clouds and haze in saturn’s troposphere · clouds and haze in saturn’s troposphere cécile...

Clouds and Hazein Saturn’s Troposphere

Cécile MerletTrinity Term 2011

Lady Margaret Hall

Atmospheric, Oceanic and Planetary PhysicsDepartment of PhysicsUniversity of Oxford

Word count =12 824Page count = 63

Contents

1 Introduction 1

2 VIMS Near-IR Spectrum of Saturn 62.1 VIMS data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2 Thermal Spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.3 Reflection Spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82.4 Scattering Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122.5 Aerosol-free Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

3 Tropospheric Clouds 173.1 Cloud Vertical Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

3.1.1 Fitting strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183.1.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3.2 Limb-darkening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223.2.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223.2.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

4 Tropospheric Haze 324.1 Sensitivity to Model Parameters . . . . . . . . . . . . . . . . . . . . . . . . 334.2 Optical and Physical Haze Properties . . . . . . . . . . . . . . . . . . . . . 344.3 Vertical Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

4.3.1 Sensitivity Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414.3.2 Opacity clearing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

5 Conclusion and Future Work 43

Bibliography 46

ii

List of Figures

1.1 Historical Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2 The Cassini Cruise Trajectory to Saturn. . . . . . . . . . . . . . . . . . . . 31.3 Both Sides View of the Cassini Spacecraft. . . . . . . . . . . . . . . . . . . 41.4 VIMS-VIS and VIMS-IR Layouts. . . . . . . . . . . . . . . . . . . . . . . . 5

2.1 VIMS Image and Spectra of Saturn at Near-IR Thermal Wavelengths . . . 72.2 Altitude Sensitivity of the the VIMS Thermal Spectrum of Saturn . . . . . 82.3 VIMS Near-IR Reflection Images and Spectrum of Saturn. . . . . . . . . . 92.4 Comparison of VIMS and SpeX Near-IR Reflection Spectra of Saturn . . . 102.5 Weighting Functions for Aerosol-free Opacity . . . . . . . . . . . . . . . . . 112.6 Henyey-Greenstein Phase Functions . . . . . . . . . . . . . . . . . . . . . . 132.7 Doubling Method for the Computation of Radiance in a Multiple Scattering

Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152.8 Effects of Rayleigh Scattering on Saturn’s Near-IR Spectrum . . . . . . . . 16

3.1 Vertical Locations and Compositions of Clouds in Saturn’s Atmospherebased on Thermochemical Equilibrium Theory . . . . . . . . . . . . . . . . 17

3.2 Retrieval of Cloud Pressure Level and Particle Size for the Compact CloudModel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3.3 Limb-darkening Observations of Saturn. . . . . . . . . . . . . . . . . . . . 233.4 Limb-darkening Analysis for a Compact Ammonia Cloud Model . . . . . . 253.5 Limb-darkening Analysis for a Compact Ammonium Hydrosulfide Cloud

Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263.6 Limb-darkening Analysis for an Extended Ammonia Cloud Model. . . . . . 283.7 Limb-darkening Analysis for a Continuous Aerosol Model. . . . . . . . . . 303.8 Limb-darkening Analysis for a Compact Ammonia Cloud and a Compact

Haze Layer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.1 Sensitivity of Nemesis to Atmospheric Parameters . . . . . . . . . . . . . 344.2 Retrieval of Optical and Physical Haze Properties . . . . . . . . . . . . . . 364.3 Retrieval of Opacity for Various Aerosol Sizes . . . . . . . . . . . . . . . . 384.4 Retrieval of Particle Size for an Ammonia Haze . . . . . . . . . . . . . . . 394.5 Complex Refractive Indexes of Haze Candidates . . . . . . . . . . . . . . . 404.6 Sensitivity of Saturn’s Reflection Spectrum to Haze Pressure Levels . . . . 414.7 Tests on the Haze Vertical Profile . . . . . . . . . . . . . . . . . . . . . . . 42

iii

Chapter 1

Introduction

“The primary goal of Cassini-Huygens is to conduct an in-depth exploration of the Satur-nian System”, NASA (1989)

Visible to the naked eye, Saturn was observed before the beginning of Christianity by themost ancient civilizations as a moving point on the sphere of fixed stars1. Nevertheless,it is really with the start of telescopic astronomy in 1609 that Saturn was seen not asan ordinary round planet but as a much more complex system. In July 1610, Galileopointed his telescope, an instrument with an aperture of around 2cm, 20× magnificationand imperfect optics, at Saturn and noted that the planet was triple (North, 1974). Heshared his discovery in the most cryptic way by the use of an anagram: “smaismrmilme-poetaleumibunenugttauiras”. Combined in the right order, the letters revealed their truemeaning: Altissinum planetum tergeminum observavi [I have observed the highest planetto be triple-bodied].

It is only in 1655 that the Dutch astronomer Christiaan Huygens correctly interpretatedGalileo’s observations with the discovery of Saturn’s rings. He also discovered, in the sameyear, Saturn’s largest moon, Titan. A few years later, the Italian-French astronomer Gio-vanni Domenico Cassini would discover the satellites Iapetus (1671), Rhea (1672), Tethys(1684) and Dione (1684), as well as the division between the A ring and B ring of Saturn(1675).

The development of modern astronomy and space exploration has made it possible toenhance our knowledge of the Saturnian system. The first space missions to Saturn,Pioneer 11 (1979), Voyager 1 (1980) and 2 (1981), only flew rapidly by the planet, itsmoons and its rings. They provided a unique insight into the Saturnian system, but theyalso raised many new questions to planetary scientists. Following the Galileo mission toJupiter (1995-2000), the Cassini-Huygens mission to the ringed world was the next logicalstep.

1The very first recorded observations of Saturn were made by the Assyrian an Babylonian civilizationsin ∼ 800 BC.

1

CHAPTER 1. INTRODUCTION 2

Figure 1.1: Left : Saturn as sketched by Galileo in 1616. Middle: Portrait of GiovanniDomenico Cassini (1625-1712). Right : Portrait of Christiaan Huygens (1629-1695).

The Cassini-Huygens spacecraft is a joint NASA/ESA/ASI2 interplanetary mission aimedat studying the Saturnian system. The spacecraft is made of two main components: theCassini orbiter, which surveys Saturn, its moons and its rings, and the Huygens probe,which landed on Titan’s surface on 14 January 2005. NASA provided the Cassini orbiter,the launch vehicule, the mission operations and the telecommunication systems via theDeep Space Network (DSN). The development of the Huygens probe was carried out byESA. The Italian Space Agency was in charge for the supply of the orbiter’s hardwaresystems, as well as instruments for both the orbiter and the probe.

The Cassini-Huygens spacecraft was launched on 15 October 1997, using a Titan IVB-Centaur rocket at Cape Canaveral Air Force Station, Florida. On its way to Saturn, thespacecraft performed four gravity assist flybys of Venus on 26 April 1998 and 24 June1999, Earth on 18 August 1999, and Jupiter on 30 December 2000 (see Fig. 1.3). TheCassini-Huygens spacecraft reached Saturn’s satellite Phoebe on 11 June 2005, 19 daysbefore sucessfully entering Saturn’s orbit on 1 July 2004. During the first four yearsof mission, also known as the Prime Mission, Cassini orbited 74 times around Saturn,executed close targeted flybys of Titan (×44), Enceladus (×3), Phoebe, Hyperion, Dione,Rhea and Iapetus, and made distant flybys of other major moons. The huygens probewas released from the Cassini orbiter on 25 December 2004, and entered Titan’s upperatmosphere 20 days later. The Prime Mission ended on 30 June 2008 and was extendeduntil September 2010 under the name of Cassini Equinox Mission as the mission coincidedwith Saturn’s equinox. The mission is currently known as the Solstice Mission and willend at the summer soltice of Saturn’s nothern hemisphere in 2017.

2The ASI is the Italian Space Agency (Agenzia Spaziale Italiana)


Figure 1.2: The Cassini cruise trajectory to Saturn.

With a total mass at launch of 5636 kg3 and height of 6.87 m, Cassini-Huygens is one of thelargest interplanetary spacecraft ever built, and the third heaviest one after the two SovietUnion spacecraft Phobos4. The Cassini orbiter is equipped with 12 science instruments,which have been designed to meet 27 science investigations. Half of these instrumentsundertake in situ measurements of the plasma, dust and magnetic field in the Saturniansystem: CAssini Plasma Spectrometer (CAPS), Cosmic Dust Analyze (CDA), Ion andNeutral Mass Spectrometer (INMS), MAGnetometer (MAG), Magnetospheric IMagingInstrument (MIMI) and Radio and Plasma Wave Science (RPWS). The other half con-sists of remote sensing instruments operating at optical and radio wavelengths: CompositeInfraRed Spectrometer (CIRS), Imaging Science Sunsystem (ISS), UltraViolet ImagingSpectrograph (UVIS), Visible and Infrared Mapping Spectrometer (VIMS), RADAR, andRadio Science Subsystem (RSS). The Huygens probe carries 6 science instruments: GasChromatograph and Mass Spectrometer (GCMS), Aerosol Collector and Pyrolyzer (ACP),Descent Imager and Spectral Radiometer (DISR), Huygens Atmospheric Structure Instru-ment (HASI), Doppler Wind Experiment (DWE) and Surface Science Package (SSP).

The topic of this thesis is to investigate the physical and optical properties of aerosolsin Saturn’s atmosphere, as well as their vertical structure, which is addressed using theVIMS instrument. Therefore, the technical capabilities and main science objectives of theVIMS instrument are briefly described here.

3The total mass includes the propellant mass at launch of 3132 kg.4The two spacecrafts Phobos 1 and 2 were sent to Mars and its moons Phobos and Deimos on July

1988. They weighed each 6220 kg with their orbital insertion hardware attached.


Figure 1.3: Both sides view of the Cassini spacecraft.

The Visual and Infrared Mapping Spectrometer (VIMS) is an imaging spectrometer mea-suring scattered and emitted light from surfaces and atmospheres in the 0.35-5.1 µmwavelength range. It consists of two distinct 64×64 pixel cameras in the visible andnear-infrared: VIMS-VIS and VIMS-IR. Each pixel corresponds to a spectrum with 352wavelengths, 96 in the visible and 256 in the infrared. VIMS-VIS operates in the 0.35-1.0µm spectral range, and can reach a maximum spectral and spatial resolution of respec-tively 1.46 nm and 167 µrad in the high-resolution mode. In the normal mode VIMS-VISoperates at a resolution of 7.3 nm (5 pixel summing) and 500 µrad (3×3 pixel summing).The VIMS-VIS total field of view (FOV) is 2.4◦×2.4◦, however the visible channel coversan effective FOV of 1.8◦×1.8◦ in order to match the VIMS-IR total FOV. The infraredchannel VIMS-IR covers the 0.85-5.1 µm wavelength range. The instrument has a maxi-mum spectral resolution of 16.6 nm and covers an effective FOV of 500 µrad. The opticsof the visible and infrared channels are thermally isolated from the Cassini spacecraftand cooled down in order to reduce the effects of noise. VIMS-VIS and VIMS-IR aremounted together on the same palette so that they are optically aligned. Data from boththe visible and infrared channel are combined into the instrument’s electronics, and so asingle spectrum is produced simultaneously in the whole 0.35-5.1 µm wavelength range.

