Gusev crater has a complex history beginning with its formation by an impact in the Noachian Period and subsequent evolution, including erosion and deposition associated with Ma'adim Vallis, which cut through Gusev crater from the south. Various wind-related features are found on the crater floor, including albedo patterns, bedforms (such as barchan dunes), and elongated hills possibly reflecting erosion by the wind. Comparisons of the orientations of these features with predictions from global and atmospheric models show wind patterns as functions of season and time of day. Strongest global winds occur out of the southeast in the late fall and early winter and appear to be responsible for the formation of a large dark streak that drapes 100 km across the crater floor. On an intermediate-scale, bright wind streaks correlate with nighttime winds that flow into the crater from the rim and outlying high areas; these streaks are interpreted to be deposits of dust settled from the atmosphere during times of positive atmospheric static stability. Smaller features include various dark streaks and duneforms, the orientations of which are inferred to represent local zones of high wind-shear surface stress. These features also tend to correlate with strong afternoon upslope winds caused by the uneven heating of the atmosphere across the crater. Thus the large dark streak reflects global-scale wind patterns, the smaller dark features and duneforms represent flow of strong winds out of the crater in the afternoon, and the small bright streaks represent deposition of dust associated with nighttime flow of wind into the crater. The correlation between the wind regime predicted by the models and the orientations of the aeolian features gives confidence that the models are essentially correct. Because Gusev is a prime candidate landing site for landers, including the Mars Exploration Rovers, understanding the interplay of the atmosphere and the surface (including the erosion and deposition of windblown particles) is important in assessments of the science potential for these missions. Despite the abundance of wind-related features on the floor of Gusev, only about 1.2% of the proposed landing ellipse area is covered with organized bedforms (e.g., dunes), although mantles of sand and dust are also likely to be present. On the basis of an analysis of random landing points within the ellipse, traverse distances to ridges, ejecta from small craters, and other potential sources of rocks are within 150 m traverse distance for 95% of the touchdown points.
 Winds might be the dominant process on Mars for modifying the surface in the current climatic regime. Wind-related surface features are seen from orbit and the ground, and include duneforms, erosional features, and various albedo surface patterns that change with time, apparently in response to short- and long-term variations in the wind. Consequently, understanding the formation of wind-related deposits and landforms is important in the interpretation of orbital remote sensing data and measurements obtained from future landed missions.
 The floor of Gusev crater is a candidate site for the Mars Exploration Rovers (MER), launched in 2003 [Anderson and Parker, 2002; Golombek et al., 2003; Carbol et al., 2003]. High-resolution images from the Mars Orbiter Camera (MOC) [Malin et al., 1998] show numerous wind-related features within Gusev crater. Comparisons of these features with winds predicted by numerical atmospheric models can provide insight into the geological evolution of Gusev crater in the post-lacustrine stage and its suitability as a site for the MER and other landers, and is the topic of this report.
 Results from these studies show that Gusev crater formed by an impact during the Noachian Period as part of the final stages of heavy bombardment early in Mars' history. Subsequently, the crater and the surrounding terrain were degraded by fluvial dissection, groundwater sapping, and flooding from channels that were precursors to Ma'adim Vallis, leading to deposition of materials on the floor of Gusev crater. Hesperian-age eruptions from Apollinaris Patera, a volcanic center north of Gusev, possibly emplaced pyroclastic materials (Figure 2) and lava flows on the northern flanks of the crater [Scott et al., 1993]. Concurrently, the channels of Ma'adim Vallis and other fluvial systems were established, leading to continued deposition in Gusev crater. Episodic flooding, deposition, and erosion appear to have continued into the Middle Amazonian Epoch. An elevated terrain extending 35 km onto the floor of the crater from the mouth of Ma'adim Vallis is thought to represent fluvial delta deposits emplaced in a paleolake environment. At times, flow through Ma'adim might have been blocked, creating locally flooded segments of the channel. Partial drainage of Gusev crater probably occurred through a breach in the northwestern part of the crater rim.
