 AN ANALYSIS OF MARTIAN DUST ACROSS THE LENGTH OF THE PHOENIX MISSION USING TRIAXIAL ELLIPSOIDS.  E. L. Mason and M. T. Lemmon, Texas A&M University, Department of Atmospheric Sciences   Introduction: The Phoenix Lander mission extended across a time when frontal dust storms caused by the recession of the polar cap were extensive. Optical depth values ranged from 0.1 to 0.9, generally decreasing across the length of the mission. The lander's Surface Stereo Imager performed several cross-sky brightness surveys to constrain the size distribution and scattering and absorption properties of the airborne dust in the Martian northern polar environment. More recently, new models have been applied to the Phoenix data to explore different shapes that can adequately model light scattering by the dust.  Martian atmospheric dust is not spherical and likely contains irregularly shaped particles. This irregularity adds complexity to models that determine radiative heating of the atmosphere. Studies in the past have been successful in constraining particle size. More recent work[1] has corrected particle sizes down from 1.6 µm[8] and 1.5 µm[4] to 1.4 µm and 1.3 µm, respectively, which is consistent with returned results shown in Figure 3. Shape, however, has not been well determined. Optical properties of the dust have been modeled using many different shapes, including oblate and prolate spheroids and cylinders.  More recently, triaxial ellipsoids have shown to be a good fit to the observed data for I/F values as well as polarization[6], which is an improvement over other shapes. Work using terrestrial observations[10] shows that an assortment of triaxial ellipsoids provide a good analog for the scattering properties of dust aerosols. In addition, a database[5] has been developed containing single-scattering properties of irregularly shaped dust particles with pre-defined microphysical and optical parameters. The tabulation allows for quick and efficient use of the results and can be applied to the Martian atmosphere. Using the database, single scattering properties adapted for the Martian atmosphere can be applied to a radiative transfer model to help determine bulk scattering properties of the medium at the Phoenix landing site. In addition, polarization of dust could play a unique role in scattering radiation and therefore in heating the atmosphere. This heating is important for determining weather on Mars. Model: The model uses the DISORT[7] radiative transfer code to determine sky radiance with a triaxial ellipsoid addition to constrain dust sizes and bulk scattering properties. The model is an update from previous work[8] and uses 2 parameters instead of 5 parameters to fit modeled I/F to the observations. The model varies randomly a distribution of triaxial ellipsoids defined by three geometric lengths and provides scattering properties using the tabulated database for terrestrial dust[5]. It fits a standard phase function to the observed data through multiple iterations using a Levenberg-Marquardt least-squares inversion. The phase function as well as the effective particle size and variance are returned. Results: Fits to the single scattering phase function shown in Figure 1 have produced a well-constrained forward scattering peak for seven different data sets across the Phoenix mission (solid). These data sets are compared to previous work[4][8] for Sol 081 (dashed).  Figure 1: Fits to single-scattering phase function showing seven data sets from the mission. The first seven sets (solid color lines) are compared to a phase function from Sol 081 modeled according to reference [8].   Figure 2: Polarization plots comparing different shape models applied to the Phoenix cross-sky survey on Sol 081. Green dots are the observed values and blue dots are the modeled triaxial ellipsoid values.  Polarization results for Sol 081 show good agreement between triaxial ellipsoids and observations as compared to other shapes. As show in Figure 2, the polarization is weak in magnitude and not specifically Rayleigh-like in shape. Data show a negative branch around 150° (not shown), and it is distinctly non-Mielike in complexity. In addition, the first two moments for a normalized gamma distribution were returned. The effective particle size, a, is shown in Figure 3 for the seven data sets. Values did not vary significantly between the sets, and the results were consistent with previous work [1].     Figure 3: This plot represents the effective particle size, a, in microns returned from initial runs using the triaxial ellipsoid model discussed above. A 10% error bar is shown. Sizes are plotted for each of the seven data sets represented in Figure1.  Future Work: Initial results using triaxial ellipsoids show promise in terms of producing a good fit to the observed data. We will continue to apply this model to other data sets across the Phoenix mission to determine the extent to which triaxial ellipsoids work, and explore how this shape is important to radiative heating.   Short Summary: Irregular dust How I wish you were a sphere Ellipsoids work, though  References: [1] Wolff M. J. and Clancy R. T. (2003) JGR, 108, 5097. [2] Clancy R. T. et al. (2003) JGR, 108, 5098. [3] Johnson et al. (2003) Icaurus, 163, 330. [4] Lemmon M. T. et al. (2004) Science, 306, 1753.  [5] Meng et al. (2010) Aerosol Sci., 41, 501-512.        [6] Lemmon M. T. et al. (2014) DPS presentation.     [7] Stamnes et al. (1988) Appl. Opt., 27, 2502.          [8] Tomasko et al. (1999) JGR, 104, 8987. [9] Mason E. L. et al. (2014) AGU poster presentation. [10] Bi L. et al. (2009) Appl. Opt., 48, 114-126.  