The VIMS instrument has been measuring the spectrum of Saturn at various latitudesand longitudes and for many emission, solar and azimuthal angles. Therefore, it allowsus not only to map atmospheric features all over the surface of Saturn, but also to con-strain numerical models by considering a specific area of the planet observed at differentviewing geometries. VIMS can also perform measurements of Saturn’s spectrum at a veryhigh spatial resolution. Thus the instrument capabilities has made it possible to trackdistinctive features such as storms in the planet’s atmosphere. The list of science objec-tives for the VIMS instrument is extensive. In particular, VIMS observations of Saturn’s


Figure 1.4: VIMS-VIS (left) and VIMS-IR (right) layouts.

atmosphere are expected to provide new insight on:i. The structure of clouds and haze in the troposphere and stratosphere, as well as theiroptical and physical properties.ii. The nature of tropospheric aerosols by spectral detection of cloud features.iii. The distribution of gas species, including condensables (such as ammonnia and per-haps water) and disequilibrium species (such as phosphine).iv. The bolometric bond albedo.v. The role of solar energy in the dynamics of the planet.vi. The three-dimensional distribution of lighting.

The analysis of VIMS data in Oxford has been undertaken using the numerical code Neme-sis (Irwin, 2008). The Non-linear Optimal Estimator for MultivariatE spectral analySISmodel was initially designed for the analysis of CIRS spectra of Saturn, but is also appli-cable to any planets and many other instruments, such as VIMS. NEMESIS is a retrievalalgorithm which undertakes two main tasks. First, the algorithm calculates the syntheticspectrum for an assumed atmospheric structure through the radiative transfer equation.Then, the synthetic spectrum is directly compared to the measurements and NEMESISadjusts the atmospheric structure so as to match more closely the observations. Thisprocess is iterated a number of times until convergence of the solution. NEMESIS is par-ticularly useful for studying the cloud and haze distribution in Saturn’s atmosphere sincethe algorithm can model both the reflected and thermal radiation in a multiple-scatteringatmosphere. NEMESIS can also compute synthetic spectra for many viewing geometries,including limb observations of the planet.

In the next section, we review the main characteristics of the infrared spectrum of Saturnmeasured by the VIMS instrument, as well as its relevance for the investigation of aerosolproperties in the troposphere of the planet.

Chapter 2

VIMS Near-IR Spectrum of Saturn

2.1 VIMS data

VIMS raw cubes can be downloaded directly from the Planetary Data System (PDS)which stores scientific data from NASA planetary missions. VIMS data are stored intoarchive volumes containing the raw spectral images, the instrument pointing and naviga-tion, as well as the calibration files and software. VIMS data were released for the firsttime to the whole scientific community on 1 July 2005. To date there are 43 volumesavailable, and the next release is expected on 1 October 2011.

The data calibration was performed by using the calibration pipeline provided by thePDS. The software ISIS was used to add all the backplane data in the cubes. They pro-vide information on the geometry of observations such as the phase angle, emission angle,solar incidence angle, as well as the latitude, longitude, pixel resolution and line resolu-tion. In addition, the VIMS team at the University of Arizona developed continuouslytheir own calibration pipeline. However this version of the pipeline is not available and weonly have a few cubes calibrated with the updated algorithm. Extensive testing was madeto compare VIMS data calibrated with the PDS and the University of Arizona pipelines.By using new flatfields files measured in 2009, both pipelines provide identical calibrateddata. The new flatfields files were incorporated in the PDS pipeline to calibrate all VIMSdata.

The spectral cubes are provided without error measurements. The photon noise has beenestimated to be less than 1 DN, which corresponds to a relative error on the flux of around0.1 %. However, folding in all of the other errors (especially, absolute calibration errors ofhow the instrument responds), the measurement errors are more likely to be a few percent.In particular, the measurement errors should be at least as large as the solar irradianceuncertainty in the near-infrared. Colina et al. (1996) estimated this error to be around3-5 %. The measurement errors are greater where water absorbs in particular since thecalibration was done on Earth where water exists (K. Baines, personal communication,November 2008). Therefore, it is generally agreed to set up the measurement errors at10% of the mean flux (Sromovsky et al., 2010).

6

CHAPTER 2. VIMS NEAR-IR SPECTRUM OF SATURN 7

2.2 Thermal Spectrum

The spectrum of Saturn between 4.6 and 5.1 µm is formed mainly by the thermal radiationemitted from the deepest layers of the planet’s atmosphere and absorbed by the two mainopacity sources, i.e. aerosols and gases. In this spectral region, the gaseous absorption isdominated by phosphine (PH3) with smaller contributions from ammonia (NH3), arsine(AsH3), germane (GeH4), deuterated methane (CH3D), carbon monoxide (CO) and water(H2O) (Fletcher et al., 2009).

Figure 2.1: VIMS image and spectra of Saturn at near-IR thermal wavelengths. Left:VIMS image at 5.06 µm extracted from the VIMS spectral cube CM1524383985 andacquired on 22 April 2006 on the nightside. The spectral cube was calibrated with thePDS pipeline. The thermal image of Saturn shows significant latitudinal structure, aswell as the presence of bright and dark spots. Here the dark pixels correspond to regionof high opacity where the internally generated heat radiation is absorbed by aerosols andgases. In contrast, bright pixels correspond to region of relatively low opacity. Right:Typical VIMS thermal spectra at different northern latitudes extracted from the samecube CM1524383985.

The weighting functions for a cloud-free and a typical cloud model are shown on Fig.2.2. For a cloud-free atmosphere, only gases are responsible for the absorption of thermalradiation. Most of the gaseous absorption occurs at around 4-5 bar in the upper limitof the wavelength range investigated here. Note that the thermal spectrum of Saturnis still sensitive to pressure levels up to 500 mbar, especially at lower wavelengths. Theweighting functions for a tropospheric cloud located at 2 bar1 are shifted higher up inthe atmosphere at the exact position of the cloud. Therefore, VIMS data in the 4.6-5.1µm wavelength range are particularly valuable for sounding the cloud structure and gasdistribution at deep tropospheric levels.

1Thermochemical models predict the formation of ammonia ice between 1 and 2 bar (Atreya et al.,2004)


Figure 2.2: Altitude sensitivity of the VIMS thermal spectrum of Saturn. left : Theweighting functions are calculated for a cloud-free atmosphere. The dark colours showthe pressure levels at which gaseous absorption is significant. At a given wavelength,the thermal spectrum is not sensitive to pressure levels below the peak of the weightingfunction. right : One compact cloud is present at 2 bar. The weighting functions areshifted to the altitude of the cloud layer. In this case, the tropospheric cloud contributesthe most to the opacity sources, and any other aerosol or gas opacity below the 2 barlevel would contribute very little to the thermal spectrum.

We focused our analysis of the VIMS thermal spectrum of Saturn on retrieving cloud andgas opacities from nightside measurements. There is still a signal at dayside in Saturn’sthermal spectrum. Nevertheless, dayside data are more difficult to interpret becausephysical phenomena such as scattering of sunlight can have a significant impact on thethermal spectrum of Saturn. NEMESIS can perfectly model the scattering of sunlightby the tropospheric clouds2, however it would add another level of complexity and newdegeneracies in the model.

2.3 Reflection Spectrum

The near-infrared spectrum of Saturn at 0.8-3.5 µm is made of six distinctive peaks (seeFig. 2.3). These features are the result of strong absorption of the continuum3 at 0.90,1.01, 1.17, 1.40, 1.75, and 2.25 µm (Clark et al., 1979). Methane is the main gaseousabsorber with many absorption bands in the 0.8-3.5 µm wavelength range. Ammonia,hydrogen, and phosphine also contribute to the absorption of the continuum at thesewavelengths (Baines et al., 2005 ; Butler et al., 2005). In particular, 1-0 band pressure-induced absorption due to molecular hydrogen dominates the 2.0-2.5 µm spectral region.

2Tropspheric gases also scatter the sunlight but at shorter wavelengths than those considered here.3The continuum refers to the spectrum at wavelengths where there is no absorption. It is also the

continuous solar spectrum reflected by a non-absorptive surface and measured at the viewing geometryof VIMS instrument.


Figure 2.3: VIMS near-IR reflection images and spectrum of Saturn. Top: Near-IR imagesof Saturn at 1.07 µm (left) and 1.40 µm (right). These images were extracted from theVIMS cube CM1469004529 acquired on 23 March 2004. The bright pixels in the 1.07 µmimage correspond to the reflection of sunlight by the planet’s haze and rings. In contrastto Saturn’s image at 5.06 µm, radiation at 1.07 µm is homogeneously distributed overthe whole dayside area of the planet (here, on the left side of VIMS image). The sunlightscattered by tropospheric aerosols is absorbed at 1.40 µm by gaseous methane located athigher altitude levels. So the planet appears black at this wavelength, while scattering ofsunlight by Saturn’s rings is still measured. Bottom: Near-IR spectrum of Saturn from0.8 to 3.5 µm. The spectrum was extracted from the same spectral cube as the imagesabove. The selected spectrum was measured at a latitude of 4◦N and at an emission angleof 61◦. There is no signal at nightside from 0.8 to 3.5 µm. This dayside spectrum wasobtained at a relatively high solar angle of 51◦ and at an azimuthal angle of 58◦.


Note that the signal between 1.60 and 1.65 µm has a very low quality due to an overlap offilters at 1.64 µm, and so this spectral range can be ignored in further analysis of VIMSdata. There is also another joint at 2.98 µm. Sromovsky et al. (2010) noticed anotherVIMS artefact in the thermal spectrum of Jupiter. They observed an enhancement ofradiance at 2.3 µm probably due to light scattered by the grating of the infrared channel.This feature is not present in the VIMS spectrum of Saturn. Indeed, no significant signalin the gaseous absorption bands was measured by the VIMS instrument except in themethane absorption band at 1.01 µm. The VIMS spectrum was compared to Saturn’sreflection spectrum measured by the SpeX instrument mounted on the NASA InfraredTelescope Facility (IRTF) in Mauna Kea (see Fig. 2.4). A similar feature is observed inthe SpeX spectrum which means that the high reflectance at 1.01 µm is not an artefactspecific to VIMS instrumentation. This feature is likely to be due to the presence ofscatterers at higher altitude levels where methane absorption is less significant4.

Figure 2.4: Comparison of VIMS and SpeX near-IR reflection spectrum of Saturn. Left :The two VIMS spectra were extracted from the spectral cube CM1564373828 acquired on29 July 2007 at a latitude of 20◦S and 35◦N. SpeX spectra were measured on May 2007and co-added along the central meridian (Rayner et al., 2009). All spectra are scaled tothe VIMS spectrum measured at a latitude of 20◦S. The change in reflectance at 1.50µm is due to latitudinal variations of aerosol opacity. The high reflectance in the firstabsorption band at 1.01µm is observed in all VIMS measurements and also in the Spexspectrum. Therefore, this feature is not a VIMS artefact and is probably formed byscattering of sunlight by upper level hazes. Right : Retrieval of aerosol concentration fromVIMS specta at a latitude of 20◦S and 35◦N. The aerosol opacity retrieved at 35◦N latitudeis significantly less than at 20◦S. The cumulative optical depth increases slowly at 35◦Nlatitude, therefore the reflection spectrum is also sensitive to deeper atmospheric levels.In particular, the increase of opacity around 2 bar might affect the relative intensity ofthe six reflection peaks.

The source of emission for the continuum is scattering of sunlight in Saturn’s atmosphere,so spectral features at 0.8-3.5 µm are only present in dayside spectra. The sunlight iscontinuously scattered back into the direction of the spacecraft, except at specific wave-

4Stratospheric aerosols have smaller sizes and so they can only be detected at lower wavelengths.Therefore, no signal in the other absorption bands was measured at higher wavelengths.


lengths corresponding to gaseous absorption bands5. Therefore, the 0.8-3.5 µm spectrumand the individual peaks can be equally referred to as the reflection spectrum and thereflection peaks.

