Martian Surface Environment

The Martian surface is a 3.72 m/s2 gravity field [1] under a thin, almost pure carbon dioxide atmosphere at roughly 1 percent of Earth sea-level density, permanently loaded with suspended dust, with no global magnetic field and only a few tens of g/cm2 of atmospheric column shielding the ground.

Planetary and orbital parameters

Quantity	Value	Source
Gravitational parameter GM	42828 km³/s²	[1]
Mean equatorial radius	3396.2 km	[1]
Mean polar radius	3376.2 km	[1]
Oblateness term J2	0.00196045	[1]
Sidereal rotation period	88642.44 s (24 h 37 min 22 s)	[1]
Mean solar day (sol)	88775 s (24 h 39 min 35 s)	[1]
Surface gravity	3.72 m/s²	[1]
Orbital period	686.98 Earth days, 668.6 sols	[1]
Obliquity	25.19 deg	[1]
Orbital eccentricity	0.0934	[1]

Season is expressed as areocentric solar longitude L_s, with L_s = 0 at northern spring equinox, and the engineering atmosphere model is parameterized on L_s, latitude, longitude and local time rather than on calendar date [1]. Eccentricity of 0.0934 places southern summer near perihelion, so the perihelion half of the year carries both higher insolation and, as set out below, the dust storm season.

Atmosphere

Near-surface composition in Mars-GRAM is 94.8 percent CO2, 3.1 percent N2 and 1.9 percent Ar by mole fraction [1]. Measured surface pressure at the Viking lander sites ranged over 700 to 1100 Pa across the year [2]; at Gale crater, at an elevation of about 4.4 km below datum, Curiosity recorded 400 to 1200 Pa. The annual swing is driven by CO2 condensing onto and subliming from the seasonal polar caps, which is an exchange of total atmospheric mass rather than a local weather signal. The gradient with elevation is steep enough to be measured by a rover in motion: an elevation gain of 184 m over 1526 sols produced a pressure fall of about 15 Pa at Gale [2]. The atmospheric scale height at Gale is 11.1 km [10].

Diurnal pressure tides at equatorial sites have amplitudes of roughly 1 to 2 Pa near aphelion and about 5 Pa near perihelion, with the amplitude tracking dust opacity [2].

Design analyses are run against Mars-GRAM, which outputs temperature, pressure, density, wind components and constituent fractions [1]. It is a composite of three data sources: the NASA Ames MGCM from the surface to 80 km, the Michigan MTGCM from 80 km to 170 km (or 240 km for the TES-derived map years), and the Stewart thermospheric model above 170 km, whose base sits at the 1.26 nbar level near 125 km. Stored data are gridded at 5 km in altitude, 7.5 deg in latitude for MGCM and 5 deg for MTGCM, and 30 deg in L_s. The 2024 release upgraded the MOLA topography model to read the higher resolution Planetary Data System data, defaulting to 4 pixels per degree with 16, 32 and 64 pixels per degree also selectable [1]. Solar activity enters through F10.7, with 70 and 130 (10⁻²² W/cm² at 1 AU) as the low and high inputs and 200 available for the TES map years.

Mars-GRAM returns a perturbed density as rho’ = rho0 (1 + Rc’ PF PU), a correlated random walk about the mean profile [1]. The dispersion, not the mean, is what sizes landing ellipses and control authority. The model has been maintained since version 1.00 of 20 May 1988 [1].

Mars-GRAM sample output [1]: pressure against height from the surface to 600 km, spanning roughly fourteen decades and joining three separate data sources. Public domain (NASA).

Aerodynamic regime

Mean surface density is about 0.020 kg/m³, near 1 percent of Earth sea level [14]. For rotary-wing flight this places blade sections at chord Reynolds numbers of order 10⁴ against order 10⁶ for terrestrial rotorcraft, while the low speed of sound pushes tip Mach numbers high enough that the same section is simultaneously in a low-Reynolds and a compressible regime [13][14]. Wind tunnel measurement of Mars rotor airfoils is therefore done at combined conditions such as Re = 20,000 and M = 0.60, in a chamber held at 1 kPa and 288 K [13]. Peak sectional lift-to-drag ratio for an optimized Mars airfoil under those conditions is approximately 26, reached near 4.5 deg angle of attack, against values several times higher for a conventional airfoil at terrestrial Reynolds numbers.

