Studies of Jovian Atmospheric Structure and Coloring Agents using Hyperspectral Imaging

Paul Strycker

Jupiter's atmosphere presents us with a remarkably diverse scene of clouds. They exhibit contrast in brightness, appear in a range of white-to-red colors, and are organized into well-defined bands that encircle the planet and storms that are larger than Earth. The coloring agents, or chromophores, that are embedded within Jupiter's vertical aerosol structure have not been identified and are poorly characterized. In this dissertation, we present two studies of chromophores in the context of the jovian atmospheric structure. In the first study, we analyzed multispectral images acquired with the Wide Field Planetary Camera~2 on-board the Hubble Space Telescope. We employed a radiative transfer code to retrieve single scattering albedo spectra, SSA, for particles in Jupiter's tropospheric haze at seven wavelengths in the near-UV and visible regimes. The data consisted of images of the 2008 passage of Oval BA to the south of the Great Red Spot. We derived SSA for locations that were selected from 14 weather regions, which spanned a large range of observed colors. All SSA curves were absorbing in the blue, and SSA increased monotonically to approximately unity as wavelength increased. We found accurate fits to all SSA curves using an empirically derived functional form: SSA = 1 - A*exp(-B*lambda). The best-fit parameters for the mean SSA curve were A=25.4 and B=0.0149 for lambda in units of nm. We performed a principal component analysis (PCA) on our SSA results and found that one or two independent chromophores were sufficient to produce the variations in SSA. A PCA of reflectance spectra, I/F, for the same jovian locations resulted in principal components (PCs) with roughly the same variances as the SSA PCA, but they did not result in a one-to-one mapping of PC amplitudes between the SSA PCA and I/F PCA. We suggest that statistical analyses performed on I/F image cubes have limited applicability to the characterization of chromophores in the jovian atmosphere due to the sensitivity of I/F to horizontal variations in the vertical aerosol distribution. In the second study, we collected and analyzed hyperspectral images of Jupiter from 470--900 nm in 2-nm increments. We acquired our data with the New Mexico State University Acousto-optic Imaging Camera (NAIC) using the Astrophysical Research Consortium 3.5-meter telescope at Apache Point Observatory. We characterized the photometric response of the NAIC instrument and developed an automated data reduction procedure for NAIC data. We retrieved SSA for four jovian locations and confirmed the applicability of the functional form found in the HST data analysis. We performed statistical analyses---PCA, non-negative matrix factorization, and vector quantization---on our I/F hyperspectral image cubes. These results are also consistent with one, possibly two, chromophores in the jovian atmosphere.