[Previous] | [Session 62] | [Next]
W.H. Farrand (Space Science Institute), J.F. Bell III (Cornell University), R.V. Morris (NASA Johnson Space Center), M.J. Wolff (Space Science Institute)
Multispectral images of Mars were collected using the HST WFPC2 instrument over a 10 day period around the 1999 opposition. Four sets of multispectral images were collected, each separated by ~90° of longitude. The 12 channel data, collected using narrow or medium bandpass filters, extends from 255 to 1042 nm. Spatial resolution of the data is ~19 km/pixel near the sub-Earth point. In this work, we seek to map spectrally distinct global surface units by utilizing recently developed techniques, used initially with airborne imaging spectrometer data. The techniques were used to successively reduce the dimensionality of the data, to identify spectrally unique classes in the reduced data space, and to map those classes over each of the four image sets. In order to mitigate undesired atmospheric scattering effects, this analysis was conducted on only the 588 to 1042 nm data. Portions of the limb and those pixels most affected by the atmosphere were also masked out, using high water ice cloud opacity determined by spectral mixture analysis using all 12 channels. We see no evidence in these images for either local or regional dust storm activity.
Noteworthy class aspects include: (1) The deepest 953 nm absorption occurs within the north polar sand sea; (2) Several bright region classes were discernable, with differences in the depth of the 953 nm band and in the spectral slopes between 588 to 673 nm and 953 to 1042 nm; (3) North and south Acidalia Planitia are spectrally distinct: the south appears more contaminated by bright materials and has stronger absorption at 835 nm; (4) A dark class southeast of the Valles Marineris was identified that has a strong absorption at 835 nm; and (5) There are distinct intermediate albedo regions within Elysium Planitia with relatively high reflectance at 835 nm.
The author(s) of this abstract have provided an email address for comments about the abstract: firstname.lastname@example.org