 The floor of Gusev crater has been modified by a wide variety of processes, including erosion by late-stage fluvial activity, wind erosion and deposition, impact cratering, and the receipt of mass-wasted materials from the crater walls [Kuzmin et al., 2000]. Comparisons of features observed on the crater floor with periglacial landforms seen on Earth suggested to Cabrol et al.  that some of the deposits were probably subjected to freeze-thaw conditions, and that putative paleolakes within the crater were ice-covered [Grin and Cabrol, 1997]. Collectively, these processes had the potential to generate a wide variety of physically and chemically altered materials, including clastic particles in the size ranges amenable to wind transport (fine dust approximately a few microns in diameter to coarse sand a few millimeters in diameter [Greeley and Iversen, 1985]).
 Our analysis of the Viking Orbiter Infrared Thermal Mapper (IRTM) [Kieffer et al., 1977] and Mars Global Surveyor (MGS) Thermal Emission Spectrometer (TES) [Christensen et al., 2001] data for Gusev crater suggests that much of the floor is mantled with dust and sand. Although the thicknesses of the deposits cannot be determined directly, the preservation of features such as craters smaller than a few tens of meters in diameter (visible in high-resolution MOC images) suggests that recent deposits are less than a meter thick. The infrared data also suggest that the particle sizes and abundances of rocks increase across the floor of the crater toward the mouth of Ma'adim Vallis.
1.2. Gusev as a Potential Landing Site
 The current strategy for Mars' exploration includes orbiters, landers, rovers, and the eventual return of samples to Earth, all governed by a set of prioritized science goals and objectives [Greeley, 2001]. A principal goal is centered on the search for evidence of life and the characterization of present or past habitats conducive for organic evolution [McCleese, 2001]. Because the geological history of Gusev crater is generally considered to have involved substantial quantities of liquid water, which might have been present for long periods of time, it has been considered a primary target for astrobiological exploration [e.g., Cabrol et al., 1994, 1996a; Landheim et al., 1994].
 The MER are scheduled to land on Mars in early 2004 to begin extensive surface operations at two sites. Because of various engineering constraints, such as the elevation and latitude of the sites, the potential locations for the MER landings are limited [Golombek et al., 2001]. Through a process involving the scientific and engineering communities, potential sites for the MER have been narrowed to a handful [Golombek et al., 2002, 2003], one of which is Gusev crater [Cabrol et al., 2002, 2003; Grin et al., 2002].
 A principal concern for safe landings is the near-surface wind shear. Unfortunately, very little data are available for such winds on Mars. Although wind data were collected during the descent of the Viking and Mars Pathfinder (MPF) landers, these represent only three places on Mars for three brief periods of time. Meteorology data were also collected from the landers for the lifetimes of the missions, but all three sites represent rocky plains that are not representative of all potential sites, especially those being considered for the MER landings. Of particular concern are sites situated adjacent to significant topographic relief, including sites on the floors of canyons or within craters, such as Gusev. Because data are lacking for these terrains, planners must rely on atmospheric model predictions of the near-surface winds for the times of the proposed landings. Initial mesoscale model runs for various candidate MER landing sites suggest that local winds could be a problem [Rafkin, 2001; Toigo and Richardson, 2001; Kass and Schofield, 2001; Kass et al., 2003; Rafkin and Michaels, 2003]. Of particular concern are slope winds that vary in direction and strength as a function of time of day.
 In this report, we outline the wind patterns predicted from atmospheric models, describe the wind-related features observed in images of Gusev crater and compare them with the predicted winds, and discuss the implications of the results for understanding the surficial geology of the crater floor and the site as a potential for landed missions.
