Valles Marineris cloud trails



[1] Distinctive cloud trails are identified in Mars Reconnaissance Orbiter Mars Color Imager (MARCI) images over specific locations associated with Valles Marineris and Noctis Labyrinthus and at perihelion solar longitudes (LS = 230°–260°). High-contrast surface shadows are well defined, as cast from their eastern margins, supporting altitude and optical depth determinations. These relatively high altitude clouds (40–50 km) exhibit narrow latitudinal widths (25–75 km) in comparison to extended longitudinal dimensions (400–1000 km). MARCI multispectral imaging of cloud surface shadows in five wavelength channels (260, 320, 437, 546, and 653 nm) yields the wavelength dependence of cloud extinction optical depth, revealing a range of small cloud particle sizes (reff = 0.2–0.5 μm) and moderate cloud optical depths (0.03–0.10 visible and 0.1–0.2 ultraviolet). Local time and temporal sampling characteristics of MARCI cloud images indicate that these clouds develop very rapidly in afternoon hours (1300–1500 LT), reach their full longitudinal extents within <2 h time scales, and often reoccur on successive afternoons. Mars Global Surveyor Mars Orbital Camera imaging in previous Mars years indicates these clouds are annually repeating. These observed characteristics suggest a cloud formation mechanism that is specific to ∼50 km horizontal and vertical scales, transports water vapor and dust upward from lower levels, exists during the afternoon, and is likely associated with the mesoscale atmospheric circulations induced by the near-equatorial canyons of Mars. Cloud particles formed in such updrafts would then be rapidly transported westward in the strong retrograde zonal circulation of the subsolar middle atmosphere in this season.

1. Introduction

[2] The existence of water ice clouds in the Mars atmosphere has long been recognized, based upon extensive observations ranging from historical ground-based telescopic reports [Martin et al., 1992] to spacecraft imaging [Leovy et al., 1973; French et al., 1981] and spectroscopic [Curran et al., 1973; Smith et al., 2001] studies. Such clouds in the lower Mars atmosphere (below 40 km altitudes) have been shown to reflect meteorological conditions [e.g., Kahn, 1984], as well as affect the global transport of atmospheric water [Clancy et al., 1996a; Montmessin et al., 2004, 2007a]. Our knowledge of discrete high-altitude clouds is currently limited to the specific locations (equatorial, over two longitude corridors) and LS ranges (10°–40°, 120°–160°) associated with the Mars Equatorial Mesospheric (MEM) clouds [Clancy et al., 2004, 2007]. Clouds occurring over the same LS ranges and locations have been spectroscopically identified from OMEGA (Observatoire pour la Mineralogie, l'Eau, les Glaces et l'Activite, on Mars Express) spectral imaging as CO2 rather than water ice in composition [Montmessin et al., 2007a]. High-altitude clouds over these same locations and times are also observed in THEMIS (Thermal Emission Imaging System, on Mars Odyssey) multicolor imaging [McConnochie et al., 2006]. All of these clouds, whether identical in origin or not, occur within very cold atmospheric conditions that may be consistent with CO2 saturation temperatures.

[3] We report a distinctly different type of Mars high-altitude cloud in this study, based upon recent Mars Reconnaissance Orbiter (MRO) imaging observations taken from the Mars Color Imager (MARCI) [Malin et al., 2001]. As these clouds appear related to the Valles Marineris canyon system and they are very linear in form, we refer to them as Valles Marineris cloud trails (abbreviated herein as VMCT). Based upon the significant difference (≥50 K) between CO2 saturation temperatures and typical perihelic atmospheric temperatures associated with their altitude and season of occurrence, they are almost certainly water ice in composition. Their localized occurrence suggests a potential connection to mesoscale circulation patterns associated with the Valles Marineris canyon, which has been considered as a special dynamical province [Rafkin and Michaels, 2003]. MARCI multiwavelength (260–650 nm) imaging observations are presented for multiple VMCT observed over May–June of 2007, toward characterization of their spatial and physical properties. Although it is not currently possible to demonstrate an unambiguous origin for these clouds, we consider their observed behaviors in the context of recent simulations of mesoscale circulation associated with Valles Marineris.