A limit on the vertical location of the scatterers can be inferred by mapping the weightingfunction for an aerosol-free atmosphere (see Fig. 2.5). Assuming no aerosols are presentin Saturn’s atmosphere, gaseous absorption only contributes to the total opacity of theplanet’s atmosphere, and so the two-way weighting functions6 show the pressure levels atwhich gases absorb the reflected sunlight. At the reflection peaks, absorption is significantbelow 1 bar, and so only scatterers located above this level can contribute to the reflec-tion spectrum. In between the reflection peaks, absorption starts above the tropopause at∼100 mbar and extends throughout the lower stratosphere up to ∼10 mbar. Therefore,the scatterers are located deeper than the 10 mbar level, or else they would contributesignificantly to the spectrum in the absorption bands7. The analysis of Saturn’s spectrumat 0.8-3.5 µm is then particularly relevant for surveying the nature and vertical distribu-tion of scatterers located between 1 and 10 bar. Towards this end, a scattering modelneeds to be properly defined and implemented in Nemesis.

Two-way Weighting Function Aerosol-free model

1.0 1.5 2.0 2.5 3.0 3.5Wavelength (µm)

10.000

1.000

0.100

0.010

0.001

Pres

sure

(bar

)

0

0.5

1.0

Figure 2.5: Weighting functions for aerosol-free opacity. The weighting functions at eachwavelength were derived from the two-way transmission, i.e. the transmission to andfrom each atmospheric level. The calculation was made assuming that no aerosols arepresent in Saturn’s atmosphere, so the only source of opacity is gaseous absorption. Thedot-dashed lines show the position of the reflection peaks. The weighting functions werenormalized to 1.

5The sunlight is still scattered back from upper level hazes, but this is not detected in the 0.8-3.5 µmspectrum except at 1.01µm

6The two-way weighting functions are derived from the transmission functions to and from the reflec-tive layers.

7Aerosols can be present higher in the stratosphere, but Fig. 2.4 shows that they are only detected at1.01µm and contribute very little to other wavelengths.


2.4 Scattering Model

In planetary atmospheres, particles are sufficiently far away from each other to be con-sidered as independent scatterers. Therefore, Saturn’s atmosphere can be treated as asum of small volume elements with individual scattering properties (Hansen et al., 1974).The theory describing light scattering within these small volume elements is known assingle scattering. Light scattering in the entire atmosphere is then computed by multiplescattering algorithms. Rayleigh scattering and Mie scattering are both implemented inNemesis.

Rayleigh scattering

Rayleigh scattering applies to small particles with size much less than the incident wave-length of the incident radiation. This condition implies that individual particles can betreated as being in a homogeneous external electric field. Another condition for Rayleighscattering to apply is that the ratio between the wavelength and the complex refractiveindex should be larger than the particle size. The incident radiation polarizes the par-ticles’s own field. The dipole moment induced by the incident radiation is proportionalto the external electic field. The second condition means that the time of polarization inresponse to incident radiation is short compared to the period of the incident wave. TheRayleigh phase matrix8 is as follows (Chandrasekhar, 1950):

34(1 + cos2 α) −3

4sin2 α 0 0

−34sin2 α 3

4(1 + cos2 α) 0 0

0 0 32cosα 0

0 0 0 32cosα

where the phase angle α is the angle between the incident and emission angles. Rayleighscattering theory is particularly relevant for understanding light scattering by moleculargas.

Mie scattering

Light scattering by spherical particles with size of the same order of the incident wave-length is described by Mie theory. The Mie phase function is a function of the sizedistribution and complex refractive index, n = nr + ini of particles. The refractive in-dex is a complex number and carries important information on the optical and physicalproperties of particles. The real and imaginary parts of the index represent respectivelythe phase velocity of radiation and absorption by scatterers. Mie scattering theory pro-vides an analytical solution for spherical particles interacting with plane waves, thereforescatterers are modelled as single spheres in numerical codes. Nemesis computes the radi-ation absorbed and scattered by spherical particles with various size distributions. Herethe scatterers are assumed to have a standard gamma size distribution

8Radiation polarized by Rayleigh scatterers is represented by a phase matrix in the I, Q, U, and VStokes vector. The solution is for isotropic Rayleigh particles.


n(r) ∝ r(1− 3b)/b exp(− r

ab),

where a and b are respectively the mean radius and the variance.

Figure 2.6: Henyey-Greenstein phase functions. Top: Simple Henyey-Greenstein functionsfor forward (g = 0.5), isotropic (g = 0.01), and backward scattering (g = −0.5). Bottom:Combined Henyey-Greenstein functions. The three phase functions have both forwardand backward peaks with variable relative strengths. By tuning the values for f , g1, andg2, a wide range of phase functions can be modelled including the specific case where bothforward and backward scattering are simultaneously present. Note that for f = 0, 1, thephase functions are identical to the single Henyey-Greenstein functions.

The Mie phase function is calculated by Nemesis and then fitted with the Henyey-Greenstein function in order to make the computation of radiance more efficient withinthe retrieval code. The Henyey-Greenstein phase function is given by

p(θ) =1

4π

1− g2

(1 + g2 − 2g cos θ)3/2,

where θ is the scattering angle9 and g the asymmetry factor.

The Henyey-Greenstein function can account for strong forward, strong backward, andisotropic scattering (see Fig. 2.6). However, the function cannot properly model all thepossible phase functions. In particular, there is no Henyey-Greenstein phase function

9Here the definition of the Henyey-Greenstein function implies that forward scattering occurs forθ ∈ [−π

2 ,π2 ], and backward scattering for θ ∈ [−π,−π

2 ] ∪ [π2 , π].


that can describe both strong forward and backward scattering simultaneously. Sincemost atmospheric scatterers are likely to have both forward and backward scatteringpeaks, Nemesis uses instead the combined Henyey-Greenstein function

p(θ) =f

4π

1− g21

(1 + g21 − 2g1 cos θ)3/2

+1− f4π

1− g22

(1 + g22 − 2g2 cos θ)3/2

,

where f is a fraction between 0 and 1, and {g1, g2} are the asymmetry factors of the twocontributing functions.

The Rayleigh and Mie phase functions are calculated for each small volume element withinSaturn’s atmosphere. A multiple scattering code is needed to compute the radiation emit-ted at the top of the atmosphere after being scattered by all volume elements. Nemesis

uses a doubling or adding method which consists in calculating the reflection, transmissionand source function for individual layers. Adjacent layers are then combined to derive thethe reflection, transmission, and source functions for the two layers. This process can beiterated until all the layers are successively integrated in the numerical code.Consider one homogeneous layer with bottom and top boundaries indexed by 0 and 1

respectively (see Fig. 2.7). For any incident, emission and azimuth angle (θ′, θ, φ), thecoefficients of reflection and transmission, and the source function due to thermal emissionwithin the layers are given by rαβ, tαβ and Jαβ for {α, β} ∈ {0, 1}. The flux emitted atthe top and bottom layers are given by

I+1 (µ, φ) =

Z 2π

0

Z 1

0I+0 (µ′, φ′)t01(µ, µ′, φ′ − φ)µdµdφ +

Z 2π

0

Z 1

0I−1 (µ′, φ′)r10(µ, µ′, φ′ − φ)µdµdφ + J01(µ, φ)

I−0 (µ, φ) =

Z 2π

0

Z 1

0I+0 (µ′, φ′)r01(µ, µ′, φ′ − φ)µdµdφ +

Z 2π

0

Z 1

0I−1 (µ′, φ′)t10(µ, µ′, φ′ − φ)µdµdφ + J10(µ, φ)

where µ = cos θ and µ′ = cos θ′. In Nemesis, the integral equations are replaced by thematrix equations

I+1 = t01I

+0 + r10I

−1 + J+

01

I−0 = r01I+0 + t10I

−1 + J+

10

Similar equation as those shown above can also be written for the second layer bounderedby the surfaces indexed by 1 and 2. Then the flux emitted by the combination of thelayers is

I+2 = t02I

+0 + r20I

−1 + J+

02

I−0 = r02I+0 + t20I

−1 + J+

20

where the coefficients of reflection and transmission, and the source function for thecombined layer are

r02 = r21 + t12(E − r10r12)−1r10t21

t02 = t12(E − r10r12)−1t01

J+02 = J+

12 + t12(E − r10r12)−1(J+

02 + r10J−21)


Figure 2.7: Doubling method for the computation by Nemesis of radiance in a multiplescattering atmosphere.

2.5 Aerosol-free Atmosphere

The main scatterers in planetary atmospheres are molecular gas and particles in con-densed phase, commonly referred to as aerosols. Depending on the physical and chemicalprocesses that lead to their formation, aerosols can also carry many other names such asclouds, haze, dust,etc. In Saturn’s atmosphere, clouds are expected to form at deepertropospheric levels than those considered in this analysis (Atreya et al., 1999). On theother hand, hazes might form in the troposphere above the cloud layers, and also in thestratosphere (Ortiz et al., 1995). Thus, the generic term of haze will be used to refer toaerosols located between 1 and 10 bars.

The synthetic reflection spectrum for an atmosphere free of aerosols was generated byNemesis (see Fig. 2.8). The only scatterers present in this atmospheric model are thenmolecular gas, which can deviate the sunlight through Rayleigh scattering. Rayleightheory states that the ratio between emitted and incident radiation decreases with wave-length. The synthetic spectrum shows two distinctive signals at the spectral position ofthe first two peaks. The radiances for these peaks are respectively ∼ 4 and ∼ 2 µWcm−2sr−1µm−1. Comparison with VIMS spectra measured at the same latitude and lon-gitude, and for the same solar and emission angles shows that Rayleigh emission frommolecular gas is small comparing to measurements . Therefore, this effect cannot accountfor the 0.8-3.5 µm spectrum and so aerosols are the main scatterers at these wavelengths.


Rayleigh Scattering Spectrum in the Near-IR of Atmospheric Molecular Gas

0.80 1.07 1.33 1.60 1.86 2.13 2.39 2.66 2.92 3.19 3.45Wavelength (µm)

0

10

20

30

40

50

60

70

Radi

ance

(µW

cm

-2 sr

-1 µ

m-1)

Measured by VIMSModelled with NEMESIS

Figure 2.8: Effects of Rayleigh scattering on Saturn’s near-IR spectrum. The spectrumwas extracted from VIMS image cube CM1524386400 measured on 16 April 2006 ata latitude of 19◦S. The solar and emission angles are respectively 76◦ and 38◦. In thischapter, analysis of Saturn’s reflection spectrum is based on this particular data set unlessotherwise stated.

Chapter 3

Tropospheric Clouds

The knowledge of cloud and haze distribution in Saturn’s atmosphere is particularly valu-able since aerosols are key ingredients in many atmospheric processes such as the radiativetransfer, the energy balance, microphysical reactions and stratospheric photochemistry.Aerosols can also be used as passive tracers of wind motion. Theoretical expectations ofaerosol formation and structure are based mainly on thermochemical equilibrium theoryand photochemistry. Thermochemical models predict the formation of clouds by con-densation of a gaseous phase into an aqueous or solid solution. In Saturn’s atmosphere,oxygen, nitrogen, and sulphur are expected to combine with hydrogen to form clouds inthe troposphere, from the tropopause down to the 20 bar pressure level (Lewis, 1969 ;Weidenschilling and Lewis, 1973 ; Atreya et al., 2004). We mentioned in the previoussection that the thermal spectrum of Saturn between 4.6 and 5.1 µm is not sensitiveto opacity sources located below 5 bar. Ammonia (NH3) and ammonium hydrosulphide(NH4SH) are the only molecules expected to condense above this pressure level (see Fig.3.1).