2. Wind Regime
 Results from two numerical models of Martian atmospheric circulation were used in our study. The NASA Ames General Circulation Model (GCM) of Haberle et al. [1993, 1999] was used to derive estimates of winds on global scales as a function of Martian season. The Mars Regional Atmospheric Modeling System (MRAMS) of Rafkin et al.  and Rafkin and Michaels  was used to predict local winds in and around Gusev crater as a function of time of day for two seasons (summer and winter). Because the MRAMS is not a global model and simulates only several Mars days at a time, it uses GCM results to produce its initial condition and time-dependent boundary conditions.
2.1. General Circulation Model
 The GCM is three dimensional, consisting of a 25 by 40 cell computational grid of Mars with 13 vertical layers extending from 0 to 47 km altitude, the spacing of which increases with height. A typical run consists of 1200 simulated hours, during which the seasonal date is advanced at a rate appropriate for Mars' orbit. It incorporates topography, albedo, and thermal inertia of the surface. The near-surface boundary layer profile is predicted for each cell, from which the magnitude and direction of the surface wind shear stress are diagnosed and shown graphically as vectors. Because most winds in the vicinity of Gusev crater are predicted to be below “threshold” for sand and dust entrainment, we used only the strongest winds (greatest 5 percent by magnitude) in our comparisons with wind-related surface features.
Figure 3 shows the GCM values for 8 periods (2 intervals for each season) for the two geographic cells within which Gusev crater occurs. The shear stress vectors reflect the strongest winds, which occur in late southern fall and winter and blow from the south-southeast to the north-northwest. The wind then diminishes in strength and reaches a minimum in the spring. Moderate winds occur in the southern summer and are oriented toward the south and southeast, in the opposite direction to the prevailing winter winds.
2.2. Mars Regional Atmospheric Modeling System
 The MRAMS [Rafkin et al., 2001] is based on the system of Pielke et al.  developed for Earth. The grid-cell size of MRAMS is variable and can be multiple-nested in size to “zoom” into local areas of interest. MRAMS is initiated using output from the GCM to provide the global-scale winds. Figure 4 shows MRAMS results from the third nested grid for Gusev crater and the surrounding area for two seasons (winter and summer). This grid consists of 55 by 55 points spaced 12 km apart, and is nested within two outer grids (not shown) with grid spacing of 60 km and 240 km, respectively. The outermost model domain is several thousand kilometers on a side. MOLA topography binned to 1/16th of a degree, and 1/8 degree TES albedo and thermal inertia are used to define the surface characteristics. The surface roughness length is set to a constant 3.0 cm.
Figure 4 shows horizontal wind vectors at a height of 14.5 m above the surface throughout the day at Ls = 143° (southern winter) and Ls = 330° (southern summer and the period for the proposed MER landing). In both seasons, during nighttime, winds generally flow downslope into the crater, especially from topographically high areas that are southeast and south of Gusev crater. Because the topography is very low on the northwest rim of Gusev (the inferred breached rim), slope winds in this area are minimal and the general wind patterns reflect the regional trend from the southeast predicted by the GCM.
 In the late morning hours, winds begin to flow out of the crater. By noon, the outward flow dominates, with reverse flow occurring up the channel of Ma'adim Vallis in the opposite direction to the winter regional winds. By afternoon (1500 hour) the dominant winds are toward the southeast, with very strong flow across the crater floor and up the channel. By late evening (2100 hour) the nighttime downslope flow from the high areas and into the crater has become reestablished.
 To first-order, the local winds in and around Gusev crater appear to be controlled primarily by slope winds and local heating. For both winter and summer, the nighttime, near-surface air over the higher elevations cools more rapidly than that at lower elevations (e.g., the crater floor) and flows downslope due to gravity. Similarly, the near-surface air over the surrounding higher elevations heats more rapidly than that within the crater during the day, creating a pressure gradient that leads to upslope winds.
