Descent Imaging Science

Figure 1a: Satellite view of Antarctic Ice-free valleys at 80 m
per pixel (260 ft/pixel). Box shows location of Figure 1b.

Why do some places look boring and other places look exciting?

Primarily, the way a surface looks depends on the strength of the materials that make up that surface, the strength of the processes that are eroding it, and how long these processes have been operating at that location. On Mars, the strength of processes appears to have varied substantially from place to place and over time. Processes acting today (for example, wind transport of dust and sand) appear to be relatively weak, while ancient processes appear to have been much more dramatic (for example, the catastrophic flood that affected the Pathfinder landing site). However, some areas, like layered materials within Candor Chasma, are largely devoid of impact craters, implying an erosional process of substantial capability for which there is little other evidence. The true nature of martian erosional processes is just now being deciphered with the help of high resolution images from Mars Global Surveyor.

Descent imaging provides a bridge between orbiter pictures, that tell us about regional and global scales, and lander images of very small, "micro-scale" attributes of the planet. The pictures above, to the right, and below illustrate this change in perspective.

Figure 1 shows the relationship between an orbiter image typical of Viking coverage of Mars (in this case, Figure 1a, above, shows an 80 X 80 km portion of a Landsat frame at 80 m/pixel) and the first image of a descent sequence (Figure 1b, to the right, a portion of an aerial photograph). At a resolution of 8 m/pixel and covering an area about 8 km on a side, the location of the descent image is reasonably visible in the orbiter data. Images taken early in a descent provide both this crucial link to orbiter observations, and the context for all subsequent frames. Of course, with Mars Global Surveyor Orbiter Camera (MOC) images that are even better than 8 m/pixel, we are already making the connection. However, the beauty of descent imaging is that it gets closer and closer and closer. At some altitude it is becomes better than MOC, and then keeps getting better.

Figure 1b: Aerial photograph of Taylor Valley, Antarctica, at
8 m/pixel (26 ft/pixel). Boxes show subsequent figures.

In the specific case of the images in these illustrations, the increase in resolution from 80 m/pixel to 8 m/pixel spans several important transitions in the types of features seen. It is clear in the orbiter image, for example, that the area is mountainous and has glaciers moving down various valleys. Details of the glaciers cannot be seen at orbiter resolution, but become obvious in the 8 m/pixel data. Note the snout of the Taylor Glacier (lower center, Figure 1b), which shows ablation pitting along streamlines of shear and morainal debris that run parallel to the long axis of the glacier. Note, too, details of the mountain walls, including landslides and debris slopes, and the evidence of the flow of water (e.g., small stream channels).

Figure 2: Sequence of progressively higher resolution images of the Upper Taylor Glacier,
Lake Bonney, and Bonney Riegel terminal moraine. Figure 1b shows the locations of these
images in a lower resolution view.

Figure 2 (above) shows a sequence of six images at successively higher resolutions of 5, 2.5, 1.2, 0.6, 0.3, and 0.16 m/pixel, equivalent to images taken roughly 65, 40, 26, 18, 13, and 9 seconds prior to landing. As resolution improves the ground changes dramatically. At low resolution the pictures show great diversity of features, while at high resolution the scene looks much simpler (some might even say boring!). However, this trend should not be interpreted to mean that the high resolution images are uninteresting. Rather, the type of things that one can tell from such images is different from the types of things one can tell from lower resolution data. Boulders and sand, as an example, can just be discriminated at about 1.2 m/pixel, but become much more prominent at resolutions of 0.3 m/pixel and better. The number of boulders at different sizes, and there distribution across the surface, tells us about the processes that formed the surface, as well as the processes that have modified the surface and transported the sand and rocks to the places we see them. In this example, the boulder population indicates, independent of any other association, that these materials were eroded and transported by glaciers. The distribution indicates both the direction of flow of the glacier and how much erosion was occuring at the base of the glacier. In this case, the motion was from left to right (shown by the location of the boulder field), and the glacier was wet-based (shown by the size of debris in the ground moraine on which the boulders sit). Since the glaciers in the present environment are dry-based, the debris is remnant from an earlier period.

Figure 3: Oblique view of Bonney Riegel
moraine, with the shore of Lake Bonney in
the upper right corner. Image is taken from
a helicopter about 80 m (262 ft) above the
ground. The white arrows indicate the
location of a 7 m (23 ft) diameter circle
of stones used to mark a safe landing
location for helicopters.

Two "serendipitous" observations can also be made from these images, relating to liquid water. First, the dark stains in some valley walls and floor channels indicate that liquid water was flowing on the surface in the very recent past. Indeed, relatively simple calculations that take into account the seasonal heating and high evaporation rates in the Antarctic ice-free valleys show that the moisture is only a few weeks old. Second, the dark band around the ice-covered lake can be seen in several of the higher resolution images to be a liquid water moat (indicated by ripples and the reflecting properties of the surface). Simple calculations suggest that such moats are very short-lived.

Figure 3 (left) shows an image taken from about 80 m (162 ft) above the surface, but looking obliquely. The bottom edge of the picture is about 30° from straight down; the top of the picture is about 15° from the horizontal. 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 center foreground a helicopter landing pad made by placing rocks in a circle approximately 7 m (23 ft) in diameter (white arrows). The field-of-view of the Mars Descent Imager covers the ground from directly below the spacecraft to about 30° from the horizon.

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 good evidence against the idea that in nature features seen at one scale are similar to those seen at other scale (that is, that they show "fractal" behavior). Some scientists have argued, through various examples, that "self-similarity" is a fundamental attribute of geology. Others contend that the types and style of geologic processes and materials clearly vary with scale (for example, that the mechanisms responsible for breaking individual grains of sand are very different from those responsible for the shape of river valleys). The sequence of images shown here attests to this latter 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.

©2005 Malin Space Science Systems, Inc.