Figure 3.1: Vertical locations and compositions of clouds in Saturn’s atmosphere basedon thermochemical equilibrium theory (Atreya et al., 2004)

17

CHAPTER 3. TROPOSPHERIC CLOUDS 18

3.1 Cloud Vertical Structure

3.1.1 Fitting strategy

The first part of the analysis of VIMS thermal spectrum focused on the retrieval of aerosoland gas opacities from one single spectrum at a given viewing geometry. Data were ex-tracted from nightside spectral cubes at low emission angles. The last condition makesthe computation of radiance by Nemesis more efficient since limb-darkening effects arenegligible. The temperature and para-ortho ratio profiles were derived from CIRS obser-vations at similar latitude and emission angle as for the selected VIMS spectra.

Sensitivity tests showed that the atmospheric parameters most affecting the Saturn’sthermal spectrum are the aerosol opacity, and the phosphine and ammonia gaseous abun-dances. In the spectral range considered for this analysis, Saturn’s spectrum is sensitiveto phosphine and ammonia at pressure levels where they have homogeneous distribution.Therefore scaling factors to their a priori values in the deep troposphere were retrievedfrom the VIMS thermal spectrum.

A Mie scattering model was used to generate the cloud absorption coefficient, the singlescattering albedos and the scattering phase functions at all wavelengths. The Henyey-Greenstein parameters were fitted to the Mie phase function. The complex refractiveindexes for ammonia (Martonchik et al., 1984) and ammonium hydrosulfide (Howett etal., 2007) are known from laboratory measurements1. The five optical parameters werecalculated for a few wavelengths between 4.6 and 5.1 and for particle sizes ranging from100 nm to 10 µm, and then implemented in the retrieval algorithm.

The cloud opacity, phosphine and ammonia abundances were retrieved simultaneouslywith Nemesis for various cloud models:

i.: Vertical structure: two models were tested in this analysis, one compact and one ex-tended structure for the cloud. For the compact cloud model, the aerosol density is zeroeverywhere except at one pressure level. The retrieval code derived the cloud opacityat this specific pressure level which provides the best fit to the observations. For theextended cloud model, there are two free parameters, the base pressure and the extensionof the cloud2

1The presence of ammonia and ammonium hydrosulphide clouds is expected from thermochemicalequilibrium theory. However no spectral signatures have been detected for either of these molecules. Thefirst results show that VIMS near-IR spectrum of Saturn is mainly sensitive to the vertical distribution ofaerosols and the particle size. Since there are less constraints on the complex refractive index, the valuesfor ammonia and ammonium hydrosulphide have been used so far for the analysis of the VIMS thermalspectrum.

2The extension above the pressure level at which condensation occurs can be defined either by theheight of the cloud in kilometres or by a fractional scale height ranging from 0 to 1.


ii.: Composition: Retrievals were performed assuming the cloud is made of either ammo-nia or ammonium hydrosulphide particles

iii.: Interaction radiation-matter: Radiation can be scattered by clouds and gases locatedabove the source of thermal emission. This effect was tested by fitting the VIMS spectrumfor scattering and purely absorptive opacity sources in the troposphere.

iv.: Number of clouds: Thermochemical equilibrium models predict the formation of twoclouds above the 5 bar level. Therefore, the analysis of the VIMS thermal spectrum wascarried out for 1 and 2 aerosol layers.

3.1.2 Results

The results for one compact cloud made of ammonia ice are shown on Fig. 3.2. We foundthat we can fit the VIMS thermal spectrum equally well for both a scattering and a purelyabsorptive atmosphere. In both cases, we also found that the cloud base pressure needsto be less than 3.2 bar in order to fit the observations properly. The top boundary ofthe cloud is less constrained, except for a scattering atmosphere. In that case, we couldnot fit the VIMS spectrum with clouds above the 2.5 bar pressure level. Best fits areobtained for a cloud pressure base at 2.7 bar for both purely absorptive and scatteringaerosols, however there are less constraints on the particle size. In a purely absorptiveatmosphere, only particles with sizes ranging from 300 nm to 2 µm can model the VIMSspectrum. There are no such constraints for a scattering atmosphere in the range of sizeinvestigated, and particles of all sizes are potential candidates in the retrieval problem.

Similar values for the mean opacity of the compact cloud, and for the deep volume mix-ing ratios (VMR) of phosphine and ammonia3 were retrieved for the range of parametersproviding a good fit to the observations. For a cloud at 2.7 bar made of purely absorptiveparticles of one micrometre size, optimum fit to the thermal VIMS spectrum was obtainedfor a mean opacity of 10.9, a deep phosphine VMR of (7.03 ± 0.36) × 10−6 and a deepammonia VMR of (2.48±0.40)×10−3. For a scattering cloud, these values are respectively14.5, (6.84 ± 0.35) × 10−6 and (2.03 ± 0.34) × 10−3. The values retrieved for the deepphosphine VMR are in good agreement with recent measurements from Cassini/CIRS.Fletcher et al. (2009) derived an average value of (6.4± 0.4)× 10−6 from the analysis ofhigh-resolution spectra of Saturn in the 600-1430 cm−1 wavenumber range. Nevertheless,typical values for the deep ammonia VMR are of one order smaller than the values derivedhere. Davis et al (1996) retrieved a value of 10−4 from ISO observations of Saturn in thefar-infrared.

3The mean opacity is averaged over all the wavelengths. As seen on Fig. 3.2 there is actually littlevariation of the optical depth at the cloud pressure level. Deep VMR values for phosphine and ammoniacorrespond to pressure values above ∼500 mbar and ∼2 bar respectively.


We found that the optical depth value retrieved by the algorithm depends on the cloudmodel. Clouds detected by remote sensing instruments have optical depth of typically afew units, so the values found here are very likely to be overestimated. This is due tothe simplicity of the compact cloud model. Since the aerosol concentration is assumed tobe zero everywhere except at the base pressure level, Nemesis retrieved a high opticaldepth so as to account for the absorption and scattering of thermal emission. We alsofound that the optical depth for a scattering cloud is higher than for a purely absorptivecloud. This was observed for all cloud models and can be explained as follow: Thermalemission scattered by the troposphoric cloud is spread over many emission angles, and sothe VIMS instrument for a specific viewing geometry measures less flux than for a purelyabsorptive cloud.

As mentioned in the previous section, many cloud models were tested in our analysis ofone single thermal spectrum. In particular, we managed to fit the VIMS spectrum equallywell for a compact and an extended cloud. It should be noted that an extended cloudadds another free parameter in the aerosol model. In Nemesis, the cloud extension canbe parametrized by either a fractional scale height between 0 and 1 or a cloud heightin kilometres. Significantly smaller values of the optical depth were retrieved for an ex-tended cloud. Nevertheless, the range of atmospheric parameters that can possibly fit theobservations is larger than for a compact cloud, and so the extended cloud model turnsout to be highly degenerate.

Thermochemical theory predicts the condensation of ammonium hydrosulphide below theammonia ice layer at ∼ 5 bar (see Fig. 3.1). At this pressure level, clouds are difficult todetect due to gaseous absorption (see Fig. 2.2). Nevertheless, the ammonium hydrosul-fide cloud is expected to extend up to the 3 bar level where aerosol opacity contributessignificantly to the thermal spectrum of Saturn. For that reason, the same analysis of onesingle thermal spectrum was carried out for a compact and extended ammonium hydro-sulphide cloud. The conclusions drawn for an ammonium hydrosulphide cloud are verysimlar to those for an ammonia cloud model, therefore we cannot distinguish between thetwo models. Furthermore, neither ammonia nor ammonium hydrosulphide cloud featureshave been spectrally detected so far. So we cannot exclude the possibility of a more com-plicated model for the complex refractive indexes. This obviously adds a higher level ofdegeneracy in the retrieval problem.

Finally, we did not constrain either the number of clouds necessary to match the ob-servations. In conclusion, the retrieval problem is highly degenerate and many possiblesolutions can be found to fit one single VIMS thermal spectrum. One way to break thedegenerencies of the retrieval problem is to look at independent measurements simulta-neously. This is done by combining spectra of the same area in Saturn’s atmosphere butobserved at different viewing geometries.


Figure 3.2: Retrieval of cloud pressure level and particle size for a purely absorptive (left)and scattering (right) cloud model. The Saturn’s spectrum used here was extracted fromthe VIMS cube CM1524383985 (see Fig. 2.1), and measured at a latitude of 0.2◦N andemission angle of 10◦. The results are shown for one compact cloud made of ammonia ice.The two figures at the top correspond to the goodness-of-fit [χ2/(number of wavelengths)]as a function of the cloud pressure level and particle size. Optimum fits to the observationsare obtained for typical χ2/nλ values . 1.0, so values above 5.0 are omitted here. Thesynthetic and VIMS spectra are shown on the middle panel for optimum values of thecloud pressure level (Pb = 2.7 bar) and particle size (radius = 1 µm). The bottom panelshows the cumulative optical depth from the top of the atmosphere to all pressure levels.


3.2 Limb-darkening

3.2.1 Methodology

Many cloud models have been found to fit individual VIMS spectra perfectly well. Inthis section, we focus on the analysis of spectra emitted from the same region of Saturnbut viewed at different angles. This extra information is expected to constrain the cloudstructure and composition. As the planet rotates, the VIMS instrument keeps measuringspectra and so the same area on the planet can be observed at different emission angles.Therefore the time difference between two measurements of the flux emitted by the samearea of the planet, but at different emission angles is relatively small4. The variations ofaerosol structure and abundance are expected to be negligible over such a short period oftime, and so spectra can be considered as independent measurements of the same physicalquantity.

We show in Fig. 3.3 how thermal spectra of the same area of Saturn viewed at differ-ent emission angles can be seen as independent measurements and so why they can betreated simultaneously in order to constrain the atmospheric structure of the planet. Fourspectra were extracted from the VIMS spectra cubes CM1523880874, CM1523879088,CM1523877318 and CM1523875542. The cubes were all acquired on 16 April 2006 be-tween 10:15:22.999 and 11:44:14.964 (UTC). All spectra were emitted from a latitude of21◦S and a longitude 113◦ in the System III West, but measured at different emissionangles ranging from 27◦ to 76◦. The cloud opacity, as well as the phosphine and ammo-nia VMRs were retrieved with Nemesis from the spectrum at 27◦ emission angle. Asexpected from the previous section, we managed to fit quite well the low emission an-gle spectrum for various cloud structure and composition. The atmospheric structure ofSaturn was updated with the retrieved values of cloud opacity and gaseous VMRs, andthe synthetic spectra a higher emission angles were then calcucated. This time, we couldnot fit any of the high emission angle spectra measured with the VIMS instrument. If allspectra were non-independent measurements of the same physical quantity, the solutionfound at lower emission angle should also be a solution at higher emission angles. As seenon Fig. 3.3, this is clearly not the case, and so the analysis of each spectum individuallyprovides independent sets of solutions to the retrieval problem. By retrieving Saturn’scloud opacity and gas VMRs from all spectra simultaneously, we can find the commonsolutions and hopefully exclude many atmospheric structures that were found to fit VIMSspectra individually.

4In this report, the measurements of VIMS specta of the same area viewed at different emission anglesare only a few hours. Therefore, the aerosol opacity and volume mixing ratios of tropospheric gases areunlikely to undergo large variations over such a short time. Over larger time scales, the atmosphericstructure in belts and perhaps zones might vary and so we avoided analysing such data simultaneously.