Winds and the boundary layer

Quantity	Value	Source
Measured wind speed span, Curiosity sensor	0 to 150 m/s	[2]
Wind speed accuracy, Curiosity	±1 m/s below 20 m/s, ±4 m/s above	[2]
Expected wind speeds at Jezero	0.5 to 30 m/s	[8]
Convective vortex pressure drop	0.1 to 10 Pa, characteristic 0.6 Pa	[4]
Convective cell pressure fluctuation	0.1 Pa, period approx 100 s	[4]
Saltation shear stress threshold	0.01 ± 0.0015 N/m²	[11]

Wind speeds peak in daytime at every landed site, following convective mixing, and the seasonal maximum at the Viking Lander 2 site (47.9 N) falls in late northern winter near L_s = 345 deg [2].

InSight logged approximately 6000 convective vortex encounters in its first 400 sols, against a detection threshold of 0.3 Pa; the amplitude distribution follows a power law with exponent 3.4, so large-amplitude vortices are rare but not negligible [4]. Large-eddy simulations matched to that record used a model top at 10 km, roughly twice the expected daytime boundary layer depth.

The threshold for sustained saltation, calibrated from seasonal dune sand flux at Nili Patera, is a surface shear stress of 0.01 ± 0.0015 N/m² [11]. Sand flux there is about three times larger in southern summer than northern summer, following the perihelion insolation and wind cycle. That threshold sets both the surface mobility of loose deposits and the wind conditions under which deposited dust can be removed from a horizontal surface.

Dust

Column dust optical depth is the standard scalar for atmospheric dust loading. Gridded climatologies normalize it to the 610 Pa pressure level so that values from sites at different elevations are comparable [3]. Retrievals assume a particle effective radius of 1.06 µm with effective variance 0.3, and infrared absorption-only opacities are converted to visible extinction opacity by a factor of 2.6.

Condition	Column optical depth	Source
Aphelion season, L_s 0 to 180	tau < 1; interannual sigma 0.03 to 0.05	[2]
L_s 240 to 250 window	0.35 ± 0.08 (IR absorption, 610 Pa)	[3]
Global storm, global mean peak	approx 4 at L_s 205 to 210 (MY34)	[9]
Global storm, local values	5 to 10	[9]
Historic maxima	tau > 3 at L_s approx 300 (MY12) and L_s approx 280 (MY28)	[2]

Dust storm season runs L_s 185 to 300 [9]. Two windows of enhanced regional activity recur each year, at L_s 210 to 240 and L_s 320 to 340, with a third at L_s 150 to 170 observed over Gusev crater [2]. During the MY34 planet-encircling event dust was lofted to 80 km altitude [9].

Surface optical effects during a global storm are measurable directly. At Gale, horizontal line-of-sight extinction reached 1.15 km⁻¹ on sol 2088, against a pre-storm seasonal amplitude of 0.06 km⁻¹ and mean peak values of 0.11 km⁻¹ [10].

Dust loading also compresses the diurnal thermal cycle. During the 1977 global storm the diurnal air temperature amplitude at Viking Lander 1 fell from 42 K to 12 K near L_s = 275 deg [2]. Suspended dust absorbs and scatters in the visible, reducing direct-beam flux and raising the diffuse fraction, and warms the middle atmosphere, which changes the density profile at entry altitudes.

Deposition onto horizontal surfaces is cumulative and, absent wind at or above the saltation threshold, monotonic [11]. Removal by gusts and vortex passage is stochastic and is not a budgetable mechanism.