3. Wind-Related Surface Features
 Surface features associated with winds in the Gusev crater area can be classified as: a) albedo patterns, b) deposits of windblown materials (e.g., various bed forms, such as dunes), and c) features resulting from wind erosion. All available imaging data for the area were analyzed, including those from Mariner 9 (obtained principally in 1972), the Viking Orbiters (taken between 1976 and 1980), the Russian Phobos 2 spacecraft (1989), and Mars Global Surveyor (MGS, beginning 1997 to the present). Collectively, these data sets span a period of some 3 decades representing a variety of Martian seasons. Our descriptions of the features are based principally on MOC images, the coverage of which is incomplete for the study area (Figure 3). Consequently, the descriptions of the locations and orientations are biased by the available images.
3.1. Albedo Patterns
 Low-albedo and high-albedo areas that appear, disappear, or change in size or shape with time were first recognized on Mariner 9 images and were attributed to aeolian processes [Sagan et al., 1972]. Termed variable features, they have no apparent topographic relief and could involve the deposition and removal of a layer of particles as thin as a few microns [Wells et al., 1984]. In general, as reviewed by Greeley et al. , high-albedo variable features are thought to be deposits of very fine particles (i.e., dust), whereas low-albedo features are 1) coarser particles (i.e., sand), or 2) surfaces that have been swept relatively free of dust to expose a dark substrate, or 3) deposits of dark material. The term “albedo” is used to describe relative brightness, rather than given in quantitative terms because the MOC images have not been radiometrically calibrated.
 The most prominent albedo pattern in Gusev is a dark zone that extends from the inferred delta deposits of Ma'adim Vallis northwest across the floor of the crater for more than 100 km. This pattern is very pronounced in Viking Orbiter images (Figure 5) but in MGS TES data and MOC wide-angle images, taken some 23 years later, the center of the pattern is brighter. Phobos 2 Termoscan [Selivanov et al., 1998], MGS TES, and Mars Odyssey THEMIS data [Milam et al., 2003] show differences in albedo and thermal properties of the surface in the middle part of the pattern, suggesting changes on the surface. We attribute the changes to aeolian processes. The overall trend of the albedo pattern is consistent with the prevailing and strongest global-scale winds from the southeast, which are enhanced by winds flowing down the channel during the nighttime. Prevailing winds are defined as those that are predicted by the GCM (Figure 3) to be the strongest within the confidence of the model. As described below, the eastern part of this zone also includes very complex patterns of dark streaks when viewed in high resolution, the orientations of which indicate formative winds from the west for the dark elements.
 In addition to the broad low-albedo pattern that extends across the crater floor, smaller patterns are seen in association with small craters and other landforms, including ridges (Figure 6). These features, termed wind streaks, are thought to serve as wind vanes indicative of the prevailing wind directions at the time of their formation. Bright wind streaks are seen 1) as extensions of the broad albedo pattern described above and 2) in a cluster in the north part of the Gusev floor. They are as long as 8 km seen on Viking Orbiter images, but also include very small features a few hundred meters long seen on some MOC images in the same area. High-resolution MOC images show that some bright streaks are complex, with dark mottling and small dark streaks within the general zone of high albedo.
 Dark wind streaks are found in several locations on the crater floor, but are most numerous within the broad albedo zone described above. Nearly all of these features are oriented approximately normal to the west rim of Gusev crater, indicative of winds from the west and consistent with downslope nighttime winds.
 A prominent set of dark wind streaks occurs in the northwest part of the crater floor, within a partly “flooded” 34-km impact structure (Figure 7b), provisionally named Zutphen. These streaks are as long as 950 m and are uniformly oriented toward the northwest, a direction consistent with both the regional prevailing winds (predicted by the GCM; Figure 3) and with wind flow out of the crater during the daytime (predicted by MRAMS; Figure 4). It is also noted that the MOC image showing these streaks was taken at a season when the global-scale wind is from the northwest (late fall, early winter), a direction opposite to that of the wind streaks. Although we do not know when the streaks formed, this suggests that the local daytime winds in this area are more important in terms of interaction with the surface than the global-scale winds.