Figure 3.3: Limb-darkening Observations of Saturn. Top: VIMS images CM1523880874(left) and CM1523875542 (right) µm at 5.06 µm acquired on 16 April 2006 at 11:44:14.964and 10:15:22.999 (UTC) respectively. The red squares show the same area [latitude =21◦S , longitude = 113◦ (System III West)] on the night side of the planet observed atan emission angle of 27◦ (left) and 76◦ (right). Middle-left : VIMS thermal spectra ofSaturn at a latitude of 21◦S and a longitude of 113◦, measured at an emission ange of 27◦,45◦, 65◦ and 76◦. The four spectra are extracted from the VIMS cubes CM1523880874,CM1523879088, CM1523877318 and CM1523875542 measured on 16 April 2006. Thecubes CM1523879088 and CM1523877318 were acquired at 11:14:28.976 and 10:44:58.987(UTC) respectively. Middle-right : Goodness-of-fit as a function of the emission angle forvarious cloud models. The cloud opacity as well as the phosphine and ammonia VMRswere retrieved from Saturn’s spectrum at 27◦ emission angle, and their values updatedin the atmospheric structure of the planet. The synthetic spectra at higher emissionangles were then calculated and directly compared to VIMS data. The atmosphericstructure retrieved from the spectrum at low emission angle provides very poor fits tothe observations at higher emission angles. So spectra of the same area measured atdifferent emission angles can be seen as independent measurements and should be analysedsimultaneously in order to narrow the constraints on Saturn’s atmospheric structure.Bottom: VIMS and synthetic spectra of Saturn for various cloud models at an emissionangle of 27◦ (left) and 76◦ (right).


3.2.2 Results

One compact ammonia cloud model

Saturn’s thermal spectra at a latitude of 21◦S and a longitude of 113◦ measured at 27◦

and 76◦ emission angles were analysed simultaneously. The two spectra were extractedfrom the VIMS spectral cubes CM1523880874 and CM1523875542 acquired on 16 April2006 (see Fig. 3.3). The aerosol model was first assumed to be one single compact cloudmade of ammonia ice. The aerosol opacity, and the phosphine and ammonia VMRs wereretrieved with Nemesis for various particle sizes and cloud pressure levels (see Fig. 3.4).For purely absorptive aerosols, the lowest χ2/nλ values are obtained for a cloud pressurelevel between 2 and 3 bar. We show on Fig. 3.4 the synthetic spectra at low and highemission angles for a cloud pressure level at 2.3 bar and a particle size of 1 µm. The fit tothe observations is quite poor especially at higher wavelengths. Therefore, VIMS spectraat low and high emission angles cannot be modelled simultaneously by one purely absorp-tive cloud made of ammonia ice. This was not the case for a single spectrum measuredat one specific viewing geometry. Indeed, while we found many possible solutions to theretrieval problem from the analysis of a single spectrum (see Fig. 3.2), by combining dataat different emission angles we managed to discriminate the compact cloud model madeof purely absorptive ammonia ice.

The same analysis was carried out for one compact cloud made of scattering ammoniaice. The minimum χ2/nλ values correspond to a cloud pressure either less than 2 bar ormore than 1.5 bar. The synthetic spectra are shown for a particle size of 1 µm, and acloud pressure level at 0.94 bar and 2.3 bar. The fit to the observations is slightly betterthan for purely absorptive ammonia ice. In particular, we managed to get a better fit athigher wavelengths between 5.0 and 5.1 µm. Therefore, these results show that scatteringeffects are very likely to be important in the wavelength range considered here. The nextstep is to refine the atmospheric structure, and especially the aerosol model, in order tomatch VIMS observations more closely.

One compact ammonium hydrosulfide cloud model

The atmospheric structure was modelled with one compact cloud made of ammoniumhydrosulphide particles. Both purely absorptive and scattering aerosols were consideredas in the previous section. The conclusions that can be drawn for the ammonium hydro-sulphide cloud are similar to those for an ammonia cloud. The main difference is thatminimum χ2/nλ values for scattering ammonium hydrosulfide particles are obtained fora cloud pressure level between 2 and 3 bar. Otherwise, even though scattering aerosolsprovide a better fit to the observations, the compact cloud model still fails at exactlymatching VIMS spectra at low and high emission angles simultaneously.


Figure 3.4: Limb-darkening analysis for a compact ammonia cloud model. Top:Goodness-of-fit as a function of the cloud pressure level and the particle size for a purelyabsorptive (left) and scattering (right) cloud. The synthetic and VIMS spectra at lowand high emission angles are shown for optimum χ2/nλ values on the bottom panels.


Figure 3.5: Limb-darkening analysis for a compact ammonium hydrosulfide cloud model.Top: Goodness-of-fit as a function of the cloud pressure level and the particle size for apurely absorptive (left) and scattering (right) cloud. The synthetic and VIMS spectra atlow and high emission angles are shown for optimum χ2/nλ values on the bottom panels.


One extended ammonia cloud model

Our first revision of the cloud model was to consider an extended cloud. Indeed, Lewiset al. (1969) predict from thermochemical theory that clouds should extend above thepressure level at which condensation starts. Although most of the cloud opacity is foundat the base pressure, the layers above might contribute significantly to the absorptionand scattering of thermal emission. For simplicity, we often considered a compact cloudmodel and managed to fit quite well VIMS spectra individually. Nevertheless, since wecannot fit VIMS spectra simultaneously at different viewing angles, the next logical stepis to consider an extended cloud model. The cloud vertical distribution is set by thebase pressure and the cloud extension in kilometres. The analysis of limb-darkening datawas carried out for an ammonia and an ammonium hydrosulphide cloud made of purelyabsorptive and scattering aerosol particles. The cloud opacity, as well as the VMRs ofgaseous ammonia and phosphine were retrieved with Nemesis for a wide range of particlesize, cloud base pressure and cloud extension.

The results are shown on Fig. 3.6 for an extended ammonia cloud made of scatteringparticles. The best fits were obtained for a deep cloud at pressure higher than 2.5 barwith an extension of at least 25 kilometres, and for a particle size between 1 and 3 µm.We show the synthetic spectra at low and high emission angles for a base pressure of 2.7bars, an extension of 25 kilometres and a particle size of 2 µm. The fit to the observationsis significantly better than all the compact cloud models considered so far. We alsoshow the aerosol concentration in particles per litre retrieved for this specific atmosphericstructure. The concentration at the base pressure was scaled to its theoretical valuederived by Atreya et al. (2004) from thermochemical theory (see Fig. 3.1). This was doneby assuming an homogeneous particle mass of 1 µg, and so the units are now expressedin g.l−1 The aerosol concentration derived from our analysis decreases significantly slowerthan predicted by thermochemical theory. In particular, we found an aerosol concentrationof 10−3 g.l−1 at 500 mbar while Atreya et al. predicted a value around 10−6 g.l−1. Asmentioned above, 25 kilometres is the minimum cloud extension providing a good fit to theobservations, and so 10−3 g.l−1 is also a minimum estimation of the aerosol concentrationvalue at 500 mbar. By assuming an extended cloud made of ammonium hydrosulphideparticles, similar conclusions can be drawn and the aerosol concentration at 500 mbar isstill overestimated by comparison with theoretical models. Finally by considering purelyabsorptive ammonia or ammonium hydrosulphide particles, we could not fit the VIMSspectra simultaneously at low and high emission angles. This is not surprising since weconcluded in the previous section than scattering is very likely to be necessary for limb-darkening analysis. The results for a scattering extended cloud show that a source ofopacity above the cloud base is needed in order to fit all spectra simultaneously. Thehigh value retrieved here for the cloud extension suggests that this opacity source is to befound at higher altitudes in the troposphere. Photochemical and microphysical modelspredict the existence of haze at pressure levels around a few hundred millibars, and so atropospheric haze was tested as a potential candidate for the secondary opacity source.


Figure 3.6: Limb-darkening analysis for an extended ammonia cloud model. Top:Goodness-of-fit as a function of the cloud pressure level and the particle size for a scat-tering (right) cloud. The results are shown for a cloud extension of 5, 10, 15, 20, 25 and30 kilometres. Bottom-left: Synthetic and VIMS spectra for an ammonia cloud at 2.7bar with an extension of 25 km and particle size of 2 µm. The aerosol concentration inparticles.l−1 retrieved for this atmospheric structure is shown on the bottom-right figure.The profile was converted in g.l−1 by assuming an homogeneous particle mass of 1 µg.The ad hoc value for the particle mass was chosen so that the concentration at the basepressure would be similar to the value derived from thermochemical models (see Fig. 3.1).


Continuous cloud model

One difficulty is that we have little knowledge of the location and composition of the haze.Also, the cloud at deep topospheric levels is probably the main opacity source and can beparametrized by many free parameters, such as the refractive indexes, the pressure base,the cloud extension and the particle size. We started our analysis by considering a simpleaerosol model: the particles are assumed to have a complex refractive index of 1.4+0.001i

and a size of 1 µm. Indeed, the value used here for the refractive index is typical for mostmaterials and especially for ammonia and ammonium hydrosulfide particles. Finally, ourresults from both the analysis of single and limb-darkening data provided an optimum fitto the observations for a particle size of around one micrometre.

The aerosol vertical distribution was initially set to a continuous a priori profile. Theaerosol concentration at each pressure level, as well as the phosphine and ammonia VMRswere retrieved for various a priori profiles. The results are shown on Fig. 3.7. Two mainopacity sources were retrieved with Nemesis: One deep tropospheric cloud with a basepressure between 2 and 3 bar, and a haze layer below the tropopause at around 200 mbar.Similar profiles were retrieved for all tested a priori profiles and provided a very close fitto VIMS spectra both at low and high emission angles. The non-cumulative optical depthat each pressure level shows that the deep tropospheric cloud contributes the most to theopacity sources, and so the features apparent on Saturn’s images at ∼ 5 µm are associatedwith the deep cloud. Nevertheless, our results show that Saturn’s thermal spectrum isstill sensitive to the haze located higher up in the troposphere.

Two compact clouds model

One main issue with the continuous model is that Nemesis derives the aerosol concentra-tion at each pressure level which increases significanty the number of free parameters inthe retrieval algorithm. To break some degenerencies inherent to the continuous aerosolprofile, we considered the cloud and haze model as follows: One compact ammonia cloudmade of 1 µm scatterers and one compact haze layer made of scattering particles with1.4 + 0.001i refractive index and 1 µm size. We retrieved the cloud and haze opacity,and phosphine and ammonia VMRs for a wide range of pressure levels for both the cloudand haze layers. The results are shown on Fig. 3.8. The best fits to the observationsare obtained for a deep ammonia cloud between 2.5 and 4 bar, and for a haze layer atpressures higher than 500 mbar.

In conclusion, the thermal spectrum of Saturn is sensitive to the tropospheric haze, and sowe need to know the haze structure in order to better constrain the physical and opticalproperties of the clouds at deeper tropospheric levels. To that end, we oriented ouranalysis to the near-infrared reflection spectum of Saturn which is particularly relevantfor the study of tropospheric haze.


Figure 3.7: Limb-darkening analysis for a continuous aerosol model. The complex refrac-tive index and particle size are respectively 1.4+0.001i and 1 µm. Top-left: A continuousaerosol distribution (solid line) is retrieved from three different a priori profiles (dot-ted line). The fit to the VIMS spectra at low and high emission angles are shown forthe three aerosol profiles retrieved with Nemesis. Bottom: Cumulative (left) and local(right) optical depth.


Figure 3.8: Limb-darkening analysis for a compact ammonia cloud and a compact hazelayer. The cloud and haze particles have a size of one micrometre. The haze particleshave a complex refractive index of 1.4 + 0.001. Top: Goodness-of-fit as a function of theammonia cloud and haze layer. Bottom: Synthetic and VIMS spectra for an ammoniacloud at 2.7 bar and a haze layer at 0.16 bar.

