INTRODUCTION
Results presented here are from Mare Moscoviense (26 degrees north and 147 degrees east, diameter 221 km). The processed data are from three out of the five bands acquired through the UltraViolet-VISible (UVVIS) camera.
DISCUSSION
Grayscale images represent brightness measured in two dimensions over a surface. These images act as a first-order discriminant of surface composition based on overall albedo or brightness. For example, on the Moon, the feldsic lunar highlands are bright relative to the mafic basalts deposits (mare) which are dark in Fig. 1. The boundaries between the mare and highlands units are easily delimited. In contrast, multispectral images record brightness in three dimensions, the X and Y directions are the same as before, and the third dimension is band depth. Reflectance is controlled by composition of the rock type and surface maturity. Different compositions have a characteristic pattern of reflection and absorption over the UVVIS spectrum. Therefore, by measuring surface brightness over a variety of discrete wavelengths, it is possible to identify the presence, absence, and access qualitative composition of surface units. For example, iron preferentially absorbs light around 1000 nm, whereas titanium preferentially absorbs light near 415 nm. If an absorption of light at either of these wavelengths is measured, 415 or 1000 nm, then it is inferred that the element with that characteristic absorption is present. Furthermore, the strength of the absorption is proportional to the amount of that particular element.
Three bands (415, 750, and 1000 nm) from the Clementine data have been used to create false-color image Figure 2. Images from the three bands are ratioed (750/415, 750/1000, and 415/750) and control the red, green, and blue components of the color composite image respectively. The image ratios reduce brightness variations due to albedo and topography, and the resulting image contains brightness variations due primarily to changes in lunar soil composition and maturity. The red channel represents areas that are low in titanium, or high in glass content, the green channel is sensitive to the amount of iron in the surface, and the blue channel reflects the surfaces with high titanium or bright slopes and albedos that are not compensated by using the image ratios. Lunar highlands appear red because they have accumulated glassy agglutinates produced during the bombardment of micrometeorites (maturation). Also red in the false-color image are pyroclastic deposits because of their naturally high-glass content. The yellow-green area in the mare is the combined effect of concentration of mafic minerals (green) and the glass in the soil produced by maturation (red). The blue unit in the mare is relatively higher in titanium compared to the mare unit to its immediate north.
In order to extract quantitative abundance of titanium, groundtruth sites must be used to calibrate the sensor. Titanium composition of soils returned from Apollo missions 11, 12, 14, 16 and the Soviet's Luna 16 and 24 have been compared to the UV/VIS (415/750 nm) reflectance ratio from Clementine. Return samples from each of these missions have had their chemical abundance precisely measured in the lab. A plot of element weight percent versus 415/750 ratio reflectance value displays the correlation between the two components Figure 3. This empirically derived titanium equation based on the Clementine data is only applicable for mature mare surfaces, so the input from the 415- and 750-nm filters had to be trimmed of the highlands areas before they were used in the calculations. Any reflectance value above an empirically determined limit is set to null in the 750-nm image and corresponding pixel in the 415-nm image. The output grayscale image from the equation was then binned into intervals that represent titanium weight percent from 0 through 2.5 and a look-up table designed to convey the data at 0.5 percent intervals [blue 0-0.5; cyan 0.6-1; green 1.1-1.5; yellow 1.6-2; and red 2.1-2.5]. The resulting pseudo-color image was then overlaid on top of the 750-nm grayscale image so the reader could observe the spatial relation of the two mare units two each other and the surrounding highlands, Fig. 4.
The colordrape image provides the perspective of where the eruptive centers for each of the mare units are located. The low-titanium units have erupted around most of the outer edge of the basin and have flowed downhill to the basin center. On the other hand, the high-titanium units emanated from only the eastern side of the basin. The vents for the high-titanium unit are also stratigraphically higher, which supports the model that as the basin is loaded with volcanics, compression increases in the center of the basin and extension increases at the edges.
CONCLUSIONS
Borrowing information from each of the datasets, I have compiled a geological map of Mare Moscoviense and the area surrounding it, Fig. 5. Combining the information from all four datasets supplies a large amount of geological information that aids in accurately interpreting the complex processes that have produced the features one is studying. The grayscale mosaic provides contact relations between highlands and mare. Also apparent are structures and bright-rayed craters which are helpful for deciphering the color-composite image. The color-composite image reveals the surface maturity and composition. What appears as a single unit in the grayscale image is separated into two different chemical units based on multispectral reflectance. The colordrape yields where the lavas erupted from and which direction they flowed. Last but not least, the titanium image produces absolute numbers that we can use for direct comparison with other units and use to calculate abundance of titanium-oxide mineral phases present.