 Long narrow dark streaks seen on Viking Orbiter images of Mars were suggested by Grant and Schultz  to be the tracks from vortical atmospheric activity. In confirmation of this idea, these albedo features are now generally recognized to be tracks from the passage of dust devils [Edgett and Malin, 2000], during which dust is removed from the surface to expose a darker substrate. Such features are seen in the south-central part of the Gusev crater floor (Figure 6d). Smaller tracks are a few hundred meters long by 10–20 m wide. Larger features have lengths that exceed the width of the MOC images (approximately a few kilometers), but are still only tens of meters wide. The azimuths of the tracks vary considerably, but they are generally oriented west-to-east. Although there is no definitive means to determine which wind direction was involved in their formation (east to west, or west to east), small dark streaks emanating from a ridge in the same area as the tracks would suggest winds from the west toward the east, consistent with afternoon slope winds.
 An unusual set of dark wind streaks occurs in the middle of the Gusev floor, within the eastern part of the dark zone shown in Figure 5. These streaks (Figure 6e) appear to be combinations of dust devil tracks and dark crater-associated wind streaks. Several hundred streaks of this type occur in the area, the orientations of which suggest generally west-to-east winds, correlative with afternoon winds predicted by the MRAMS. We propose that the elements of these streaks that do not appear to be directly due to the existence of a local topographic obstruction result from complex vortical winds, including dust devils, occurring during conditions of atmospheric instability (daytime). As shown in Figure 6e, not all surfaces in the area are affected, as indicated by the many small impact craters that lack dark streaks. We suggest that local vortices (both dust devils and horizontal vortices shed from crater rims) are spawned by erratic winds that occur in some areas but not others.
3.2. Deposits of Windblown Material
 A variety of windblown deposits occurs on Mars, including mantles of dust and sedimentary layers, some of which could be lithified or indurated. Such deposits could exist in and around Gusev crater, but their forms cannot be distinguished from sedimentary deposits of non-aeolian origin with the available data. The most readily recognizable windblown deposits are features inferred to be various bedforms, such as dunes and ripples. Although we follow the dune classification of McKee  for Earth, many of the dune-like features seen on Mars are too small to be identified definitively and classified; consequently, we refer to the ambiguous features as duneforms. Most dunes on Earth are typified by an asymmetric cross section with respect to the wind and have a gentle upwind slope and a steeper downwind slope, termed the slipface, which forms at the angle of repose for sand (i.e., 33–35°). In general, duneforms indicate the presence of sand-size material and sufficiently strong winds for sand transport in saltation at the time of dune formation. Duneforms on the crater floor of Gusev (Figures 8–10) include isolated or clustered features, fields of duneforms, intercrater dunes, barchanoids dunes, possible reversing dunes, and a set of features which might represent paleodunes.
 Isolated duneforms and clusters of duneforms are common. They range in size from 20 to 210 m along their crests (average length is 60 m) and are typically about 15 m from the edge of the windward side to the base of the slip face. Dunes of this form are represented in Figure 8c showing the southwest part of the crater. Although these features are small, they appear to have orientations indicative of winds from the northwest, consistent with the dark wind streaks seen in the same area. As shown in Figure 8d for another area, local topography can influence the orientations of some small dunes.
 Many of the small impact craters superposed on the floor of Gusev serve as traps for windblown sand and contain sets of dunes (Figure 9). The axes of the dunes within the craters suggest winds consistent with those inferred from the wind streaks adjacent to the craters in the same general area, but because the wind patterns in, over, and around craters are complex [Kuzmin et al., 2001], the orientation of inter-crater dunes should not be used to assess general wind patterns.
 Aside from dunes within craters, a few sets of features in the Gusev area might be considered fields of dunes (Figure 10). One set forms a patch about 1 kilometer across on the surface just north of the Gusev crater rim crest. Individual features in the set consist of mounds and duneforms that are generally smaller than 20 m across. Some of the duneforms are aligned in transverse sets, similar to dunes in fields on Earth. Spacings of the individual duneforms increase on some margins of the field, also comparable to terrestrial dune fields.