Chapter 4

Tropospheric Haze

Karkoschka et al. (1996, 2005) analysed Saturn’s images at methane bands and nearbycontinuum wavelengths between 0.6 and 0.96 µm. These images acquired with the HubbleSpace Telescope between 1991 and 2004 were used to retrieve the aerosol vertical structurein the troposphere and stratosphere of the planet. They found that the best fits to theobservations were obtained for a tropospheric haze layer located above the condensateclouds.

Most theories of the formation of stratospheric haze involve photochemical processes andbombardment by energetic particles, as well as microphysical reactions such as the nu-cleation, sedimentation and coagulation of haze particles. However, little is known onthe mechanisms leading to the formation of haze in the troposphere of Saturn. To betterunderstand these mechanisms, it is necessary to know the optical and physical propertiesand vertical stucture structure of the tropospheric haze.

Following the work done by Karkoschka et al., only a few papers investigating the prop-erties of the upper tropospheric haze have been published (Baines et al., 2006 ; Ortiz etal., 1996 ; Peres-Hoyos et al., 2005). Nevertheless, these studies focus on the analysisof visible and near-infrared images in a few filters, and so many ad-hoc assumptions aremade in the atmospheric stucture in order to fit properly the observations.

In this section, we used VIMS data sets in the near-infrared between 0.8 and 3.5 µm toanalyse the particle haze size, opacity, complex refractive index, and vertical distibutionin Saturn’s upper troposphere. As seen on Fig. 2.5, the near-infrared reflection spec-tum of Saturn is sensitive to pressure levels above the condensate clouds, and so VIMSspectra are particularly relevant for the investigation of tropospheric haze. VIMS reflec-tion spectra are made of 146 spectral points and were acquired at many different viewinggeometries and observation times. Therefore, VIMS data offer a unique opportunity tobetter constrain the haze properties in the upper troposphere of Saturn.

32

CHAPTER 4. TROPOSPHERIC HAZE 33

4.1 Sensitivity to Model Parameters

Temperature

The ratio to the initial reflection spectrum after perturbing the temperature is below1.1 and increases with wavelength. This dependency with wavelength is due to thermalemission which dominates Saturn’s near-IR spectrum at higher wavelengths. For the firstfour peaks, the variations in temperature have very little effect on the spectrum (less than1%). However, changes up to 10% of the radiance are observed for the last two peaks.The effects on the spectrum are particularly significant at the edge of the peaks wherethe signal is low, and for large temperature variations, which are less likely to occur inSaturn’s atmosphere. Therefore the temperature profile will be set to the profile retrievedfrom CIRS thermal spectra measured at similar times.

Complex refractive index

Tests on the real and imaginary parts of the refractive index were carried out for nr ∈ [1, 2]

and ni ∈ [0, 0.1]. These ranges of values embrace refractive indexes for most materials atany wavelength, and so no specific composition of tropospherice haze is a priori assumed.The ratio is very close to 1.0 for most values of ni1. Most importantly, the ratio to theinitial spectrum is nearly uniform for most imaginary refractive indexes and so there isno apparent correlation with wavelengths. That is to say, variations in the imaginaryrefractive index should affect the overall spectrum uniformly. On the other hand, theratio for the real refractive index can vary between 0.9 and 1.4 for the same value of theindex. Thus the effects on the spectrum can be large and have a wavelength dependency

Aerosol size

The aerosol size significantly affects the near-IR reflection spectrum at lower wavelength.There is a clear wavelength dependent effect of the size perturbation on the ratio, andthis is observed for all radii ranging from 100 nm to 1 µm2. For one specific particle size,the change in radiance can vary between 0% and up to 600%. Consequently, the particlesize affects both the amplitude and shape of the reflection spectrum, while the complexrefractive index has the main effect on the amplitude3. The problem is partly uncorrelatedsince the aerosol size can be tuned to match the shape of the synthetic spectrum toVIMS spectrum; then adjustment is made by varying the complex refractive index. Theproblem is only partly uncorrelated since both the real and imaginary refractive indexesare unknown parameters in the model.

1The largest variations are seen for "extreme” values of the refractive index. e.g. an imaginary indexof 0.1 leads to significant changes of the reflection spectrum, however it is statistically unlikely that thisvalue should be relevant since most materials have very small imaginary index below 0.1.

2Here also typical values for aerosol sizes were used. Note that contrary to the refractive index, thedependency with wavelengths is not a statistical bias since all the radius show a similar trend.

3This also applies to the real part of the complex refractive number. Even though variations in radianceup to 50% are observed, this is for extreme values of the index and is still relatively small compared toeffect of the particle size.


Sensitivity of Saturn’s Near-IR Reflection Spectrum to Temperature

1.0 1.5 2.0 2.5 3.0Wavelength (µm)

0.95

1.00

1.05

1.10

Ratio

-20

-15

-10

-5

0

5

10

15

20

Sensitivity of Saturn’s Near-IR Reflection Spectrum to Real Refractive Index

1.0 1.5 2.0 2.5 3.0Wavelength (µm)

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

Ratio

1.0

1.2

1.4

1.6

1.8

2.0

Sensitivity of Saturn’s Near-IR Reflection Spectrum to Imaginary Refractive Index

1.0 1.5 2.0 2.5 3.0Wavelength (µm)

0

2

4

6

8

10

Ratio

0

10-5

10-4

10-3

10-2

10-1

Sensitivity of Saturn’s Near-IR Reflection Spectrum to Particle Size

1.0 1.5 2.0 2.5 3.0Wavelength (µm)

0

2

4

6

8

Ratio

100 nm

300 nm

500 nm

700 nm

900 nm

2 µm

4 µm

10 µm

Figure 4.1: Sensitivity of Nemesis to atmospheric parameters. A synthetic reflection spectrumwas generated with Nemesis for an atmospheric profile made of tropospheric aerosols. Aerosolswere assumed to have a size of 1 µm with a variance of 0.05, and a refractive index of 1.4+0.001i.The temperature profile was retrieved from the CIRS thermal spectrum of Saturn measured atabout the same time period (Fletcher et al., 2010). A volume mixing ratio (VMR) of 4.5× 10−3

was chosen for methane (Flasar et al., 2005). For each perturbation of the temperature, realrefractive index, imaginary refractive index and aerosol size, the synthetic spectrum was re-calculated and then compared to the initial spectrum. The four figures show the ratio betweenthe re-calculated and initial radiances at the six reflection peaks. Perturbations of temperatureare expressed as a shift ∆T ∈ [−20, 20] homogeneously applied to the initial profile. The colorbarsfor the other parameters show their actual values after perturbation.

4.2 Optical and Physical Haze Properties

The haze vertical distribution was retrieved from a continuous a priori profile with Neme-

sis. The input parameters in the retrieval code relevant for the analysis of Saturn’snear-IR spectrum are: the extinction coefficient, the scattering albedo, and the threeHenyey-Greenstein coefficients f , g1, and g2. Each parameter is expected to vary withwavelength; so assuming that there are N wavelengths in the spectrum, the retrieval prob-lem has 5N free parameters. It is unlikely that all sets of parameters would fit the VIMSspectrum, but it is practically impossible to try all of them4. Furthermore, the problemmight be highly degenerate and so many solutions could potentially fit the data. This

4The full 0.8-3.5 µm VIMS spectrum was retrieved. There are 146 wavelengths in VIMS spectrum,which gives more than 730 degrees of freedom to the retrieval problem for this particular data set. Also,we do not know a priori the values for each parameter which means that there is an infinite number ofparameter sets.


is even more true for a large spectrum spread over many wavelengths, such as Saturn’sreflection spectrum. A high number of measurements further increases the constraints onthe retrieval solutions but also introduces a wavelength dependency of the parameters inthe model. Finally, very little is known about the actual values of the parameters andtheir dependency with wavelength, and so it is very likely that no distinction could bemade amongst all possible solutions.

Another approach was used in this analysis, in order to break the degeneracies inherent tothe retrieval problem. The extinction coefficient, the scattering albedo, and the Henyey-Greenstein coefficients can be pre-calculated in Mie theory assuming a particular sizedistribution and complex refractive index. The number of degrees of freedom then reducesto 1+2N instead of 5N (The aerosol size accounts for one degree of freedom). It can stillbe a large number especially if a full spectral range is retrieved with Nemesis, but moreimportantly there is some information available on these three parameters:

i. Atmospheric aerosols have typical sizes ranging from a few nanometers to a few microns.ii. The imaginary refractive index has a typical value below 1 (see Fig. 4.5).iii. The real refractive index of most materials ranges between 1 and 2.iv. The complex refractive index shows low variability with wavelength and so is uncor-related to the the aerosol size at first order.v. The aerosol size has the most effect on both the amplitude and the shape of the near-IRreflection spectrum.

The first step consisted in fitting the shape of VIMS reflection spectrum. Towards thisend, the synthetic spectrum was calculated with Nemesis for all aerosol sizes rangingfrom 100 nm to 10 µm, and then compared to the VIMS spectrum. The real refractiveindex was set up to 1.4, which is a typical order value shared by many materials. Sincethe range of physical values for the imaginary part of the refractive index is typicallymuch larger than for the real part, both the imaginary refractive index and the aerosolsize were retrieved simultaneously5 (See Fig. 4.2).

5In the ideal case, the real refractive index, the imaginary refractive index, and the aerosol size shouldbe varied all simultaneously. Although this is actually a work in progress, the time order for runningNemesis on the full 0.8-3.5 µm spectrum is around one day.


Variation of the Goodness-of-fit with Aerosol Size and Imaginary Refractive Index

10-7 10-6 10-5 10-4 10-3 10-2 10-1

Imaginary refractive index

0.1

1.0

10.0

Aer

osol

size

9

9

10

10

15

15

15

20

20

20

30

30

30

30

40

40

40

50

50

50

50

60

60

60

60

70

70

80

80

90

90

Variation of the Goodness-of-fit with Aerosol Size and Real Refractive Index

1.00 1.07 1.15 1.23 1.32 1.41 1.52 1.62 1.74 1.87 2.00Real refractive index

0.1

1.0

10.0

Aer

osol

size

9

9

10

10

10

15

15

15 15

20

20

20

20

30

3030

40

40 40

40

40

50 50

50

50

5060

60

60

60

60

70 70

70

70

70

80 80

80

80

80

90 90

90

90

90

100 10

0

100

110 11

0

Retrieval of Aerosol Vertical Profile from VIMS Near-IR Reflection Spectrum

0.80 1.07 1.33 1.60 1.86 2.13 2.39 2.66 2.92 3.19 3.45Wavelength (µm)

10

20

30

40

50

Radi

ance

(µW

cm

-2 sr

-1 µ

m-1)

Measured by VIMSModelled with NEMESIS

HAZE MODEL*************particle size = 700 nmrefractive indice = 1.4 + 0.001*i

10-4 10-3 10-2 10-1 100 101 102 103 104

Aerosol concentration (cm-3)

10.000

1.000

0.100

0.010

0.001

Pres

sure

(bar

)

Figure 4.2: Retrieval of optical and physical haze properties. Top-left : Goodness-of-fit asa function of aerosol radius and imaginary refractive index. The real index is set to 1.4.Parameters for which the χ2/nλ value is above 100 are ignored since optimal fit is foundfor values below ∼ 10. Top-right : Goodness-of-fit as a function of aerosol radius andreal refractive index. The imaginary index is set to 0.001. Middle: Synthetic spectrumcalculated with Nemesis for aerosol size of 700 nm and a refractive index of 1.4 + 0.001i.Bottom: Retrieved haze vertical distribution from a continuous a priori profile (solid line).The retrieval errors are shown in dotted lines.


