Verification of Quality and Compatibility for the Newly Calibrated Clementine Nir Data Set

Introduction: In February 1994, the Clementine spacecraft began orbiting the Moon with a suite of instrumentation that generated millions of images of the lunar surface. This instrumentation package included an ultraviolet-visible (UVVIS) imager that collected 5 spectral wavelengths (0.415, 0.75, 0.90, 0.95, and 1.00 μm) and a near-infrared (NIR) imager that collected 6 spectral wavelengths (1.10, 1.25, 1.50, 2.00, 2.60, and 2.78 μm) of information [1]. Over the past decade, calibration of the NIR data set has been troublesome due to largely varying dark frame signals with time and orbit and uncertainties on gain and offset values [2, 3]. Several previous attempts to calibrate the Clementine NIR data set have been made [2, 4-7]. However, these calibrations were not easily automated. Recently, the USGS, Flagstaff released these data with a near final calibration [8]. Here we show a preliminary study to verify the quality and compatibility of this data set using Earth-based telescopic data and previous calibrations of the Clementine NIR and UVVIS data sets [5, 6, 9]. Data: Newly calibrated Clementine NIR data were downloaded from the USGS Astrogeology Research Program website. These data are available in a simple cylindrical projection with a resolution of 500 meters/pixel. Previously calibrated NIR Clementine data of the Aristarchus crater region of the Moon was obtained from [5, 6] for comparison. The data from [5, 6] is characterized by a robust NIR pixel-by-pixel data calibration using Earth-based data. Currently, [5, 6] serves as the current gold standard for automated NIR Clementine data calibration. Earth-based telescopic data taken from Mauna Kea Observatory of the Hadley-Apennine region of the Moon was also obtained for comparison with space borne data [9]. For a summary on how these data sets were calibrated see [5, 6, 9]. Earth-Based Telescopic Data: Earth-based telescopic data of the Moon was implemented for use as a ground truth comparison. These data have high spectral resolution and provide an accurate shape to the continuum. A preliminary comparison was conducted using mean spectra of nine localities in the Hadley-Apennine region. The majority (8 of 12) of mean spectral comparisons shows good similarity and are within respective 1σ standard deviations. A comparison between spectra of the Aristillus central peak shows a goodness of fit of acceptable quality (Figure 1). Spectral comparisons outside of the 1σ standard deviation are speculated to be attributed to imprecise knowledge of the geographic positions available in the Earth-based telescopic data set.