Journal of Geophysical Research: Planets

Mars: Aeolian features and wind predictions at the Terra Meridiani and Isidis Planitia potential Mars Exploration Rover landing sites



[1] Orientations of 11,497 aeolian (wind related) features at the potential Mars Exploration Rover (MER) landing site in Isidis Planitia reflect formative winds from the northwest and are consistent with strong morning winds predicted by a regional-scale atmospheric model for Mars. Orientations of 3296 aeolian features mapped at the Terra Meridiani potential MER site suggest multidirectional wind patterns. The results for the Terra Meridiani site are consistent with gentle winds which change direction as a function of Martian season (predicted by a general circulation model) and erratic winds that occur diurnally (predicted by a regional-scale atmospheric model). Results for both sites give confidence that the predictions from the atmospheric models are basically correct when time of day and local topography are taken into account.

1. Introduction

[2] In the absence of known active volcanism, tectonic activity, or other surface processes, the action of wind appears to be the dominant agent of modification on Mars. Various lines of evidence suggest that much of the surface is mantled with windblown deposits of sand and dust. The type, location, and orientation of wind-related features, such as duneforms, provide clues to the interaction of the surface and the atmosphere, and can give insight into the validity of models of the atmosphere.

[3] The Mars Exploration Rovers (MER), named Spirit and Opportunity, will enable observations of windblown features from the ground. However, as discussed by Golombek et al. [2003], wind during the landing process is an important safety consideration for the MER and in the absence of direct measurements (which will not be made), we must depend on predictions from models. Potential landing sites have been identified for the MER in Isidis Planitia, centered at 4.31°N, 87.91°E and Terra Meridiani, centered at 2.07°S, 353.77°E [Golombek et al., 2003].

[4] In this paper, we describe the aeolian features in and near the proposed landing ellipses, map the orientations of features, and compare the results with winds predicted from models of the atmosphere. The wind orientations, given here as azimuths and as shown in the figures, refer to the directions toward which the winds blow.

[5] The primary data used in our analyses of aeolian features consist of images taken from Mars Global Surveyor (MGS) by the Mars Orbiter Camera (MOC; Malin et al. [1998]). Aeolian features were classified using the scheme used for Gusev crater [Greeley, 2003], and include dunes of various morphologies and albedo patterns (Figure 1). Six categories of dunes are included: (1) barchan dunes, characterized by their crescent-shaped planform, in which the “horns” point in the downwind direction, (2) asymmetric dunes that have an asymmetric cross section, in which the steeper slope is considered to represent the slipface and the downwind direction, (3) duneforms inferred to be traverse bedforms in which the axes are oriented perpendicular to the formative winds; similar features seen elsewhere on Mars could be granule ripples [Zimbleman and Wilson, 2002], (4) dunes within craters, (5) dunes within craters, in which the dune geometry (either barchan in planform or with asymmetric cross section) is used to infer the formative wind direction, and (6) dunes that are in close proximity to topographic features such as scarps, which might have influenced the wind patterns and subsequent orientations of the dunes.

Figure 1.

Images of the primary classes of wind-related features used to assess wind regimes in Isidis Planitia and Terra Meridiani; all images are oriented with north to the top. (a) Crescent-shaped dunes inferred to be barchans in Isidis Planitia; their planforms suggest formative winds from the northwest. (b) Inferred transverse duneforms in Terra Meridiani, in which the dune axes are oriented normal to the formative winds; the asymmetric cross section suggests formative winds from the lower left toward the upper right. (c) Duneforms accumulated in impact craters, which appear to serve as “traps” for windblown sand, as indicated in this example in Terra Meridiani. (d) The location and orientation of duneforms influenced by local topography, as shown here in Terra Meridiani, in which the hills on either side of the dune field appear to control the local winds and subsequent duneforms. (e) Dark wind streak in Isidis Planitia; formative wind is inferred to be from the lower left toward the top of the image. (f) Bright wind streak in Terra Meridiani, in which the formative wind is inferred to be from the upper left toward the lower right.

