2.1. Landing Site Context
    2.2. Science Studies from Descent Imaging

2.1. Landing Site Context

Among the most important questions to be asked about a spacecraft sitting on a planetary surface is "Where is it?" Radiometric tracking and orbit determination (both spacecraft-to-Earth and spacecraft-to-spacecraft) and integration of inertial reference system variations (accelerometers tied to inertial measurement units) provide answers to this question to varying degrees of accuracy, but at best can only tell the position to perhaps a few hundred meters. Surface imaging of features also visible from orbit can be used to pinpoint lander positions to a few tens of meters or better, provided that such features are found. However, if the orbiter image resolution is insufficient to see features visible to the lander, or the local, meter-scale relief is too great (so the lander cannot see very far), or the surface is relatively featureless, or the surface has many features but they all look the same, then the lander cannot be located. The Viking Landers provided good examples of such circumstances. Through a combination of 20 m/pixel, relatively low-sun orbiter photography, excellent radiometric tracking from Earth over a long period of time combined with good inertial position measurements during landing, and fortuitously landing near craters large enough to be seen on the horizon in lander images, VL-1 was located to within 50 m[1]. However, despite good inertial position measurements during landing and good radiometric tracking data both during the descent and for a number of weeks thereafter, the homogeneously rugged local relief, nearly featureless horizon, and the lack of spatially variable landforms in the 40 m/pixel orbiter images defeated all attempts to determine the location of the VL-2 to better than 10 km.

Why is it important to know "exactly" where a lander is located? The principal reason is context. It is necessary to determine if the locale is representative of the region, and, indeed, of the entire planet. It is usually not possible, just from a lander's perspective, to tell the difference between what is visible, and what is just over the horizon. The locale may be anomalous; this must be determined before general interpretations can be made. Knowing that local meteorology is affected by a nearby escarpment, or that the lander sits on ejecta from a nearby crater, is important both for local interpretation, and for extending it farther afield. The context of relating lander observations to those seen from the orbiter is also important. The simplest, and most obvious, example is to place surface imaging into the context of orbiter images (extending and linking crater and boulder size/frequency relationships, extending surface observations of eolian bedform wavelength/amplitude/particle size attributes to larger scale, etc.). Other examples include relating color and/or albedo boundaries seen in orbiter data down to lander scales (which is particularly difficult to do from the surface owing to the extremely oblique viewing geometry of the lander instruments), and providing validation of models used to calculate rock abundance and other granulometric properties of the surface from thermal emission measurements.

Descent imaging can also provide a context for operations after landing. For example, the final few images should cover the area around the lander out to 10 meters or more at spatial scales of a centimeter or better. Such images can be used to plan sampling activities and/or mobility unit traverses, both initially before lander imaging, and complimentary to those data once they are received. The easily interpreted, overhead perspective provides such planning activities considerably greater flexibility, and permits more rapid planning as well. Advanced techniques in computer graphics and data visualization have been used to merge lander images with distance measurements, derived from stereoscopic images or laser rangefinding, in efforts to mimic the overhead perspective. However, the inability to see surfaces hidden from direct view from the lander perspective is an essentially fatal flaw in such efforts. The simplest, most comprehensive way to achieve overhead viewing is from a descent camera.

2.2. Science Studies from Descent Imaging

The scale at which processes modify a planet's surface are dependent on the vigor of the processes and the timescales over which they act. For Mars, the vigor of environmental processes has varied with time: recent phenomena appear to be relatively weak (e.g., wind transport of dust and sand), while ancient processes appear to have been much more vigorous (e.g., channel formation by catastrophic flood), although some processes are exceptions to this general rule (e.g., the occasional contemporary mass movement). Based on cratering relationships (both the number of craters on surfaces and the degree of degradation of the ensemble of craters), a crude relationship between size and age can be formulated: features a few meters across are unlikely to be more than a few millions of years old, while those hundreds of meters across are unlikely to be younger than a few hundred of millions of years old. This relationship suggests that features visible in descent images will cover a range in ages from hundreds of millions of years to a young as a few years in age.

JPG = 105 KBytes
GIF = 493 KBytes
Figure 1: Left: Portion of Landsat image showing Antarctic Dry Valleys at 80 m/pixel. Right: Aerial photograph of area outlined in Landsat image, at 7.5 m/pixel, also indicating location of nested descent images. Note comparison of landforms at this scale ratio of roughly 10:1.

Relationships between scale and time also occur on Earth, and can be used to illustrate the study of temporal relationships with descent images. Figures 2 and Figure 1 shows the relationship between a typical orbiter image (in this case, a portion of a Landsat frame on the left) and the first image of a descent sequence (on the right). At a resolution of about 8 m/pixel and covering an area over 8 km on a side, the location of the descent image is reasonably visible in the 80 m/pixel orbiter data. The first descent image provides both this crucial link to the orbiter observations, and the context for all subsequent frames.

In the specific case of these figures, the increase in resolution from 80 m/pixel to 8 m/pixel spans several important transitions in geomorphic interpretation. The orbiter data show that the area is mountainous, with glaciers moving down various topographic gradients. Details of, for example, the glacial flow cannot be seen at this resolution, but become obvious in the 8 m/pixel data. Note the snout of the Taylor Glacier (lower center, image on right), which shows ablation pitting along medial streamlines of shear and morainal debris. Note, too, details of the mountain walls, including mass movements, and evidence of liquid flow (e.g., small stream channels).

JPG = 115 KBytes
GIF = 597 KBytes
Figure 2: Sequence of images at a nesting scale ratio of 2:1, showing changes in landforms as a function of scale.

Two "serendipitous" observations, relating to liquid water, can be made using these images. First, the dark stains in some valley wall- and floor-channels indicates that liquid water was flowing on the surface in the very recent past. Indeed, given relatively simple calculations, it is possible to show that the moisture is only a few weeks old. Second, the dark band around the ice-covered lake (Lake Bonney, right side, center) can be seen in several of the higher resolution images to be a liquid water moat. Again, relatively simple calculations suggest that such moats are ephemeral.

Figure 3 shows an image taken from about 80 m above the surface, but looking obliquely (the near field is viewed at an emission angle of 30° , and the far field at about 75° ). The advantages of oblique viewing, in particular the ability to gain from a single image some knowledge of subtle relief in the scene and to provide a more familiar view, are clearly evident in this image. Note in the lower left foreground a helicopter landing circle approximately 7 m in diameter.

JPG = 45 KBytes
GIF = 213 KBytes
Figure 3: View of area shown in last two images of Figure 2, showing advantages of oblique viewing.

One of the more interesting observations that can be made from this sequence of images is that the scene content varies dramatically with scale. This is prima facie evidence against the idea that nature is scale-invariant (i.e., that it isn't fractal-like). Many geologists have disagreed with the mathematicians and geophysicists who contend that self-similarity is a fundamental attribute of geology. Geologists contend that the types and style of geologic processes and materials clearly vary with scale (i.e., the mechanisms responsible for breaking individual grains of sand are very different from those responsible for the shape of river valleys), and the sequence of images attests to this view. There are clearly several points in the continuum of scales where the surface takes on distinctly different properties (the last two frames in Figure 2 are very different from the first two frames). To the extent that these surfaces reflect different processes and materials, an analogous sequence on Mars will provide considerable insight into similarities to and differences with terrestrial conditions.

To MARCI Table of Contents

To Previous Section

To Top of This Section

To Next Section

Return to Mars Page

Return to MSSS Home Page