2. MARCI Imaging Observations

[4] The MARCI wide-angle imaging experiment employs separate visible and ultraviolet optics, both illuminating a single interline transfer 1024×1024 CCD detector [Malin et al., 2001]. MARCI obtains wide-angle (WA) (180° cross-track field of view) images in a push frame mode from the near-polar, solar synchronous (1500 LT at nadir) orbit of MRO. Continuous MARCI WA imaging provides full dayside latitudinal coverage in overlapping longitudinal strips that enable construction of daily global, multispectral image maps of Mars. At nadir, MARCI visible (437, 546, 604, 653, and 718 nm) and ultraviolet (260, 320 nm) filter images are characterized by 1 and 8 km surface resolutions, respectively. These spatial resolutions degrade away from the nadir toward limb viewing, associated with projection foreshortening. MARCI radiometric calibrations were performed with detailed preflight laboratory characterizations of the flight instrument, post launch observations of the Earth and Moon, and coordinated MARCI and Hubble Space telescope comparisons in July of 2007 [Bell et al., 2009]. The daily global imaging capability of MARCI is designed to support both climate and surface science objectives [Malin et al., 2008]. In particular, MARCI extends the long-term polar cap and dust storm monitoring capabilities formerly provided by Mars Global Surveyor (MGS) Mars Orbiter Camera (MOC) wide-angle imaging [e.g., Cantor et al., 2001; Thomas et al., 2005; James et al., 2007; Cantor, 2007]. MARCI ultraviolet and multispectral visible wavelength coverages add unique ozone and surface mineralogical measurement capabilities in support of Mars climate monitoring objectives [Malin et al., 2008].

[5] Here, we employ MARCI images obtained in 260, 320, 420, 550, 600, and 650 nm filters during the early southern summer season (Mars solar longitude, LS = 230°–260°) of 2007. Daily global image maps for the MARCI 260 and 320 nm observations provide a quick-look tool for identifying VMCT occurrences as they are an automated data product and because Mars clouds appear as high-contrast features against a low UV reflectance surface (surface silicate albedo of 0.01–0.015 at 260–320 nm wavelengths [e.g., Clancy et al., 1996b]). The MARCI visible wavelength images provide a factor-of-eight gain in spatial resolution (∼1 km in nadir viewing versus 8 km in the UV channels), and a key measure of VMCT color ratios for particle size determinations. In Table 1, we list the observed occurrences of VMCT as identified in MARCI global daily image maps, including dates and locations. VMCT appear distinctly confined in locale and season, but reoccur on a variety of time scales within these bounds. Locations refer to the sharp eastern margins (and probable source regions) of the clouds, which extend westward for many hundreds of kilometers.

Table 1. MARCI Mesospheric Cloud Trail Observation Summary
  • a

    Date in 2007.

  • b

    West longitude for cloud eastern margin.

  • c

    In kilometers, primarily east to west, with slight north-to-south cant.

  • d

    Kilometers from local surface elevation (5 km uncertainty) for cloud eastern margin.

  • e

    Normal optical depth at λ = 650 and 320 nm (20% uncertainty) for cloud eastern margin.

  • f

    Estimated cloud particle size, reff (−20%/+50% uncertainty), in microns for cloud eastern margin.

  • g

    MARCI image identifier (mission phase-orbit number).

  • h

    Numbers are as follows: 1, eastern margin not observed; 2, western margin not observed; 3, poorly defined surface shadow; 4, strong longitudinal variation.