[6] Albedo patterns are qualitatively described as bright or dark in relation to the surfaces on which they occur. As reviewed by Greeley [2003], bright wind streaks are considered to be deposits of fine dust settled from the atmosphere, whereas dark wind streaks could be: (1) coarse materials (e.g., sand), (2) exposed bedrock from which bright material (i.e., dust) has been removed, (3) deposits of compositionally dark material, or (4) lag deposits of gravels, cobbles, etc. which might constitute a type of desert pavement. Each of these possibilities would indicate a higher wind surface shear stress than for the brighter surrounding terrain. Bright and dark wind streaks commonly are found in association with craters and other topographic obstructions to the wind, and their azimuths (measured as extending from the topographic features) are inferred to represent the downwind directions of the formative winds (i.e., a dark wind streak associated with a small crater “points” downwind). Some linear dark streaks are considered to be the “tracks” left by dust devils [Edgett and Malin, 2000]. The axes of these patterns are considered to indicate the orientation (east-west or west-east, etc.) of the general winds at the time of formation, but it is not possible to determine the specific direction for these winds.

[7] We analyzed global-scale wind predictions derived from the general circulation model (GCM) of Haberle et al. [1993]. Our analyses reflect averages for four intervals of time (spring, summer, autumn, and winter) over one full Martian year. Values shown in Figures 6 and 9 reflect surface shear stresses (the critical value for the transport of particles) and are for the highest 5% winds. By terrestrial analogy, these are inferred to reflect the directions when most of the dune-forming processes would occur. This enabled assessments of near-surface winds as a function of season for large areas (each cell is 7.2° longitude by 9° latitude).

[8] Winds for higher spatial resolutions were assessed based on predictions from the Mars Regional Atmospheric Modeling System (MRAMS) of Rafkin et al. [2001] for the period planned for the MER landings (∼Ls = 340°), which is southern hemisphere summer.

[9] MRAMS is initiated using output from the GCM and its grid size is variable so that areas of interest can be “zoomed” to provide more detail. As described by S. C. R. Rafkin et al. (MRAMS mesoscale model results for MER, submitted to Journal of Geophysical Research, 2003, hereinafter referred to as Rafkin et al., submitted manuscript, 2003), MRAMS uses Mars Orbiter Laser Altimetry for topography binned to 1/16th of a degree, and 1/8 degree Thermal Emission Spectrometer data for albedo and thermal inertia of the surface. The surface roughness scale is set to a constant of 3.0 cm. Wind values predicted in Figures 5 and 10 are for a height of 14.5 m above the surface.

2. Isidis Planitia

[10] The general geology and characteristics of this site are described and reviewed by Golombek et al. [2003]. Figure 2 shows the footprints of the MOC data used in our study for the landing ellipse, which is located in the southern part of Isidis Planitia and is bounded by mountains to the south. Figure 3 shows rose diagrams for the 11,497 aeolian features analyzed. Averages for the orientations of these features as mapped from MOC image observations are shown in Figure 4a.

Figure 2.

Image of potential MER landing site in Isidis Planitia, showing the footprints of the MOC frames (black and white outlines) used in this study. The landing site is located in the southern part of the Isidis basin and is bounded by the mountains visible at the bottom of the image.

Figure 3.

Rose diagrams showing the azimuths for the aeolian features in Isidis Planitia; note that the scale for the numbers of features compiled changes among the diagrams. The (a) dark wind streaks, (b) barchan dunes, and (c) asymmetric duneforms all suggest formative winds from the northwest, and would suggest that (d) the dune axes for the inferred transverse bedforms were probably formed by the same winds. Similarly, (e) duneforms within the craters that have asymmetric geometries are thought to have formed by these same winds, and (f) the other dunes within craters have axes suggestive of transverse features probably formed by winds from the northwest.

Figure 4.

