Robert A. West

The Cassini Imaging Science Subsystem returned the highest-resolution images of Saturn and its moons ever seen but could have produced substantially more had it not been for an overly lossy on-board compressor. A retroactive study later found that, had the compression algorithms been customizable and updatable during flight, data could have been compressed 3 to 30 times more than in the actual mission, without compromising the scientific goals. In this paper, we develop a machine learning framework to automate the selection of a compression algorithm catered to a given data set. We demonstrate how off-the-shelf compressors can be supplemented by previously-developed neural network compressors to effectively compress low-spatial-entropy data, using Cassini images of the atmospheric haze density of Titan as an example. For images with higher spatial entropy, we use neural networks to model and generalize wavelet-based transform coding and train them to minimize the image reconstruction error and the compressibility of the transformed coefficients. We demonstrate our compressor’s ability to preserve details of Cassini images of Tethys used in tiepoint matching when compressing to 2.6 bits per pixel (down from 4.36 bits per pixel using lossless JPEG compression), and to resolve subtle propeller-shaped formations in Cassini images of Saturn’s rings better than db5-wavelet transform coding when compressing below 1.75 bits per pixel. We then combine our network with a known optical flow-estimating network to more aggressively compress entire flyby image sequences of Tethys. Finally, we show how to train our compressor adversarially with a network that discriminates compression error from noise, allowing the compressor to learn how to compress images with more complicated modeled noise distributions.

The global energy budget is pivotal to understanding planetary evolution and climate behaviors. Assessing the energy budget of giant planets, particularly those with large seasonal cycles, however, remains a challenge without long-term observations. Evolution models of Saturn cannot explain its estimated Bond albedo and internal heat flux, mainly because previous estimates were based on limited observations. Here, we analyze the long-term observations recorded by the Cassini spacecraft and find notably higher Bond albedo (0.41 ± 0.02) and internal heat flux (2.84 ± 0.20 Wm⁻²) values than previous estimates. Furthermore, Saturn’s global energy budget is not in a steady state and exhibits significant dynamical imbalances. The global radiant energy deficit at the top of the atmosphere, indicative of the planetary cooling of Saturn, reveals remarkable seasonal fluctuations with a magnitude of 16.0 ± 4.2%. Further analysis of the energy budget of the upper atmosphere including the internal heat suggests seasonal energy imbalances at both global and hemispheric scales, contributing to the development of giant convective storms on Saturn. Similar seasonal variabilities of planetary cooling and energy imbalance exist in other giant planets within and beyond the Solar System, a prospect currently overlooked in existing evolutional and atmospheric models.

A planet’s Lyα emission is sensitive to its thermospheric structure. Here we report joint Hubble Space Telescope and Cassini cross-calibration observations of the Saturn Lyα emission made 2 weeks before the Cassini grand finale. To investigate the long-term Saturn Lyα airglow observed by different ultraviolet instruments, we cross-correlate their calibration, finding that while the official Cassini/UVIS sensitivity should be lowered by ∼75%, the Voyager 1/UVS sensitivities should be enhanced by ∼20% at the Lyα channels. This comparison also allowed us to discover a permanent feature of the Saturn disk Lyα brightness that appears at all longitudes as a brightness excess (Lyα bulge) of ∼30% (∼12σ) extending over the latitude range ∼5°–35° N compared to the regions at equator and ∼60° N. This feature is confirmed by three distinct instruments between 1980 and 2017 in the Saturn north hemisphere. To analyze the Lyα observations, we use a radiation transfer model of resonant scattering of solar and interplanetary Lyα photons and a latitude-dependent photochemistry model of the upper atmosphere constrained by occultation and remote-sensing observations. For each latitude, we show that the Lyα observations are sensitive to the temperature profile in the upper stratosphere and lower thermosphere, thus providing useful information in a region of the atmosphere that is difficult to probe by other means. In the Saturn Lyα bulge region, at latitudes between ∼5° and ∼35°, the observed brightening and line broadening support seasonal effects, variation of the temperature vertical profile, and potential superthermal atoms that require confirmation.