By plotting all the synthetic spectra against the VIMS spectrum, it can be easily seen thatany set of parameters with χ2/nλ values above 10 cannot properly fit the observations(See Fig. 4.3). Therefore, the optimum fit to the VIMS spectrum fit is obtained for anarrow range of aerosol sizes between 600 nm and 1.5 µm and for any imaginary refractiveindex below 0.001. The low value for the imaginary refractive index corresponds to highsingle scattering albedo values. Indeed, the imaginary index represents the absorption ofsunlight by scatterers, and so low absorption corresponds to high scattering strength fora given extinction coefficient. The single scattering albedo was found to vary significantlywith wavelengths between 0.8 and 3.5 µm, though most values are above 0.90. Thefact that the single scattering albedo is a variable function of the wavelength shows thedifficulty of choosing this parameter as a free variable in the model. The fitting strategyused in this analysis of VIMS reflection spectrum has proved to be efficient in reducingthe range of possible solutions to the retrieval problem, and especially strong constraintson the aerosol size were obtained6. It has been mentionned that the real refractive indexwas set to the constant value of 1.4. Although most materials have typical values around1.4 for the real index, the use of a specific value for this parameter implies that the natureof scatterers is known a priori. The synthetic spectrum was calculated for real refractiveindexes between 1 and 2 and for particle sizes between 100 nm and 10 µm. The imaginaryrefractive index was set to 0.001. The best fits to VIMS spectrum were obtained for realindexes between 1.1 and 1.8 and for particle sizes between 600 nm and 2 µm. A similarrange for the aerosol size has been retrieved while varying the real and imaginary partsof the complex refractive index. Thus the particle size is well constrained by the analysisof Saturn’s near-IR spectrum, and is quite uncorrelated with both the real and imaginaryindexes.

Another way to break some degeneracies in the model is to consider the optical depth.For each set of parameters, a different vertical profile for the aerosol concentration isfound and so this actually adds a new degenerated parameter in the model7. For example,aerosol concentrations for particle sizes between 600 nm and 1.5 µm have a similar verticalstructure, but with different absolute values especially at the peak around 200 mbars. Thisresult is somewhat expected since larger particles have a higher extinction cross-section,and so fewer particles/cm−3 are needed to fit the observations. However the aerosolconcentration and the extinction coefficient can be combined into the optical depth andit has been found that this parameter is similar for all sets of possible solutions (See Fig.4.3). Below 200 mbars, the optical depth becomes larger than 10, and so no informationon the atmospheric structure can be retrieved at these pressure levels. At the peak,the optical depth ranges between 2 and 3 and shows little variation with wavelength.Consequently, the two main parameters that can be the most accurately retrieved fromVIMS near-IR reflection spectrum are so far the aerosol size and the optical depth.

6The physical explanation for the narrow range of aerosol size solutions is currently under investigation.Preliminary tests have shown so far that particles with backward scattering (small particles) or with strongforward scattering (large particles) cannot fit properly the observations.

7It actually adds a number of degenerated parameters equal to the number of vertical levels chosen inthe model


PRESSURE = 0.00126525 bar

0.5 1.0 1.5 2.0 2.5 3.0 3.5 WAVELENGTH (MICRONS)

0.0

0.1

0.2

0.3

0.4

OPT

ICA

L D

EPTH



0.0

0.5

1.0

1.5

2.0

OPT

ICA

L D

EPTH



0

5

10

15O

PTIC

AL

DEP

TH



0

5

10

15

20

OPT

ICA

L D

EPTH



0

5

10

15

20

OPT

ICA

L D

EPTH



0

5

10

15

OPT

ICA

L D

EPTH



020406080

100120

OPT

ICA

L D

EPTH



020406080

100120

OPT

ICA

L D

EPTH



020406080

100120

OPT

ICA

L D

EPTH

400 nm 500 nm 600 nm 700 nm 800 nm 900 nm 1 µm

Figure 4.3: Retrieval of opacity for various aerosol sizes. Top: Synthetic (red lines)and VIMS (blue line) spectra for particle sizes between 400 nm and 10 µm. A complexrefractive index of 1.4+0.001i was used to generate the extinction, scattering albedo andHenyey-Greenstein phase functions. Middle: Aerosol profiles retrieved with Nemesis foreach particle size. Bottom: Cumulative optical depth calculated from the top of theatmosphere to various pressure levels.


The dependency of the real refractive number with wavelength, and both the imaginaryindex and aerosol size is still work in progress. This should be particularly relevant sincethe nature of scatterers can be inferred from their real refractive indexes8. Preliminaryresults are shown on Fig. 4.4. The haze is assumed to be made of ammonia particles andso the complex refractive indexes in the near-IR are included in Nemesis (Martonchik etal., 1984). The synthetic spectrum is calculated for various particle sizes ranging from 100nm to 10 µm and then compared to the VIMS spectrum. The best fits to VIMS data arefound for particle sizes between 500 nm and 2 µm, which are similar values to the optimumsizes retrieved for a constant refractive index. The particle size is still well constrained,even though the complex refractive index varies with wavelength. The real and imaginaryindexes of ammonia particles have typical values similar to those that have been found tofit the VIMS spectrum. Thus the synthetic spectrum for an ammonia haze matches quitewell the VIMS reflection spectrum. However the haze formation is a complex processinvolving many different components and the retrieval problem is degenerate, so it isvery unlikely that ammonia particles only account for the tropospheric haze in Saturn’satmosphere.

Figure 4.4: Retrieval of particle size for an ammonia haze

8Even though scatterers can be identified by their real imaginary part, it becomes harder if manydifferent types of scatterers are present in Saturn’s atmosphere. Haze formation is a complex processinvolving the generation and destruction of many atmospheric components, so the real index of the"mixture” might hide the individual indexes of each component.


The most interesting result from this preliminary analysis is that the variation of complexrefractive index with wavelength has little effect on the retrieval of the particle size. Thereal and imaginary indexes between 1 and 5 µm of a few haze candidates are shown onFig. 4.5. The variations with wavelength of the complex index in the near-IR are quitesignificant for the imaginary part but are small for the real part. Therefore it is verylikely that the imaginary index of aerosols should vary greatly over this spectral range,while the real part can be considered as a constant function of the wavelength.

Figure 4.5: Complex refractive indexes of haze candidates


4.3 Vertical Structure

4.3.1 Sensitivity Tests

In the previous section, it has been mentioned that Nemesis retrieved the haze verticaldistribution from a continuous a priori profile. A series of retrievals was undertaken formany a priori values and errors. The shape of the retrieved profile turned out to bealways similar to the results shown on Figs. 4.2 and 4.3. This also true for all sets ofparticle sizes and refractive indexes that are solutions of the retrieval problem. The mainfeatures of the the haze profile retrieved from Saturn’s near-IR reflection spectrum are:

i. A peak of aerosol concentration at 200 mbarii. A sharp decrease of aerosol concentration above the peak until 10 mbar.iii. An increase of aerosol concentration above 10 mbar.iv. A decrease of aerosol concentration below the peak. The retrieved profile relaxes tothe a priori profile.

Sensitivity tests on the haze pressure levels (see Fig. 4.6) show that the reflection spectrumis slightly sensitive to the haze distribution above 10 mbar. In particular, opacity is neededin the upper stratosphere in order to fit the low signal in the absorption bands. Howeverthe effect on the continuum can be completely neglected, and so the peak at 200 mbarsis contributing most of the reflectance in the near-infrared. Pressure levels between 10and 200 mbar do not contribute to the reflection peaks. This is because the aerosolconcentration is too low to have significant effect on to the reflectance. As expected, thepeak of opacity at 200 mbar contributes the most the VIMS spectrum. The sharp decreaseof Mie scatterer abundance right above the opacity peak is referred to as a “clearing”. Theexistence of a clearing above the opacity peak has been tested by considering differentsmodels for the haze vertical structure.

Sensitivity of the Near-IR Reflection Spectrum to the Haze’s Top pressure Level

0.0010 0.0025 0.0063 0.0159 0.0400 0.1004 0.2523 0.6340 1.5930 4.0023Top pressure (bar)

0

20

40

60

80

100

χ2

Sensitivity of the Near-IR Reflection Spectrum to the Haze’s Bottom pressure Level

4.002 2.463 1.516 0.933 0.574 0.353 0.217 0.134 0.082 0.051Bottom pressure (bar)

0

20

40

60

80

χ2

Figure 4.6: Sensitivity of Saturn’s reflection spectrum to haze pressure levels. Left : Thesensitivity to the haze’s top pressure level was tested by removing all points above thepressure levels shown on the x-axis. Right : The same process but for pressure levels belowthose on the x-axis. Note that both figures are not smooth at the concentration peak,implying that a thin haze layer contributes to the reflection spectrum.


4.3.2 Opacity clearing

The existence of the opacity clearing has been tested by considering the following verticalhaze profiles:

i. A compact profile: the opacity of a single aerosol layer is retrieved for all presure levels.ii. A block profile: The aerosols have a constant density between two layers, and are absentelsewhere. All sets of bottom and top pressure levels have been tested in Nemesis.The particles are assumed to have a size of 700 nm and a complex refractive index of1.4+0.001. For both models, a peak of opacity between 100 and 200 mbar is necessary tofit the VIMS spectrum (see Fig. 4.7). There is a very tight constraint on the top pressurelevel for the block model. Therefore, any synthetic spectrum calculated for a haze profilewith a high opacity above the tropopause cannot fit properly the observations. This isalso true for a haze vertical profile extending down to the deep troposphere. Therefore,the results from the block haze model suggest that the clearing above the tropopause isreal

Tests on Haze Vertical Profile: Compact Aerosol Layer

0.001 0.010 0.100 1.000 10.000Pressure base (bar)

0

20

40

60

80

100

χ2

pB = 0.16 atm

0.80 0.96 1.11 1.27 1.42 1.58 1.74 1.89 2.05 2.20 2.36 2.51 2.67 2.83 2.98 3.14 3.29 3.45Wavelength (µm)

10

20

30

40

50

Radi

ance

(µW

cm

-2 sr

-1 µ

m-1)

HAZE MODEL*************compact: pB = 0.16 barradius = 700 nmn = 1.4 + 0.001i

Tests on Haze Vertical Profile: Extended Aerosol Layer

0.01 0.10 1.00Bottom pressure (bar)

1.000

0.100

0.010

0.001To

p pr

essu

re (b

ar)

30

40

40

50

50

60

60

70

70

70

80

80

80

90

90

90

Figure 4.7: Tests on the haze vertical profile. Left: Compact aerosol layer. right: Theaerosol concentration is constant within two boundary layers, and is assumed to be zeroelsewhere.

Chapter 5

Conclusion and Future Work

VIMS thermal and reflection spectra in the near-infrared offer a new insight on the cloudsand haze structure in Saturn’s troposphere. We first started the analysis of VIMS databy looking at single thermal spectra acquired at a specific viewing geometry. We foundthat the best fits to the observations were obtained for a cloud base pressure at 2.7 bar.Nevertheless, we did not manage to distinguish between a purely absorptive and scatter-ing atmospheric model. Also, the aerosol particle size is poorly constrained, and manysizes can possibly provide a satisfying solution to the retrieval problem. It is necessary toknow the vertical structure of aerosols in order to properly calculate the optical depth.We found that both a compact and an extended cloud model could fit equally well VIMSthermal spectra, and so the optical depth values retrieved from these models differ greatly.Finally, we could not constrain either the cloud composition. The two species expectedto condense at the tropospheric levels considered here, ammonia and ammonium hydro-sulphide, were included in the atmospheric model. However, no model provided a betterfit than the other, and so we could not distinguish between both cases. The condensateof ammonia and ammonium hydrosulphide is predicted by thermochemical theories, how-ever there has been no spectral identification of these species yet. Therefore, we cannotexclude the possibility of aternative cloud compositions in Saturn’s troposphere. In con-clusion, there are many free parameters in the tropospheric cloud model and not enoughmeasurements constraints in order to find a unique solution to the retrieval problem.

