EDP Sciences
Free access
Issue
A&A
Volume 402, Number 1, April IV 2003
Page(s) 373 - 381
Section The Sun
DOI http://dx.doi.org/10.1051/0004-6361:20030027


A&A 402, 373-381 (2003)
DOI: 10.1051/0004-6361:20030027

Mars: Mapping surface units by means of statistical analysis of TES spectra

F. Altieri and G. Bellucci

CNR, Istituto di Fisica dello Spazio Interplanetario 00133 Rome, Italy
(Received 28 November 2001 / Accepted 3 December 2002 )

Abstract
Thermal Emission Spectrometer (TES) data from the Mars Global Surveyor (MGS) mapping phase have been processed to identify regions with unique spectral features for new clues on the Martian surface composition. For this purpose we have developed a procedure to search and map band absorptions related to presence of different surface minerals on a spatial scale of a few kilometers. Data used in this study cover the March 1999-July 2000 period, corresponding to $120^\circ < L _{\rm s} < 250^\circ$ on Nili Fossae, Sinus Meridiani and Valles Marineris regions, where outcrops of olivines and hematite have been identified in previous studies. We have tested the validity of our procedure on these areas and then extended our analysis to other portions of the planet. The data have been assembled in  $10^\circ \times 10^\circ$ emissivity spectra cubes with $16 \times 16$ pixels per square degree. The Principal Components Analysis (PCA) has been used to identify spectra with very low contamination by atmospheric aerosols. It has been applied in two spectral ranges: 300-550 cm -1 (~18.2-33.3  $\mu$m) and 815-1143 cm -1 (~8.7-12.3  $\mu$m). By means of PCA we have selected three principal spectral classes: spectra with a high content of dust, spectra with a high content of water-ice and spectra with a lower contamination by dust and water-ice. To identify emission spectra with interesting features likely related to surface minerals we have selected, in the last class, data with high spectral variance between 300-550 cm -1, the range where the hematite and olivine bands have been found.


Key words: planets and satellites: individual: Mars -- techniques: spectrometric -- methods: data analysis

Offprint request: F. Altieri, francesca.altieri@isfi.rm.cnr.it




© ESO 2003