). Results of the Mars Pathfinder Atmo-spheric Structure Investigation. Journal of Geo-physical Research: Planets, 104(E4):8943–8955.

Merrison, J., Gunnlaugsson, H., NÃÿrnberg, P.,Jensen, A., and Rasmussen, K. (2007). De-termination of the Wind Induced DetachmentThreshold for Granular Material on Mars usingWind Tunnel Simulations . Icarus, 191(2):568 –580.

Merrison, J. P., Bechtold, H., Gunnlaugsson, H.,Jensen, A., Kinch, K., Nornberg, P., and Ras-mussen, K. (2008). An Environmental Simula-tion Wind Tunnel for Studying Aeolian Trans-port on Mars. Planetary and Space Science,56(3):426–437.

Miller, C. M. (2011). Adhesion and the surface en-ergy components of natural minerals and aggre-gates. PhD thesis, Texas A & M University.

Mulholland, D. P., Read, P. L., and Lewis, S. R.(2013). Simulating the Interannual Variabilityof Major Dust Storms on Mars using VariableLifting Thresholds. Icarus, 223(1):344–358.

Neary, L. and Daerden, F. (2018). The gem-marsgeneral circulation model for mars: Descriptionand evaluation. Icarus, 300(Supplement C):458– 476.

Newman, C. E., Gómez-Elvira, J., Marin, M.,Navarro, S., Torres, J., Richardson, M. I., Bat-talio, J. M., Guzewich, S. D., Sullivan, R., de laTorre, M., et al. (2017). Winds Measured bythe Rover Environmental Monitoring Station(REMS) During the Mars Science Laboratory(MSL) Rover’s Bagnold Dunes Campaign andComparison with Numerical Modeling usingMarsWRF. Icarus, 291:203–231.

Pletser, V., Rouquette, S., Friedrich, U., Clervoy, J.-F., Gharib, T., Gai, F., and Mora, C. (2016). TheFirst European Parabolic Flight Campaign withthe Airbus A310 ZERO-G. Microgravity Scienceand Technology, 28(6):587–601.

Schlichting, H. and Gersten, K. (2016). Boundary-Layer Theory. Springer Berlin Heidelberg.

Schofield, J., Barnes, J. R., Crisp, D., Haberle,

R. M., Larsen, S., Magalhaes, J., Murphy,J. R., Seiff, A., and Wilson, G. (1997). TheMars Pathfinder Atmospheric Structure Inves-tigation/Meteorology (ASI/MET) Experiment.Science, 278(5344):1752–1758.

Shao, Y. and Lu, H. (2000). A simple expres-sion for wind erosion threshold friction veloc-ity. Journal of Geophysical Research: Atmospheres,105(D17):22437–22443.

Shirley, J. H. and Mischna, M. A. (2017). Orbit-spin Coupling and the Interannual Variabil-ity of Global-Scale Dust Storm Occurrence onMars. Planetary and Space Science, 139:37 – 50.

Smith, M. D. (2004). Interannual Variability inTES Atmospheric Observations of Mars during1999–2003. Icarus, 167(1):148–165.

Strausberg, M. J., Wang, H., Richardson, M. I.,Ewald, S. P., and Toigo, A. D. (2005). Observa-tions of the Initiation and Evolution of the 2001Mars Global Dust Storm. Journal of GeophysicalResearch: Planets, 110(E2).

Sullivan, R., Arvidson, R., Bell, J., Gellert, R.,Golombek, M., Greeley, R., Herkenhoff, K.,Johnson, J., Thompson, S., Whelley, P., et al.(2008). Wind-Driven Particle Mobility on Mars:Insights from Mars Exploration Rover Observa-tions at âĂIJEl DoradoâĂİ and Surroundings atGusev Crater. Journal of Geophysical Research:Planets, 113(E6).

Sullivan, R., Banfield, D., Bell, J., Calvin, W., Fike,D., Golombek, M., Greeley, R., Grotzinger, J.,Herkenhoff, K., Jerolmack, D., et al. (2005).Aeolian Processes at the Mars ExplorationRover Meridiani Planum Landing Site. Nature,436(7047):58–61.

Tirsch, D., Craddock, R. A., Platz, T., Maturilli, A.,Helbert, J., and Jaumann, R. (2012). Spectraland petrologic analyses of basaltic sands in ka’udesert (hawaii)–implications for the dark duneson mars. Earth Surface Processes and Landforms,37(4):434–448.

Tomas, J. (2006). Mechanics of Particle Adhesion.Particles on Surfaces, 8:183–229.

Wang, P. and Zheng, X. (2015). Unsteady Saltationon Mars. Icarus, 260:161–166.

White, B., Greeley, R., Leach, R., and Iversen, J.(1987). Saltation Threshold Experiments Con-ducted under Reduced Gravity Conditions. In25th AIAA Aerospace Sciences Meeting, page 621.

Wolff, M., Smith, M., Clancy, R., Arvidson, R.,Kahre, M., Seelos, F., Murchie, S., and Savijärvi,

13

H. (2009). Wavelength Dependence of DustAerosol Single Scattering Albedo as Observedby the Compact Reconnaissance Imaging Spec-trometer. Journal of Geophysical Research: Plan-ets, 114(E2).

Wolff, M. J., Smith, M. D., Clancy, R., Spanovich,N., Whitney, B., Lemmon, M. T., Bandfield, J.,Banfield, D., Ghosh, A., Landis, G., et al. (2006).Constraints on Dust Aerosols from the MarsExploration Rovers using MGS Overflights andMini-TES. Journal of Geophysical Research: Plan-ets, 111(E12).

Wurm, G., Blum, J., and Colwell, J. E. (2001).NOTE: A New Mechanism Relevant to the For-mation of Planetesimals in the Solar Nebula.Icarus, 151:318–321.

Zurek, R. W., Barnes, J. R., Haberle, R. M., Pollack,J. B., Tillman, J. E., and Leovy, C. B. (1992). Dy-namics of the Atmosphere of Mars. Mars, pages835–933.

14