 Barchan dunes and barchanoid ridges are present within channel segments of the eroded deltaic deposits at the mouth of Ma'adim Vallis (Figure 7c). Individual dunes average 50 m (horn tip to tip) by 15 m while barchanoid ridges have crests as long as 100 m. The orientation of the barchan crescents and the location of the inferred slip faces show that these dunes formed by winds blowing up-channel toward the south southeast, consistent with the strong afternoon winds predicted by the MRAMS (Figures 4 and 7d).
 Some bedforms appear to be reversing dunes (i.e., the result of two prevailing winds oriented at about 180° azimuth to each other). Figure 11 shows the area immediately northwest of the Gusev crater rim and a set of bedforms that are bilaterally symmetric across the dune crest, a geometry typical for reversing dunes. Their crests are consistent with the prevailing global-scale winter winds (south southeast to north northwest) and the reversed winds toward the southeast in the spring and summer. In this part of the Gusev area, these “reversed” winds are also reinforced by the daily daytime winds toward the northwest out of the crater. It must be recognized that these dunes might be linear duneforms (i.e., their axes would be parallel to the formative winds), but we suggest that they are transverse features, given the model-predicted winds.
 Some parts of the Gusev floor exhibit small ridges that could be paleodunes. Unlike the duneforms described in the previous paragraphs which are relatively brighter than the terrain on which they occur, the putative paleodunes have about the same brightness as the background terrain, suggesting that the entire area is uniformly mantled with dust, and that active saltation is not taking place. The crests of the features are oriented northeast-southwest, consistent with transverse dunes formed by either northwesterly or southeasterly winds (the prevailing global-scale winds).
3.3. Wind Erosion Features
 Wind erosional landforms of the size visible on orbital images include blowouts (shallow basins formed by wind deflation of particles) and yardangs (wind-sculpted hills that have the form of an inverted boat hull). Although neither blowouts nor yardangs are identified in the Gusev area, the eastern part of the crater floor shows numerous small knobs tens of meters across (Figure 12), some of which are aligned in sets oriented north northwest. These features could be indicative of wind erosion consistent with the prevailing regional winds.
4. Discussion and Conclusions
 Geological wind processes require a supply of fine particles and winds of sufficient strength to move them. Gusev crater exhibits a complex geological history involving impact, fluvial, lacustrine, and other processes capable of generating abundant sand and dust. Volcanic eruptions from Apollinaris Patera might also have contributed fine particles to the area. These interpretations are supported by remote sensing data in the mid-infrared part of the spectrum obtained by the Viking and MGS spacecraft, which suggest the presence of fine-grained material on the floor of Gusev crater.
 Various duneforms and albedo patterns show that winds of sufficient strength to redistribute fine particles have occurred in the past, and could be occurring in the current environment. The orientations of wind-related features are consistent with predictions of winds derived from models of the atmosphere at both global and regional scales. Changes in the broad albedo pattern on the floor of Gusev crater observed in the time between Viking Orbiter and subsequent missions suggest that surface modifications by the wind are currently active.
 Bright wind streaks are generally regarded as deposits of dust settled from the atmosphere. Veverka et al.  proposed that bright streaks form in periods of positive atmospheric static stability, during which craters and other topographic obstacles block the atmosphere (creating a “dead zone” in the wake of the obstacles), leading to dust deposition. This mechanism can explain cases in which the same crater can have bright streaks oriented in one direction and dark streaks oriented in the opposite direction. In general, atmospheric stability is expected to occur during the nighttime. Figure 13 (also see Figure 4) shows the relatively good qualitative agreement between the orientation of bright streaks and the nighttime wind pattern predicted from the MRAMS, consistent with the model of Veverka et al. .