6 May20°S43°W>600233°510.06, 0.120.4P07-0036312
7 May21°S≤50°W>400234°---P07-0036451
9 May20°S≤39°W>700235°---P07-0036711
26 May20°S38°W400245°---P07-0038953
27 May12°S47°W>550246°400.08, 0.230.2P07-0039082
27 May20°S43°W>700246°410.12, 0.250.3P07-0039082, 4
28 May12°S47°W>550246°---P07-0039213
28 May20°S49°W>700246°---P07-0039213, 4
31 May12°S37°W–41°W>800249°---P07-0039481
31 May22°S38°W–41°W>800249°---P07-0039481
31 May13°S≤100°W>550249°---P07-0039502, 4
1 June19°S≤35°W>600249°---P08-0039611
1 June22°S53°W>600249°---P08-0039612, 3, 4
2 June20°S42°W>1000250°510.13, 0.180.5P08-0039742, 4
2 June13°S103°W>550250°390.14, 0.240.4P08-0039762, 4
2 June11°S106°W>500250°---P08-0039762, 3
6 June22°S54°W450252°---P08-0040273, 4

3. Cloud Trail Behaviors

[6] Figures 1 and 2 present two of the more striking VMCT observed by MARCI, as viewed in the 320 nm ultraviolet channel and a color construction from visible (437, 546, and 604 nm) channels, respectively. Several key characteristics of VMCT clouds are apparent in these images; their long, narrow horizontal aspect ratios, their association with two narrow latitudinal ranges near Valles Marineris, and their substantial heights as indicated by extended surface shadows at their eastern margins. These characteristics indicate specific cloud behaviors that may be consistent with a localized, deep vertical updraft initiating and maintaining water ice cloud particle formation at the eastern margins of each cloud. Prevailing easterly (east-to-west) zonal winds at 40–50 km altitudes may subsequently transport the entrained ice clouds over ∼400 km distances to the west (and slightly south) during the several hour time scale of afternoon updraft forcing. For example, simulated zonal winds presented in the Mars Climate Database [Forget et al., 1999] predict 60–100 m/s westward zonal winds during the LS period, local time, and location of VMCT. The observed structure of a VMCT along its length may be due to quasi-cyclic time variance of the updraft strength (and thus of the quantity and/or nature of the cloud particles created) and/or horizontally propagating gravity waves initiated at the updraft.

Figure 1.

A portion of a daily global MARCI ultraviolet (320 nm) image map, obtained on 27 May of 2007 at LS = 246°, presents two high-altitude (40–41 km) cloud trails (VMCT) straddling Eos Chasma of Valles Marineris. The resolution is 8 km per pixel for the presented 0°W–90°W, 15°N–35°S region, which includes portions of three MARCI image strips obtained on consecutive MRO orbits. Surface shadows are apparent at the leading eastern margins of the extended VMCT at 12°S, 47°W and 20°S, 43°W. The western trailing margins of these clouds are clipped due to MRO spacecraft roll operations conducted during the central orbit imaging.

Figure 2.

The same VMCT presented in Figure 1 in a color MARCI image map constructed from the MARCI 437, 546, and 604 nm filter channels (as for all color images presented in Figures 3, 4, and 6). MARCI visible channels provide higher spatial resolution (∼1 km) and surface contrast to reveal Coprates, Ganges, Capri, and Eos chasmas and VMCT details.

3.1. Spatial Characters

[7] The key spatial relationships of VMCT are their linear form and restriction to specific locations in proximity with Valles Marineris or Noctis Labyrinthus. As Table 1 indicates, three general regions of VMCT occurrence are located, offset ∼200 km south of the Ganges Chasma (12°S, 37°W–47°W) and Eos Chasma (20°S, 35°W–54°W) southern boundaries, and ∼300 km south of the southern boundary for Noctis Labyrinthus (12°S, 100°W–106°W). The restricted latitudes of VMCT appearance, 11°S–13°S and 19°S–22°S, are particularly distinctive. The most frequently observed location of cloud occurrence (20°S, 43°W (Figures 13)) presented VMCT on three widely separated days (6 May, 27 May, and 2 June of 2007 (Table 1)). Figure 3 indicates the 2 June appearance of a parallel set of VMCT south of Noctis Labyrinthus, as well as southward of Eos Chasma.

Figure 3.

MARCI color mapped images of VMCT observed over Noctis Labyrinthus (12°S, 105°W) and southward of Eos Chasma (20°S, 42°W) on 2 June 2007 (MY 28) at LS = 250°.

