Statistical Variability of Chemcam Libs Spectra from Gale Crater Exploration

4 , the ChemCam team and the MSL team. Introduction: The huge spectra collection that ChemCam instrument [Wiens, 2012] has acquired since Curiosity landed at Gale crater contains a wealth of informations that are hard to decipher in details because of the rather high number of phenomena that infuence these spectra. However, the reproducibility of these spectra in similar conditions looks quite good. This work is therefore aimed at identifying situations of good reproducibility, permitting afterwards to defne quantitatively the level of instrumental reproducibility using the thus selected spectra subsets. We start at the lowest level of the measurement production. We will frst focuss on a few typical measurement sequences for which the targetted material is a priori relatively homogeneous. Among these potential targets are the ChemCam calibration targets, the drill tailings from the different MSL/Curiosity drill operations, and some identifed fne-grain soils. On these geological targets, a ChemCam measurement sequence consists usually in a small series of laser shooting locations, and on each location, in a series of laser shots, typically 30, for which each individual plasma has been analysed by the spectroscopic part of the instrument, three spectrometers eahc with 2048 channels giving three wavelength ranges each with its own spectral resolution. Available on the Planetary Data System, each acquired spectrum is presented with different levels of treatment [Maurice, 2012]. EDR fles give the raw values, as measured, which are unnormalized intensities versus channel numbers. An automated processing chain permits to denoise, to remove the so-called continuum and fnally to realign the spectra. The denoise operation permits to improve the statistical aspect the spectra. The continuum-removing operation aims at separating the continuum part (black body radiation and other fner physical phenomena) from the fuorescence part which consists in the series of all the individual spectroscopic lines corresponding to atomic or eventually molecular electronic transitions in the cooling plasma. The third operation consists in realigning these spectroscopic lines with a reference, and therefore calibrating the abscissa from channel numbers into wavelengths. This is necessary since the ChemCam spectrometers are sensitive to Curiosity internal temperature, which varies signifcantly. The resulting spectra outing the low-level chain of treatment are named RDR ones. The last low-level treatment consists in correcting for differential optical instrumental "transparency" and sensitivity, relatively to wavelengths by applying a so-called optical transfer function. The resulting spectra are named "CCS" ones. In order to assess the variability of the …