(a) Image showing the Isidis Planitia site and the average orientations of features mapped from MOC images. Arrows point in the inferred downwind directions. Formative winds appear to be dominantly from the west northwest. (b) Image showing the orientations of aeolian features mapped in Terra Meridiani; arrows point in the downwind direction and appear to be much less consistent as a set than those in the Isidis basin.

[11] Analyses of the orientations of the wind-related features suggest that: (1) the barchan and asymmetric dunes were formed by winds blowing from the northwest, (2) inferred transverse duneform axes are oriented perpendicular to the orientations of the barchan dunes, suggesting that they resulted from winds also blowing from the northwest, (3) the dark streaks indicate formative winds from the northwest, and (4) the axes of dark linear streaks (i.e., dust devil tracks) are northwest-southeast and, based on the orientations of the dunes and dark streaks, the dark linear streaks are likely to reflect winds from the northwest. Similarly, barchan and asymmetric dunes found within craters show orientations consistent with winds from the northwest.

[12] Figure 5 shows winds predicted by MRAMS for the Isidis site; the prevailing winds over the ellipse are from the northwest, consistent with the wind-related features observed on the images. Figure 6 shows GCM predictions in the area for four seasons, in which the strongest winds are modeled to be from the southwest during Ls 0°–180°. None of the GCM wind directions are consistent with the orientations mapped for the aeolian features, nor with the MRAMS predictions for the landing ellipse. However, as shown in Figure 5, MRAMS suggests that the global-scale winds predicted by the GCM for the landing period (Ls = 340°) sweep across Isidis Planitia, but the strong morning winds are interpreted to be “turned” toward the southeast, partly influenced by the mountains that bound the Isidis basin to the south.

Figure 5.

Wind directions predicted by MRAMS (from Rafkin et al., submitted manuscript, 2003) for the Isidis site for Ls = 320° (time of potential MER landing) for (a) the morning (0900) and (b) the afternoon (1500) at the same scale and for (c) the afternoon zoomed into the center of the landing ellipse (designated “IP”) at higher spatial scales. The prevailing winds are from the northwest, consistent with the orientations of most of the aeolian features (Figures 3 and 4). Note that in Figures 5a and 5b the winds enter the region from the NNE but are “turned” toward the SSE over the landing ellipse, apparently partly influenced by the mountains that bound Isidis Planitia to the south. Shading gives the wind speed at 14.5 m above the surface; arrows show wind directions.

Figure 6.

Wind directions for the Isidis Planitia region predicted by the GCM as a function of season (each cell covers 9° of longitude by 7.2° of latitude). Symbols indicate surface shear stress, in which each short slash is 5 × 10−4 N/m2, each long slash is 10 × 10−4 N/m2, and each triangle is 50 × 10−4 N/m2. Landing ellipse is shown just north of the mountains that bound Isidis Planitia.

3. Terra Meridiani

[13] The Terra Meridiani area has been described by Golombek et al. [2003] as a potential MER landing site. This area has also been referred to as the “hematite site,” based on data from the MGS Thermal Emission Spectrometer that suggest the presence of this iron mineral in the region [Christensen et al., 2000]. Aspects of the general geology are discussed by Arvidson et al. [2003] and Newsom et al. [2003].

[14] We have identified 3296 wind-related features within the areas of the landing site ellipse (Figure 7), which include barchan dunes, transverse duneforms, dark wind streaks, and bright wind streaks. The orientations of these features are shown as rose diagrams in Figure 8 and plotted on the base map image (Figure 4b). Analyses suggest that the: (1) barchan dunes reflect winds from the west, (2) inferred transverse duneforms are likely to develop by winds also from the west, (3) dark wind streaks as a group suggest formative winds from the SSE, but the total number is small, and (4) bright wind streaks suggest bimodal formative winds (one from the southeast, the other from the northwest). In comparison to the data for Isidis Planitia, there is much more scatter in azimuthal variations for the features mapped in Terra Meridiani.

Figure 7.

Image of the potential MER landing ellipse in Terra Meridiani and footprints of the MOC images used to map aeolian features.