[8] The spatial scales of VMCT are defined by narrow north-south dimensions (25–75 km at their eastern margins) versus extended east-west dimensions (400–1000 km). The eastern margins of VMCT appear relatively sharp, casting distinct surface shadows that are biased eastward by the afternoon local time of MARCI imaging (∼1500 LT). The western margins of VMCT are less distinct as their brightness and horizontal contrast diminish with distance from their presumed eastern boundary origins. VMCT trend slightly north to south from their eastern to western extents, by roughly 1 degree in latitude over 13 degrees in longitude.

3.2. Temporal Characters

[9] VMCT are observed to occur over a relatively short seasonal range, from LS = 233° to LS = 252° in 2007 MARCI imaging. Individual VMCT were identified in MGS MOC imaging during previous Mars Years 24, 25, 26, and 27 (where 2007 period of MARCI observations corresponds to MY 28), with very similar locations as VMCT presented in 2007 MARCI observations (Figure 4). A relatively late seasonal occurrence was identified by MOC imaging in MY 25, for LS = 263° [Cantor et al., 2008]. We also note the VMCT appear in recent (April 2009) MARCI imaging in MY 29. The LS range of VMCT occurrence corresponds to Mars perihelion, as the subsolar latitude transits the latitudinal range of VMCT occurrence and solar forcing is at a maximum.

Figure 4.

MGS MOC two-color image of a VMCT observed on 25 June 2005 (MY 27) at LS = 237°. MOC image identifiers are S07-02707 (red filter) and S07-02708 (blue filter).

[10] The local time of MARCI imaging varies across its wide-angle field of view by ∼1 h, and is centered on 1500 local time. By comparison, MOC wide-angle imaging was centered on 1400 local time. Consequently, all VMCT observations correspond to midafternoon local times, and cannot explicitly address the diurnal variation of these clouds. One particular MARCI observation does, however, indicate a time scale for the horizontal development of these clouds. Figure 5 presents a set of overlapping MARCI image strips obtained on 9 May 2007. The appearance of a fully formed VMCT at 20°S within the image strip centered at ∼55°W follows ∼2 h after the earlier image strip centered at ∼25°W was obtained. The cloud extends to at least the 39°W eastern margin of the 55°W centered strip, which is cut off at 33°W by the overlain 25°W centered image strip in the global map projection of Figure 5. Consequently, the VMCT present in the 55°W centered image strip was not present at the time the earlier (25°W centered) image strip was obtained, such that the full longitudinal development of the VMCT occurred in ≤2 h. VMCT are also observed to dissipate between successive MARCI image strips. Another temporal aspect of VMCT is their reoccurrence on consecutive days; for example, on 26, 27, and 28 May in 2007 (Table 1 and Figure 6). Given the significance of solar heating implied by the perihelion and southern summer season of occurrence, it is likely (but not demonstrated) that these clouds form and dissipate during afternoon local times.

Figure 5.

A fully extended VMCT (9 May 2007 at 20°S, ≤39°W (Table 1)) appears in the second MARCI 320 nm image strip but not in the previous (third) image strip, taken ∼2 h earlier. This indicates that the eastern margin/origin of this VMCT did not exist at that time and that the westward extension of the VMCT (over 700 km) occurred in less than 2 h.

Figure 6.

Three MARCI color mapped images of VMCT, obtained from consecutive days ((left) 26 May, (middle) 27 May, and (right) 28 May; see Table 1). Dark image portions on 27 and 28 May correspond to atmospheric limb/space views. VMCT often appear over several days, although they are likely to dissipate diurnally and reform during afternoon hours.

