1 Sam Bell The Moon Darby Dyar Haruyama et al. (2008) Haruyama et al. (2008) used the terrain camera (TC) on the SELENE mission to determine whether there was a large region of exposed water ice in Shackleton Crater on the south pole of the Moon. Unfortunately, the authors phrased their conclusions so that they appeared to be saying that they had ruled out the existence of any exposed ice in the crater when what they had actually ruled out was the existence of regions of completely exposed ice larger than several hundred square meters. Furthermore, their discussion of the vital argument—that there were no high-albedo regions and hence no exposed ice—was extremely limited and in no way sufficient to justify their conclusions. Shackleton crater is a roughly 10km diameter crater very near the lunar south pole, and due to the very low axial tilt of the Moon, most of the floor of the crater never receives direct sunlight, although the walls often do. It has been suggested that such permanently shadowed regions may contain deposits of water ice that could be very useful to future colonists. The premise of the experiment outlined in Haruyama et al. (2008) was to observe the crater at maximum solar illumination (an incidence angle of 3˚), when a small amount of light would be reflected off the illuminated wall, and scientists would be able to see if there were any extremely bright patches of exposed water ice. The instrument used was the TC, a 10m resolution stereo camera on the SELENE orbiter. Haruyama et al. (2008) were able to produce a rough image of the permanently shadowed regions of the crater, and it did not show any large exceedingly bright regions. They derived an average albedo of .23, with a maximum albedo of .28. From this they derived the conclusion that the maximum amount of ice within an area of 100m2 was (.28-.22)/(1-.22) divided by the factor of overestimation of the albedo. Although the authors do not give an estimation of this factor, it is probably not very high because they stated that the average albedo of the crater was typical for the region, with .22 being the average albedo of the farside. They do not actually calculate out these numbers, but it is a very simple calculation, and I obtained a value of 8% before correcting for overestimation. So the figure is probably around 7%—and certainly no lower than 4%—assuming that the albedo figures are reasonable (see below). Haruyama et al. (2008) characterized this value as “less than a few percent,” and used this to claim that there was no exposed water ice. Although the conclusion of the paper seemed to accept the possibility of extremely small deposits of ice, the abstract gave the general impression that they had ruled out the existence of any exposed ice, stating that “exposed relatively pure water-ice deposits are not on the crater floor.” The abstract even concluded with the sentence “Water-ice may be disseminated and mixed with soil over a small percentage of the area or may not exist at all.” This ignores the possibility that there is a large area of pure water ice covered with a thin blanket of regolith, a possibility that the data by no means ruled out. Another example of very deceptive wording was describing ~7% as “less that a few percent.” This was especially egregious because they did not actually give the figure of 7% or even the figure of 8% before correcting for the overestimated albedo; they merely provided the calculations for the 8% figure. It is probably fair to say that in general a reader will gloss over the calculations and assume that they mean what the authors say they do. And since most people would probably take “less than a few percent” to mean under 3% or so, this is really a dramatic2 mischaracterization of the results. Finally, there is a real difference between saying that there is no patch of exposed ice larger than 7m2 and saying that there is no patch of exposed ice. Several patches of exposed ice 5m2 in size would be very significant and are still a possibility. The vast majority of this paper was devoted to background (14 column inches), and only a very small fraction was devoted to discussion of the results (4 column inches). They also did give several images of the crater, but all of the images only reflected the total amount of light reflected; they did not provide us with an albedo map, the crucial piece of evidence and something they must have had to claim a maximum albedo. This meant that the results section was severely lacking in several respects: Albedo is not a very well-defined quantity for several reasons. To begin with, it varies dramatically with wavelength, and the authors do not specify what specific range or bands the TC operates over, although they do give the impression it takes mostly from the visual. It also usually varies with the direction of observation, especially when the light comes from a point-source like the Sun. The degree of this variation depends on the luster of the surface in question, but Haruyama et al. (2008) assume that all surfaces involved are Lambert surfaces, which diffuse reflected light evenly in all directions. They do not justify this assumption. This is quite troubling because whether or not the lunar regolith is a Lambert surface is still a subject of debate. Early ground-based observations indicated that the surface was solidly non-Lambertian, but one interpretation of the Clementine data concluded that the surface is actually fairly close to Lambertian.i,ii For ice, the assumption of a Lambert surface is even less justified; the luster of ice is usually vitreous (non-Lambertian).iii This is assumption could easily cause a serious underestimation of the albedo of the ice because the ground is being illuminated at a very low angle of incidence while it is being observed from directly above. Another major issue with the results section was the all-important value of 1.0 for the albedo of ice, which Haruyama et al. (2008) state as if it were a physical constant. Albedo is actually a function of many things, such as the composition, the surface roughness, and the amount of fractures in the crystal. It is not constant for all ice crystals. This number of 1.0 appears nowhere in the text of the cited paper,iv although there is a very hard-to-read graph dealing with particular laboratory samples, not estimates of the real composition of the ice, that appears to show