Figure 8.

Rose diagrams showing azimuths for the aeolian features analyzed in the Terra Meridiani site; note that the scales for the numbers of features analyzed changes with the category of the feature.

[15] GCM wind predictions show generally mild winds throughout the year in Terra Meridiani, and that the prevailing winds change direction as a function of season (Figure 9) with strongest winds occurring in the northern autumn and winter from the east. MRAMS predictions for the site as a function of time of day show very low wind speeds at night, with slight increases in speed as daytime heating occurs, then a return to calm winds after sunset. As shown in Figure 10, the afternoon winds are erratic, with the suggestion of local (convective?) “cells” that could reflect upwelling and downwelling air masses. Because these are the strongest winds, they most likely are responsible for the orientations of the features seen on the surface, and could account for the lack of consistent azimuths for the dunes and wind streaks.

Figure 9.

Winds in the Terra Meridiani area predicted by the GCM as a function of season, showing the location of the landing ellipse. Symbols indicate surface shear stress, in which each short slash is 5 × 10−4 N/m2, each long slash is 10 × 10−4 N/m2, and each triangle is 50 × 10−4 N/m2.

Figure 10.

Winds for the Terra Meridiani area predicted by MRAMS for Ls = 342° (time of potential landing) for (a) 0800 in the morning and (b) 1400 in the afternoon showing regional-scales patterns, and (c) local winds at 1400 centered on the landing ellipse. Local winds suggest patterns of upwelling and downwelling (local “cells”). Shading gives the wind speed at 14.5 m above the surface, arrows show wind directions, and “TM” shows the center of the landing ellipse.

4. Summary and Conclusions

[16] Aeolian features observed in the potential MER landing ellipse for the Isidis Planitia site all suggest formative winds from the northwest, and are consistent with morning winds predicted from the mesoscale MRAMS. The apparent lack of correlation with the global-scale predictions is suggested to be an artifact of the low spatial resolution of the GCM, and can be taken into account by the more detailed modeling from the MRAMS, which suggests that regional winds are redirected toward the SSE from their global pattern, perhaps by impingement on the mountains that bound the Isidis basin to the south.

[17] The orientations of aeolian features at the Terra Meridiani site have more scatter than those in the Isidis Planitia site. The features in Terra Meridiani are considered to reflect the generally low-speed winds (regardless of season) and the generation of erratic afternoon winds, which would produce randomly oriented surface features.

[18] In general, the correlation of wind-related features with predictions of near-surface winds from atmospheric circulation models (or the explanation of discrepancies based on reasonable assumptions) gives confidence that the MRAMS regional-scale atmospheric model is basically correct for the sites analyzed. It should be noted, however, that most of aeolian features seen at both sites are duneforms and are probably composed of sand-sized particles moved in saltation. By terrestrial analogy, such features reflect the dominant wind directions averaged over durations of decades or more. In contrast, features such as bright and dark streaks (particularly the inferred dust devil tracks) are more ephemeral and could form by winds that are more transitory, as might be expected from local weather.

[19] Both the model predictions and the orientations of wind-related features can be assessed by future landed systems, such as the MER. For example, ripples, duneforms, and dunes could be observed from the surface for comparison with the orientations of the features seen from orbit. Although there is no direct provision for measuring wind speeds and directions by MER, observations of potential changes on the surface, such as the erosion and/or deposition of sand and dust on and around the rovers, have the potential to provide insight into the local wind regime.


[20] This work was supported by NASA through the Mars Exploration Rovers Project, Planetary Geology and Geophysics Program, and the Mars Data Analysis Program. Charles Hewett, Data Manager for the NASA-sponsored Space Photography Laboratory at Arizona State University is thanked for assistance with the images used in this study and Lynn Neakrase is credited for the GCM processing. We acknowledge R. M. Haberle and S. C. R. Rafkin for atmospheric modeling predictions. The manuscript benefited greatly from comments by Lori K. Fenton and an anonymous reviewer.