Thermal environment

Quantity	Value	Source
Air temperature measurement span, Viking	140 to 280 K	[2]
Air temperature span, Curiosity	150 to 300 K	[2]
Seasonal air temperature amplitude, 47.9 N	approx 60 K	[2]
Seasonal amplitude, 22.4 N	approx 30 K	[2]
Seasonal amplitude, 4.6 S	approx 20 K	[2]
Peak ground temperature, 68.2 N (Phoenix)	approx 259 K at L_s 107	[2]
Minimum ground temperature at 15 mm depth, Phoenix	approx 180 K at L_s 133	[2]
Annual amplitude of mean ground temperature, Gale	approx 20 K	[2]
Thermal inertia, Gale terrain	170 to 550 J m⁻² K⁻¹ s⁻¹ᐟ²	[2]

Thermal inertia at Yellowknife Bay was 300 to 420 J m⁻² K⁻¹ s⁻¹ᐟ², within that range [2]. Atmospheric heat capacity at 600 Pa is too small to buffer the surface overnight, and convective coupling is weak, so surface temperature tracks insolation and the diurnal range is set largely by the thermal inertia of the surface material.

The same quantity has been measured on Mars and in a chamber on the same terrain classes, which is rare for a soil property. Perseverance TIRS diurnal surface temperature over a whole sol gives apparent thermal inertia of 574 for bedrock, 456 for intermediate and 348 for sandy terrain, against 535, 421 and 375 for visually equivalent soils cycled in an 8 mbar CO2 Mars chamber at the University of Malaga, all in J m⁻² K⁻¹ s⁻¹ᐟ² and all through the same sinusoidal inversion [19]. Agreement is within 8 percent, but the terrain matching is by visual resemblance in Navcam imagery rather than by measured composition, so this is suggestive rather than a calibration.

Pressure is what makes the chamber run worth doing at Mars pressure. Across four soils of 0.7 to 50 mm granularity, the spread in estimated thermal inertia between the highest and lowest soil was 4.20 percent at 1000 mbar and 42.84 percent at 8 mbar [19]. Thermal discrimination between soils is therefore about ten times more sensitive at Mars pressure than at Earth pressure, because vertical heat transport through the grain bed is markedly poorer: the maximum surface to subsurface temperature difference measured on embedded thermocouples rose from 11.6 to 14.7 C for the 3 to 5 mm soil and from 14.4 to 24.4 C for the 0.7 to 1.0 mm soil when pressure was dropped. A sandbox validation of a thermal soil assessment sensor at ambient pressure understates what it will do on Mars.

Relative humidity at Gale peaks between 04:00 and 06:00 local mean solar time and falls below 5 percent between 10:00 and 18:00 [2]. Water vapor volume mixing ratio at Gale reaches approximately 70 ppm at night in late winter through early summer and falls below 5 ppm in late autumn and winter.

The MEDA suite on Perseverance defines the achievable measurement envelope for these quantities: air temperature over 150 to 300 K to ±1 K at 0.1 K resolution, pressure over 1 to 1400 Pa to ±20 Pa across the 400 to 1200 Pa band at 0.08 to 0.32 Pa resolution, and relative humidity over 0 to 100 percent across 190 to 270 K, degrading to ±20 percent RH at 190 to 200 K [8].

Regolith and surface materials

Surface materials span consolidated bedrock, cemented duricrust, indurated soils, and loose aeolian sand and drift.

Property	Value	Source
Cohesion, Martian sand	≤ 1 kPa	[6]
Cohesion, drift and crusty-to-cloddy and blocky indurated soil	up to 11 kPa	[6]
Cohesion, duricrust from mole-pit slope stability at 30 deg friction	5.8 kPa	[6]
Cohesion from mole penetration resistance, friction 30 to 40 deg	4 to 25 kPa	[6]
Friction angle, Martian sand	approx 30 deg, lower for drift	[6]
Bulk density, sand (pre-InSight estimates)	1000 to 1500 kg/m³	[6]
Bulk density, top 400 mm at InSight (TEM-A)	1211 (+149/−127) kg/m³	[6]
Crustal rock density	2800 to 3200 kg/m³, extremes to 3700	[6]
Implied porosity	approx 63 percent	[6]
Penetration resistance, duricrust	0.5 to 1.2 MPa	[7]
Penetration resistance below 300 mm	5.3 MPa	[7]
Shear modulus	4.32 ± 1.01 MPa	[7]
Bulk modulus	9.84 ± 6.54 MPa	[7]
Young’s modulus	11.30 ± 2.87 MPa	[7]
Poisson ratio	0.31 ± 0.15	[7]