3.3. Cloud Shadow Heights

[11] As apparent in Figures 1 and 2, VMCT cast distinct shadows at their eastern margins. Due to the nearly tangential orientation of the Sun direction to the longitudinal axis of these clouds, shadows are not reliably measured along the extended length of VMCT. Nevertheless, the projected offset between the eastern margins of the cloud and cloud shadow supports altitude determinations for VMCT at their presumed source regions. Calculating cloud shadow heights from MARCI wide-angle images requires correction for the differential projections of surface (shadow) and elevated (cloud) features, associated with a strongly varying emission angle across the MARCI field of view. Cloud height determinations for five VMCT are provided in Table 1, ranging from 39 to 51 km above the local surface. The individual uncertainties in these cloud heights are large (∼5 km) associated with the difficulty in assigning cloud and shadow boundaries that are diffuse, due in part to the limited spatial resolution of MARCI imaging (∼1 km/pixel in the visible channels). Nevertheless, the average altitude of 45 km among the determinations is a reasonably accurate assignment for the typical cloud formation altitude at the eastern VMCT margins. It is possible that VMCT cloud levels decrease away from their eastern margins, as cloud particles gravitationally settle over the course of their westward transport in the strong easterly circulation at these altitudes. However, the small particle sizes estimated for VMCT (see below) do not suggest substantial fall distances over several hour time scales.

3.4. Cloud Optical Depth and Particle Sizes

[12] VMCT extinction optical depths are derived from their wavelength-dependent transmission, as calculated from surface brightness traces across MARCI ultraviolet (260 and 320 nm) and visible (437, 546, and 653 nm) cloud shadow images (Figure 7). This follows the technique of Montmessin et al. [2007b], who determined mesospheric CO2 cloud optical depths and particle sizes from OMEGA spectral imaging. In applying the Montmessin et al. [2007b] diffuse scattering corrections, we have adopted a background dust extinction optical depth of 0.6. We obtain a ratio of 0.3 for diffuse versus direct flux correction, employing DISORT [Stamnes et al., 1988] multiple scattering radiative transfer calculations. This scattering correction increases derived VMCT extinction optical depths by approximately 50%. Among the MARCI observations of VMCT, eastern margin cloud shadows were imaged with sufficient fidelity to determine the wavelength-dependent normal optical depth for five separate clouds, as presented by colored symbols in Figure 8. Cloud heights and locations are indicated in the top right corner of Figure 8, with additional information supplied in Table 1. Retrieved cloud optical depths vary by a factor of 2–3 for a given wavelength, with typical 653 nm and 320 nm extinction optical depths of 0.1 and 0.2, respectively.

Figure 7.

The observed decrease in imaged 420 nm signal (DN) across the eastern margin of a VMCT (2 June 2007 at 20°S, 42°W (Table 1)) surface shadow is displayed against the unregistered (i.e., not projected in latitude-longitude) MARCI image of this cloud. The surface trace of the plotted DN cross section is indicated in the image.

Figure 8.

The wavelength-dependent extinction optical depths, as calculated for 6 VMCT from surface shadows at their eastern margins, are compared to the wavelength dependence of extinction optical depth presented by water ice spheres of three particle sizes (reff = 0.1, 0.25, and 1.0 μm). The cloud heights and locations are indicated in the top right corner and in Table 1.

[13] The average factor-of-two increase in cloud optical depth between the red and ultraviolet imaging channels of MARCI is indicative of small cloud particles. The dashed, solid, and dotted lines presented in Figure 8 indicate the predicted wavelength dependence of extinction optical depth for water ice spheres of cross-section-weighted radii (reff) 1.0, 0.25, and 0.1 μm (assuming a modified gamma size distribution with 0.1 μm effective variance). In general, the wavelength of peak optical depth is roughly equivalent to the cloud particle size, implying quite small ice aerosol particle sizes (reff ∼ 0.4 μm) for the VMCT presented in Figure 8. Such ice cloud particle sizes are substantially smaller than previous particle size determinations of reff = 1–4 μm for Mars water ice clouds [Curran et al., 1973; Clancy et al., 2003, 2007; Wolff and Clancy, 2003; Wolff et al., 2006], with the exception of extremely small (reff = 0.01–0.10 μm) ice haze particles above 20 km altitudes reported from Mars Express SPICAM observations [Rannou et al., 2006].

