A summary of the report is available at Dr. Stanley McDaniel's web site here.
Lan Fleming
Engineer Johnson Space Flight Center
Viking Frames 35A72 and 70A13 both show features resembling an eye and mouth on the western side of the "Face" in Cydonia. Since the Mars Global Surveyor returned an image of the "Face" on April 5, 1998, it has been suggested frequently that this new image showed conclusively that the "eye" and "mouth" features in the Viking images were shadows caused by a chance combination of solar illumination angle in those particular Viking images and the placement of ridges on the Face mesa. To support this conclusion the observation has been offered that the MGS image shows ridges not visible in the lower-resolution Viking images that could, in fact, cast shadows over the areas where the "mouth" and "eye" features are observed.
However, consideration of the lighting geometries in the three images show that this claim cannot explain the presence of these features. These considerations are discussed here.
While the "Face" is very similar in appearance in the two Viking images and the solar azimuth was similar in both, the elevation of the sun was radically different. The incidence angle was 10 degrees in Viking Frame 35A72 while it was 27 degrees in 70A13. The shadow of an object cast on a horizontal surfaces with the 10 degree incidence angle of Frame 35A72 is 2.9 times longer than the shadow of the same object at the 27 degree incidence angle of Frame 70A13. Yet the width of both the "eye" and "mouth" features in the direction of the mesa's long axis in 35A72 are no more than 1.3 times greater than the widths of the same features in 70A13 as measured on comparable brightness contour traces of the two images in which the boundaries of these features appear as black lines. These contour tracings are shown if Figure 1.
Figure 1 Contour Traces of the western "eye and "mouth" features in Viking Frames 35A72 and 70A13. Azimuth direction of solar illumination is indicated by the arrows.
Both of these contour traces were generated by increasing the bounding brightness value until the contour trace of the "eye" area begins to merge with smaller nearby bounded regions.
The apparent lack of size change with incidence angle that would be expected of shadows cast by ridges cannot be dismissed as the result of the low resolution of the Viking images; the "eye" in 70A13 is approximately 6 pixels wide in the direction of the sun and the "mouth" 4.
Nor can the possibility that the postulated ridge shadows are being cast upon a sunward-facing slope rather than a horizontal plane account for the lack of size difference of the features between the images. The ratio of shadow lengths for the incidence angles of the two Viking frames reaches a minimum of 1.84 at a slope angle of approximately 30 degrees. The difference in the width of the "eye" feature between the two frames should be a clearly visible 5 pixels at this ratio.
It must also be considered that the differences in the sun azimuth between the two Viking frames might diminish the differences in the lengths of shadows cast by ridges responsible for the "eye" and "mouth" features. The long axis of the "Face" mesa is approximately 20 degrees west of north, so any ridges whose shadows were responsible for the eye and mouth features would run east to west. A more westerly position of the sun in Frame 35A72 relative to 70A13 would cause shadows of east-west ridges to be shorter than the maximum possible length for the incidence angle. Such a difference in azimuth would cause the lengths of shadows of ridges with the necessary orientations to be more similar in the two images, despite the generally longer shadows in 35A72.
The photograph support data, however, indicate that the situation is the exact opposite. The solar azimuth was 66 degrees west of north in Frame 35A72 and 84 degrees west of north in 70A13. The actual difference in sun azimuth between the two images therefore tends to increase the ratio of the shadow lengths, not decrease it.
Figure 2 shows the MGS image of the "Face." Features corresponding to the "eye" and "mouth" in the Viking images are outlined in black on the version to the right.
Figure 2 The "Face" from the first MGS Cydonia image. "Eye" and "mouth" areas outlined on the left.
A roughly almond-shaped dark patch is visible in the MGS image at the position where the "eye" appears in the two Viking frames. While this feature is rather indistinct, it seems very likely that it is at least part of the "eye" feature in the Viking images. The "brow" ridge to the north of this patch has been proposed as the feature casting the shadow giving the impression of an eye in the Viking images. However, the "eye" feature in the MGS image cannot possibly be a shadow cast by this ridge to the north because the sun is to the south. Again, this argues for the presence of a depression at the position of the "eye."
The MGS image shows more detail of the "mouth" area than of the "eye." Three ridges can be seen in this area. The upper-most of these ridges very likely was the upper edge of the "mouth" in the Viking images, but the lower two ridges were not visible. These lower ridges appear to mark the boundaries of a depression. The "mouth" area, as shown in the MGS image is certainly more complex than suggested by the Viking images, but it does display one interesting indication of symmetry not seen in the Viking images: the "hare-lip" feature formed by the lowest of the ridges, which comes to a point along the vertical center-line of the mesa.
It seems almost certain that the appearance of a mouth and eye on the "Face" mesa in the Viking images was created by either depressions or enclosures of significant depth relative to their width in the direction of the sun and not by fortuitous ridge shadows. In the MGS image, there are clear indications of such a depression in the "mouth" area, but the image only hints at the existence of the "eye" depression rather than showing it clearly and in greater detail than do the lower-resolution Viking images. For the reasons stated, it does not appear justifiable to infer that the depression does not exist. Instead it is concluded that images with a more orthogonal view and greater dynamic than that of the existing MGS image of the mesa will be required in order to gain any significant new information about this and other features of the mesa.