The definition of map units will likely be based almost entirely on the Magellan SAR images. However, some surficial materials, such as crater ejecta, are better delineated within emissivity images than by their radar backscatter characteristics, and the use of these data for characterizing such materials is important. At least five types of surficial deposits can be suggested for possible mapping:
1. Low-dielectric (high emissivity) units associated with impact craters. These units appear to be deposits of fine-grained soil that may be several tens of centimeters deep. A type example is the ejecta from Annia Faustina, which blankets rough lava flows from Gula Mons [Campbell et al., 1992]. Evidence of mantling relations is sometimes vague, but careful mapping of such units may provide important clues to regional stratigraphy.
2. High-dielectric (low emissivity) units associated with impact craters. These units are typically seen as the parabolas, which have little or no SAR signature but a dramatic low emissivity relative to the plains, and in some crater floors. The parabolas likely are thin deposits of fine-grained material with a concentration of high-dielectric mineral or metal grains (perhaps produced by impact melting and recondensation). The high-dielectric crater floors are probably impact melt sheets.
3. Low-dielectric areas associated with volcanism. Type examples can be found on the north flanks of Maat Mons, where radar-dark material appears to mantle rough lava flows. Such deposits may have been produced by eruptions of pyroclastics or ignimbrites.
4. Low-dielectric areas in the plains and some upland areas. These areas are found in a variety of settings, some of which suggest that the emissivity anomaly is due to soil formation or accumulation of fine (impact?) material. In other areas the radar return is actually higher than the neighboring lower emissivity plains. Mapping of these features is needed to define their occurrences and probable modes of formation.
5. Surfaces in the highlands with very high dielectric constants. Though some lava flows and other features have dielectric constants as high as 9, the highlands surfaces (above ~6,053 km radius) may reach values of 50 or more. The mechanism by which this high dielectric constant is produced remains controversial [Pettengill et al., 1992; Shepard et al., 1994; Campbell, 1994; Brackett et al., 1995]. Given the range of behaviors and the obvious elevation dependence of the dielectric constant, these areas should be mapped as separate rock units only when boundaries are apparent. High-pass filtering of the radar images has been shown to be a useful means for defining some rock or structural units in the high-dielectric areas.