4. Cloud Processes

4.1. Microphysical Constraints

[14] A consideration of the observed 40–50 km altitudes of VMCT in the context of water saturation conditions is best accomplished through comparisons to Mars Global Climate Model (MGCM) predictions for clouds over the LS = 230°–260° period. For this comparison, we employ the online (hence, universally accessible) Mars Climate Database [e.g., Forget et al., 1999], which provides a global suite of simulated parameters (including clouds) versus LS and atmospheric conditions. MCD simulated clouds over the LS = 240°–270°, afternoon local time of VMCT observations (and for the model MY 24 low dust conditions) are present over 30–50 km altitudes at 10°S, trending toward 40–60 km altitudes at 20°S. The water content of simulated MCD clouds also compares favorably with that determined for VMCT (∼10−5 g/cm3). However, significant MCD clouds are present at all longitudes between latitudes of 20°S and 20°N. By comparison, MARCI images show minimal cloud opacities apart from the VMCT and sporadic Arsia Mons clouds.

[15] The approximate water content of VMCT, as constrained from the cloud optical depth and particle size determinations, may be compared to estimates of atmospheric water content for the derived altitudes of VMCT. Of interest is the degree to which vertical transport of water to the 40–50 km altitude of VMCT formation is necessary to maintain the observed cloud mass. The water content of the small VMCT particles is roughly ten times larger than for 2 μm sized ice clouds with similar extinction opacities, due to their reduced extinction efficiencies. A visible (653 nm) cloud optical depth of 0.1 for such particles implies an equivalent water content of 0.1 precipitable microns (prμm). By comparison, a well mixed (constant mixing ratio with altitude) water vapor column of 10 prμm leads to 0.3 prμm of water above 40 km altitude, such that the water content of VMCT does not exceed in situ availability. Nevertheless, some vertical transport of additional water vapor from portions of the atmospheric column below the altitude of VMCT formation would be associated with the 50 km scale (horizontal) updrafts that appear to be the most likely mechanism for triggering VMCT genesis (see section 4.2).

[16] Fairly high cloud particle densities are implied for VMCT. For an assumed cloud vertical extent of 5 km, the observed cloud optical depth and particle sizes require particle number densities of 60 per cm3. Such high particle number densities follow from the observed small cloud particle sizes, which are atypical of Mars water ice clouds so far considered and suggest distinctive formation conditions. Available observations do not distinguish these conditions uniquely, but may include processes such as rapid adiabatic cooling, super saturation conditions, or high densities of small dust particles for rapid nucleation of small ice aerosols.

4.2. Dynamic Origins

[17] The short time scale (<2 h) implied for the westward extension of VMCT ice particles over >700 km distances requires substantial (≥100 m/s) easterly zonal wind velocities transporting the particles westward at 40–50 km altitudes, 10°S–20°S latitudes, and 30°W–110°W longitudes. In addition, the north-to-south trend of the VMCT implies southward meridional winds of order 10 m/s over the same regions. Such zonal and meridional velocities are reasonably consistent with MCD predictions for the locations of VMCT during the perihelion period, with the caveat that the implied easterly zonal circulation is a lower limit and may well exceed MCD predicted values. If cloud particles are primarily formed within updraft regions at their eastern margins, their evolution during westward entrainment with the easterly zonal circulation may be limited to dissipation by turbulent diffusion, vertical fall, and relatively slow water ice sublimation. Gravitational settling rates for the small VMCT particles, however, are quite slow with fall heights of the order of a few hundred meters over a 2 h period [e.g., Kahre et al., 2008]. Unfortunately, detailed and comprehensive measurements of the Martian atmosphere, at the times and locations that VMCT occur, do not exist. The available VMCT observations would benefit from a more detailed atmospheric state context, mesoscale atmospheric model simulations may begin to satisfy these needs. The following discussion, regarding a potential dynamical genesis of VMCT clouds, remains speculative due to the lack of supporting observations as well as definitive mesoscale simulations.