In our solar system, the densely cloud-covered atmosphere of Venus stands out as an example of how polarimetry can be used to gain information on cloud composition and particle mean radius. With current interest running high on discovering and characterizing extrasolar planets in the habitable zone where water exists in the liquid state, making use of spectropolarimetric measurements of directly imaged exoplanets could provide key information unobtainable through other means. In principle, spectropolarimetric measurements can determine if acidity causes water activities in the clouds to be too low for life. To this end, we show that a spectropolarimeter measurement over the range 400–1000 nm would need to resolve linear polarization to a precision of about 1% or better for reflected starlight from an optically thick cloud-enshrouded exoplanet. We assess the likelihood of achieving this goal by simulating measurements from a notional spectropolarimeter as part of a starshade configuration for a large space telescope (a HabEx design, but for a 6 m diameter primary mirror). Our simulations include consideration of noise from a variety of sources. We provide guidance on limits that would need to be levied on instrumental polarization to address the science issues we discuss. For photon-limited noise, integration times would need to be of order 1 hr for a large radius (10 Earth radii) planet to more than 100 hr for smaller exoplanets depending on the star–planet separation, planet radius, phase angle, and desired uncertainty. We discuss implications for surface chemistry and habitability.

In our solar system, the densely cloud-covered atmosphere of Venus stands out as an example of how polarimetry can be used to gain information on cloud composition and particle mean radius. With current interest running high on discovering and characterizing extrasolar planets in the habitable zone where water exists in the liquid state, making use of spectropolarimetric measurements of directly-imaged exoplanets could provide key information unobtainable through other means. In principle, spectropolarimetric measurements can determine if acidity causes water activities in the clouds to be too low for life. To this end, we show that a spectropolarimeter measurement over the range 400 nm - 1000 nm would need to resolve linear polarization to a precision of about 1% or better for reflected starlight from an optically thick cloud-enshrouded exoplanet. We assess the likelihood of achieving this goal by simulating measurements from a notional spectropolarimeter as part of a starshade configuration for a large space telescope (a HabEx design, but for a 6 m diameter primary mirror). Our simulations include consideration of noise from a variety of sources. We provide guidance on limits that would need to be levied on instrumental polarization to address the science issues we discuss. For photon-limited noise, integration times would need to be of order one hour for a large radius (10 Earth radii) planet to more than 100 hours for smaller exoplanets depending on the star-planet separation, planet radius, phase angle and desired uncertainty. We discuss implications for surface chemistry and habitability.

Abbreviations was published in Later Greek Epic and the Latin Literary Tradition on page VII.

Radiant energies of planets and moons are of wide interest in the fields of geoscience and planetary science. Based on long‐term multiinstrument observations from the Cassini spacecraft, we provide here the first observational study of Titan's global radiant energy budget and its seasonal variations. Our results show that Titan's radiant energy budget is not balanced over the Cassini era (2004–2017) with the absorbed solar energy (1.208 ± 0.008) × 10²³ J larger than the emitted thermal energy (1.174 ± 0.005) × 10²³ J. The energy imbalance is 2.9 ± 0.8% of the emitted thermal energy. Titan's global radiant energy budget is not balanced either at the timescales of Earth's years and Titan's seasons. In particular, the energy imbalance can be beyond 10% of the emitted thermal energy at the timescale of an Earth year. The energy imbalance revealed in this study has important impacts on Titan, which should be examined further by theories and models.

The formation of hazes at microbar pressures has been explored by theoretical models of exoplanet atmospheres to explain Rayleigh scattering and/or featureless transmission spectra; however observational evidence of aerosols in the low-pressure formation environments has proved elusive. Here, we show direct evidence of aerosols existing at ∼1 microbar pressures in the atmosphere of the warm sub-Saturn WASP-69b using observations taken with the Space Telescope Imaging Spectrograph and Wide Field Camera 3 instruments on the Hubble Space Telescope. The transmission spectrum shows a wavelength-dependent slope induced by aerosol scattering that covers 11 scale heights of spectral modulation. Drawing on the extensive studies of haze in our solar system, we model the transmission spectrum based on a scaled version of Jupiter’s haze-density profile to show that the WASP-69b transmission spectrum can be produced by scattering from an approximately constant density of particles extending throughout the atmospheric column from 40 millibar to microbar pressures. These results are consistent with theoretical expectations based on microphysics of the aerosol particles that have suggested haze can exist at microbar pressures in exoplanet atmospheres.