 Predictions of wind strengths from atmospheric models suggest that some winds are above particle threshold strengths, particularly as related to strong diurnal slope winds. The strongest winds predicted by the models and inferred from the orientations of the aeolian features show patterns in and around Gusev crater as functions of local topography, time of day, and Martian season. GCM results predict that strongest regional winds occur in the late fall to early winter and are from the southeast. MRAMS results suggest that strong diurnal winds occur in the late afternoon with flow out of the crater, and at nighttime after about midnight with flow into the crater. The strongest surface wind shear occurs at the mouth of Ma'adim Vallis (where the winds appear to be channeled) and on the southeast wall of Gusev crater where downslope winds are enhanced by flow from the elevated terrain southeast of the crater.
 Wind-related activity probably dominates the surface today and has been operating for a geologically long period of time. However, the preservation of impact craters smaller than 10 m in diameter throughout the Gusev area, including those on the floor of the crater, suggest that the amount of surface lowering by wind erosion is less than a few meters, assuming that the craters have not been exhumed. The windblown deposits, such as the dunes, probably consist of material that has been distributed and redistributed repeatedly throughout the history of Gusev. Potential “sinks” (i.e., sites of long-term deposition) include the floors of small craters and, possibly, interleaving of aeolian sediments with fluvial and lacustrine sediments in the past.
 The presence of windblown materials could be significant for landed spacecraft operations, data acquisition, and interpretation of results. For example, albedo patterns and various wind-related bedforms are ubiquitous and it is likely that a substantial part of the floor is covered with windblown sand and dust. Consequently, remote sensing data obtained from orbit or from the ground must take into account the potential signatures from mantles of windblown sand and dust. For scientific objectives related to surface measurements to be made on rocks, there must be a means to access such specimens, such as the mechanical arm on the MER.
 In an effort to assess the potential problems posed by aeolian materials for a landed mission, we analyzed the MOC images for the proposed MER landing ellipse. The ellipse is 96 km long by 19 km wide with the major axis oriented on an azimuth of 76°. About 58% of the ellipse is covered by MOC images. Using a random number generator and a gausian distribution within the ellipse, 100 potential landing points within the ellipse were determined, 69 of which fell within MOC images. Assuming that relatively fresh impact craters, ridges, and small knobs would afford the opportunity for the MER to have access to rocks, we then determined the shortest distance to these features from the landing points; 95 percent of all points are from 0 to 150 m, well within the capability of the MER.
 We also determined the area within the ellipse covered by bedforms (dunes, ripples). Despite the apparent abundance of these features on the floor of Gusev, only 1.2% of the ellipse is covered by aeolian bedforms at the resolution of the MOC images (1.5 to 3.5 m/pixel). However, it should be recognized that there is a strong likelihood that much additional surface area is mantled with windblown material that is not organized into distinctive bedforms, as is the case at the Mars Pathfinder and Viking landing sites.
 Most landed systems being considered for Mars, including the MER, will have instruments to determine the composition of surface materials. Windblown materials should be considered prime targets for this purpose. For example, dust settled from the atmosphere represents a global “homogenization” of weathered material, while sand (as in the bedforms) is most likely derived locally. Low-albedo areas, as indicated by various wind streaks, could represent bedrock exposures or lag deposits of coarse-grained material, from which the dust has been winnowed. Given these considerations, the identification of different windblown terrains and features could enable targeting areas to test these ideas and as a means for increasing the diversity of “samples” for compositional measurements.
 In conclusion, active winds, aeolian features, and windblown sediments constitute important aspects of Gusev crater that must be taken into account in the interpretation of the geological history and in considerations of the site for future landings. Because there appears to be agreement between the wind regime predicted by atmospheric models and the observations of the orientations of the wind-related features, we have confidence that the models are valid.
 We thank Nathan Seaver, Shane Thompson, Patrick Whelley, and Veronica Zabala for image processing, Charles Hewett of the NASA-sponsored Space Photography Laboratory at ASU for data access, Sue Selkirk for preparation of the figures, Eddie Lo for computer support, and Stephanie Holaday for final word processing. This research project was supported by NASA grants from the Mars Exploration Program, Planetary Atmospheres Program, and the Planetary Geology and Geophysics Program.