Duricrust thickness is the parameter that has been most consistently underestimated. Prior to InSight, landed observations supported a cemented layer no more than a few centimeters thick over loose sand, and hardware was designed to punch through it [6]. At Homestead hollow the crust was at least 70 mm thick from visual inspection of the excavated pit, approximately 200 mm from the length of mole extraction during back-hammering, and possibly as deep as the 370 mm final tip depth, and the walls of that pit stood at 87 deg, which is itself a cohesion measurement [6]. Orbital thermal inertia at that site showed no duricrust signature, so thermal inertia is not by itself a reliable discriminator for cemented near-surface material.

Thermal conductivity of the regolith is low and layered: 14 mW/m/K near the surface, 34 mW/m/K within the crust, and 64 mW/m/K in the sand and gravel beneath [7]. Buried hardware is therefore poorly coupled to the deep ground and remains dominated by the diurnal wave in the upper layers.

Two hazard classes follow from the material range. Low-cohesion aeolian deposits fail in bearing and allow sinkage and slip. Embedded angular rock in indurated material or bedrock does not yield under load and acts as a rigid indenter, concentrating the full contact load on a small area.

Ionizing radiation

The atmosphere provides a column depth of approximately 22 to 23 g/cm² over Gale crater, which sits 4.431 km below datum [5][12]. That column is thin enough that only protons above roughly 160 to 200 MeV reach the surface as primaries [12]. Earth’s atmosphere presents approximately 1030 g/cm² by comparison. There is no global magnetic field, and crustal remanent magnetism is localized.

Quantity	Value	Source
Surface absorbed dose rate, BON2014 GCR model	0.172 mGy/day	[5]
Surface absorbed dose rate, DLR2013 GCR model	0.177 mGy/day	[5]
Surface dose equivalent rate, BON2014	0.539 mSv/day	[5]
Surface dose equivalent rate, DLR2013	0.560 mSv/day	[5]
Atmospheric column depth, baseline and range	23 g/cm², 20.4 to 24.1 g/cm²	[5]
Dose sensitivity to that column range	within 3 percent	[5]
Solar modulation parameter, baseline	572 MV	[5]

Dose responds far more strongly to solar modulation than to atmospheric thickness: a ±20 percent change in the modulation parameter changes surface dose by about 20 percent, while the full plausible range of atmospheric column depth moves it by under 3 percent [5]. Roughly half the surface neutron field is produced by albedo from the regolith rather than in the atmosphere, so the ground under a vehicle is part of its radiation environment. Neutrons below 1 MeV contribute under 10 percent of effective dose behind moderate shielding [5]. The MSL/RAD measurement campaign underlying these model comparisons is collected in [15], and surface dose accumulation derived from the same record is treated in [16].

Modeled neutron flux at the Martian surface against kinetic energy for four regolith compositions [5]. The spread below 10 MeV is regolith albedo; the field above 100 MeV is insensitive to surface composition. Public domain (NASA).

Entry and descent consequences

Surface gravity of 3.72 m/s² [1] acting with a mean surface density near 0.020 kg/m³ [14] sets the entry problem. The atmosphere is dense enough to require thermal protection and to make aerodynamic control usable, and thin enough that parachutes alone cannot decelerate a heavy vehicle to a survivable touchdown velocity. Powered descent is required above a mass threshold, and that threshold falls with landing site elevation because a higher site offers less column to decelerate through. Both the mean profile and its dispersion for these analyses come from Mars-GRAM [1], with dust optical depth entered either as a globally uniform value of tau = 0.3, 1.0 or 3.0, or as the dust optical depth observed by TES in retrieval year 1, which had no global dust storm, or year 2, which contained a major global storm peaking at L_s = 210 [1]. Design environments for program-level requirements are specified in [17], and the prior model release is [18].