Presentation #213.03 in the session Mars Atmosphere (iPosters).
There in an inherent difficulty in retrieving atmospheric aerosol signal from near-infrared (NIR) spectra that is less of an issue for other wavebands (e.g. thermal infrared and ultraviolet) since, for Mars, it is primarily reflected light (instead of emitted) and surface minerals have active absorption bands (instead of being more uniformly dark). This means that the surface reflectance, for every point and each time period, must be a known input in order to solve to the radiative transfer equation and retrieve cloud optical depth. For the past several years we have been devising and improving a technique to retrieve surface reflectance spectral signatures without any a priori assumptions about the surface itself—our only assumption is that all surfaces can be decomposed into a finite set of constant spectral endmembers. That even if the surface reflectance changes, it would translate to a change in endmember coefficients and not the underlying endmembers themselves.
The technique starts with principal components analysis (PCA) on the spectral data itself, target transformation of a mineralogical spectral library into that PCA space, then N-FINDR and Hyperplane-based Craig Simplex Identification to find spectrally “pure” vertices surrounding the data cloud. These vertices are the spectral endmembers. Radiative transfer is performed to create test models with input parameters being ice and dust column optical depth and endmember coefficients. The model with the smallest least-squares-sense difference from the data spectrum is taken as the result and model optical depths are mapped.
The data used for this study are the NIR spectral data taken by the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) in its multi-spectral mapping mode. The spectra have far fewer points, but attempt to cover most of the planet over several orbits—on average we use about 20 consecutive sols of data to make a spectral imaging map.
We will present our technique and retrieved ice and dust cloud optical depth maps. These will be compared to previous studies to show the viability of the technique.