[18] Previous mesoscale atmospheric modeling work by Rafkin et al. [2002] and Michaels et al. [2006] suggests that the energetic mesoscale circulations associated with the large volcanic mountains of Mars may transport substantial amounts of dust and water from lower levels to the altitudes where the VMCT exist. Relatively detailed microphysical modeling of the afternoon lee cloud of Olympus Mons [Michaels et al., 2006] provides a cloud formation scenario that is perhaps similar to that of the VMCT. The daily formation of the lee cloud (located at approximately 20–30 km altitude) involves rapid cloud particle formation and growth directly associated with a deep spatially limited updraft (providing adiabatic cooling and a copious supply of transported water and dust). Particles thus generated are then transported roughly westward (downwind) by the larger-scale circulation, quickly exiting the particle growth region and becoming wholly subject to turbulent diffusion, gravitational sedimentation, and sublimation. “Dry” (i.e., without the prediction of water vapor or cloud fields) mesoscale model simulations of the Valles Marineris region [Rafkin and Michaels, 2003] exhibited strong and complex horizontal and vertical circulations associated with the canyon system. In particular, those results exhibited thermally driven vertical updrafts (of up to 40 m/s) above the canyon walls which extended ≥10 km in altitude. Together, such complex and significant circulations, while not directly effecting VMCT genesis, may nevertheless strongly perturb the overlying (and relatively nearby) atmospheric circulation in ways that may lead to localized dynamical effects at the VMCT altitude and location.

[19] Preliminary targeted mesoscale modeling (using the MRAMS model [Rafkin et al., 2001]) at the VMCT season and location by coauthor Michaels suggests that the results of such perturbations (perhaps akin to “constructive interference” in elementary physics) may indeed be important to VMCT genesis. In these simulations, perturbations by mesoscale canyon circulations appear to perturb the nearby (within a few hundred kilometers) atmosphere aloft in ways that generate significant (greater than 20 m/s), localized updrafts near where the eastern edge of the VMCT are modeled. It is hypothesized that only the strongest sector of these updraft phenomena is able to overcome significant environmental and microphysical barriers to cloud formation, producing a solitary VMCT. Modeled horizontal wind speeds at the VMCT altitude are more than sufficient to produce a downwind cloud trail within the time necessary to satisfy observational constraints. Thus this proposed mechanism is consistent with the very narrow range of observed VMCT locations (as the phenomenon is indirectly tied to specific underlying topography), the generally solitary nature of VMCT, their assumed afternoon and near-perihelion occurrence (thermally driven near-equatorial canyon updraft strength will be at a maximum in the early afternoon and near perihelion, potentially maximizing perturbations), and the observationally inferred characteristics of the cloud as a whole. Future work will explore this plausible VMCT genesis mechanism (and the clouds themselves) further.

5. Conclusions

[20] A new type of Mars atmospheric cloud is observed to form during the perihelion season (LS = 230°–260°) in proximity to the Valles Marineris and Noctis Labyrinthus canyon systems. Their form suggests cloud trails of extraordinary lengths (400–1000 km), tied to specific regions of local-scale (25–75 km) vertical updrafts that lead to the formation of small particle size (∼ 0.4μm) water ice clouds at 40–50 km altitudes above the surface. These ice clouds are transported rapidly westward by strong easterly zonal winds associated with the perihelion, southern subsolar middle atmospheric circulation over 1–2 h time scales during afternoon hours. A plausible VMCT genesis mechanism is proposed, in which strong canyon wall updrafts perturb the nearby (within a few hundred kilometers) atmosphere a complex nonlinear fashion (based on the lack of a simple dynamical explanation for the presence of relevant localized updrafts in preliminary “dry” MRAMS simulations), resulting in a focused, stationary dynamical phenomenon with large (perhaps tens of m/s) upward vertical velocities. In this scenario, only the strongest updraft within the dynamical phenomenon provides enough adiabatic cooling (and potentially water vapor and ice nuclei from lower levels) to generate a VMCT. This hypothesis may be consistent with the majority of observationally inferred characteristics of the cloud as a whole. Their correlation with specific locations (e.g., 20°S, 42°W) that do not present notable topographic features suggests a correspondence with similar locations of locally intense vertical updraft present in mesoscale simulations of atmospheric circulation over Valles Marineris. However, the nature of such a dynamical link is obscured by the substantial (200 km) latitudinal offset of these clouds and the simulated updrafts from the southern boundaries of the canyons.


[21] We are indebted to the excellent MRO and MSSS operations staff for the collection and processing of MARCI images presented here. Grant support for this work was provided by the MRO Program (under MSSS subcontract 06-0152).