We managed to break some degeneracies inherent to the retrieval problem by consideringdifferent observations of the same area of the planet but at various viewing geometries.In particular, we carried out the analysis of VIMS spectra simultaneously at low and highemission angles. We found that a purely absorptive atmosphere cannot properly fit VIMSdata and that scattering needs to be integrated in the numerical code Nemesis. Also,we managed to exclude the simple cloud model from our set of possible solutions to theretrieval problem. This applies to a compact cloud made of either ammonia or ammoniumhydrosulphide particles. We could match the observations quite well for an extended cloudwith a base presure higher than 2.5 bar and an extension of at least 25 kilometres, and fora particle size ranging between 1 and 3 micrometres. Nevertheless, a minimum extensionof 25 kilometres implies tropospheric clouds that extend up to altitude levels significantly

43

CHAPTER 5. CONCLUSION AND FUTURE WORK 44

higher than what is predicted from thermochemical models. Even though the minimumextension found here is quite unrealistic, it shows that we can fit VIMS observations withsome opacity sources located higher up in the troposphere. This was indeed confirmed byconsidering first a continuous aerosol profile and then a two compact clouds model. Inboth cases we retrieved two opacity sources: one deep tropospheric cloud between 2 and4 bar, and one haze layer below the tropopause. We used the VIMS spectum of Saturnat shorter wavelengths in order to investigate the properties of this haze layer.

The near-infrared spectrum of Saturn in the 0.8-3.5 µm wavelength range was used toretrieve the optical and physical properties, and the vertical structure of the tropospherichaze located above the condensate clouds. We retrieved a peak of opacity below thetropopause at around 200 mbar. Sensitivity tests showed that the opacity clearing abovethe peak is very likely to be a realistic feature in Saturn’s atmosphere. Nevertheless, thecumulative optical depth increases quickly below the peak and so we cannot sound thehaze structure at higher pressure levels. We also found that the real part of the refractiveindex should be between 1.1 and 1.8 and the imaginary part below 0.001 in order to machVIMS observations. We managed to constrain quite well the particle size range between600 nm and 2 µm. We are currently applying to VIMS reflection spectra the same multi-viewing analysis as described for thermal spectra, in order to narrow the range of possiblesolutions to the retrieval problem.

As mentioned above, the next step is to carry out a multi-analysis of VIMS reflectionspectra at various viewing geometries. In that way, we expect to derive a model ofSaturn’s upper tropospheric haze coherent with VIMS data and with the lowest level ofdegeneracy. The results from the analysis of VIMS reflection spectra will be combined tothermal spectra in order to better constrain the condensate cloud properties and verticalstructure. Indeed, by the setting up the haze model in the numerical code, the analysisof Saturn’s thermal spectum has less free parameters and so we should be able to excludemany cloud models that were found to fit the observations initially. Once the analysisof VIMS near-infrared spectra provides a coherent clouds and haze model in Saturn’stroposphere, the final step will be to investigate spatial and time variations of troposphericaerosols, and search for a correlation between the temperature knee and the presence oftropospheric haze.

CHAPTER 5. CONCLUSION AND FUTURE WORK 45

Schedule for the next academic year

Term PlanSeptember 2011 - Multi-viewing analysis of VIMS reflection spectra. I

am currently analysing reflection spectra measured atdifferent emission and incident angles.- Derive a coherent tropospheric haze model includingthe vertical structure, particle size, opacity, and complexrefractive index of aerosols.- Search for potential haze candidates by comparing lab-oratory measurements with the haze model retrievedfrom the multi-viewing analysis.

October 2011 - Test the possibility of a temperature knee generatedby tropospheric haze.- Investigate the spatial and time variations of the kneefrom CIRS observations, and look for possible correla-tions with the tropospheric haze.

November 2011 - Multi-viewing analysis of thermal spectra assuming ahaze model retrieved from the spectral analysis in the0.8-3.5 µm wavelength range.

December 2011 - Use the clouds and haze model coherent with the wholenear-infrared spectrum to investigate spatial and timevariations of tropospheric aerosols.

January 2012 - Dynamical study from the variations in clouds andhaze structure over space and time.

February 2012 - Theoretical study of the photochemical and microphys-ical processes involved in the formation of troposphericclouds and haze.

March-May 2012 - Write.June 2012 - Submit.July 2012 - Work.

Bibliography

[1] Achterberg, R. K., Conrath, B. J., and Gierasch, P. J. Cassini CIRSretrievals of ammonia in Jupiter’s upper troposphere. Icarus 182 (2006), 169–180.

[2] Andrews, D. G. Atmospheric physics. Cambridge U. P., 2000.

[3] Apt, J., and Singer, R. B. Cloud height differences on Saturn. Icarus 52 (1982),503–508.

[4] Atreya, S., and Wong, A. Coupled clouds and chemistry of the giant planets -A case for multiprobes. Space Science Reviews 116 (2004), 121–136.

[5] Atreya, S. K. Atmospheres and ionospheres of the outer planets and their satel-lites. Springer, Berlin, 1986.

[6] Atreya, S. K., Donahue, T. M., and Kuhn, W. R. The distribution of am-monia and its photochemical products on Jupiter. Icarus 31 (1977), 348–355.

[7] Atreya, S. K., Kuhn, W. R., and Donahue, T. M. Saturn: Troposphericammonia and nitrogen. Geophys. Res. Lett. 7 (1980), 474–476.

[8] Atreya, S. K., Wong, A. S., Baines, K. H., Wong, M. H., and Owen, T. C.

Jupiter’s ammonia clouds - localized or ubiquitous ? Planetary and Space Science53 (2005), 498–507.

[9] Atreya, S. K., Wong, M. H., Owen, T. C., Mahaffy, P. R., Niemann,

H. B., de Pater, I., Drossart, P., and Encrenaz, T. A comparison of theatmospheres of Jupiter and Saturn: deep atmopsheric composition, cloud structure,vertical mixing, and origin. Plan. & Space Sci. 47 (1999), 1243–1262.

[10] Aurnou, J., Heimpel, M., Allen, L., King, E., and Wicht, J. Convectiveheat transfer and the pattern of thermal emission on the gas giants. Geophys. J.Int. 173 (2008), 793–801.

[11] Baines, K. H., Carlson, R. W., and Kamp, L. W. Fresh ammonia ice cloudsin Jupiter. I. Spectroscopic identification, spatial distribution, and dynamical im-plications. Icarus 159 (2002), 74–94.

46

BIBLIOGRAPHY 47

[12] Baines, K. H., Delitsky, M. L., Momary, T. W., Brown, R. H., Buratti,

B. J., Clark, R. N., and Nicholson, P. D. Storm clouds on Saturn: Lightning-induced chemistry and associated materials consistent with Cassini/VIMS spectra.Planetary and Space Science 57 (2009), 1650–1658.

[13] Baines, K. H., Drossart, P., Momary, T. W., Formisano, V., Griffith,

C., Bellucci, G., Bibring, J. P., Brown, R. H., Buratti, B. J., Capac-

cioni, F., Cerroni, P., Clarck, R. N., Coradini, A., Combes, M., Cruik-

shank, D. P., Jaumann, R., Langevin, Y., Matson, D. L., Mccord, T. B.,

Mennella, V., Nelson, R. M., Nicholson, P. D., Sicardy, B., and Sotin,

C. The Atmospheres of Saturn and Titan in the Near-Infrared: First Results ofCassini/VIMS. Earth, Moon, and Planets (2006), 1–29.

[14] Baines, K. H., Momary, T. W., Fletcher, L. N., Showman, A. P., Roos-

Serote, M., Buratti, B. J., Clark, R. N., and Nicholson, P. D. Saturn’snorth polar cyclone and hexagon at depth revealed by Cassini/VIMS. Plan. & SpaceSci. 57 (2009), 1671–1681.

[15] Baines, K. H., Momary, T. W., and Roos-Serote, M. North vs South onSaturn: Discovery of a Pronounced Hemispherical Asymmetry in Saturn’s 5-MicronEmission and Associated Deep Cloud Structure by Cassini/VIMS. B.A.A.S. 38(2006), 488.

[16] Banfield, D., Barney, J. C., Gierasch, P. J., and Nicholson, P. D. Near-IR spectrophotometry of Jovian aerosols - Meridional and vertical distributions.Icarus 134 (1998), 11–23.

[17] Banfield, D., Gierasch, P. J., Squyres, S. W., and Nicholson, P. D. 2µmspectrophotometry of Jovian Stratospheric aerosols - Scattering opacities, verticaldistributions, and wind speeds. Icarus 121 (1996), 389–410.

[18] Barnet, C. D., Beebe, R. F., and Conrath, B. J. A seasonal radiative-dynamic model of Saturn’s troposphere. Icarus 98 (1992), 94–107.

[19] Bergstralh, J. T., Orton, G. S., Diner, D. J., Baines, K. H., Neff, J. S.,

and Allen, M. A. Spatially resolved absolute spectrophotometry of Saturn: 3390-8080 Å. Icarus 46 (1981), 27–39.

[20] Bezar, B., Baluteau, J. P., and Marten, A. Study of the deep cloud structurein the equatorial region of Jupiter from Voyager infrared and visible data. Icarus54 (Jun. 1983), 434–455.

[21] Bezard, B., Drossart, P., Lellouch, E., Tarrago, G., and Maillard,

J. P. Detection of arsine in Saturn. Astrophys. J. 346 (1989), 509–513.

[22] Bezard, B., Gautier, D., and Conrath, B. J. A seasonal model of the Satur-nian upper troposphere: Comparison with Voyager infrared measurements . Icarus60 (1984), 274–288.

BIBLIOGRAPHY 48

[23] Birnbaum, G., Borysow, A., and Orton, G. S. Collision-induced absorp-tion of H2-H2 and H2-He in the rotational and fundamental bands for planetaryapplications. Icarus 123 (1996), 4–22.

[24] Bjoraker, G. L., Larson, H. P., and Fink, U. A study of ethane on Saturnin the 3 µm region. Astrophys. J. 248 (1981), 856–862.

[25] Bohren, C. F., and Huffman, D. R. Absorption and scattering of light by smallparticles. Wiley, New York, 1983.

[26] Borysow, A. Modelling of collision-induced infrared-absorption spectra of H2-H2

pairs in the fundamental band at temperatures from 20 to 300K. Icarus 92 (1991),273–279.

[27] Borysow, A. Modelling of collision-induced infrared-absorption spectra of H2-H2

pairs in the fundamental band at temperatures from 20 to 300K. Icarus 106 (1991),614.

[28] Borysow, A., and Frommhold, L. Collision-induced rototranslational absorp-tion spectra of CH4-CH4 pairs at temperatures from 50 to 300K. Astrophys. J. 318(1986), 940–943.

[29] Borysow, A., and Frommhold, L. Theoretical collision-induced rototransla-tional absorption spectra for the outer planets: H2-CH4 pairs. Astrophys. J. 304(1986), 849–865.

[30] Borysow, A., Frommhold, L., and Birnbaum, G. Collision-induced roto-translational absorption spectra of H2-He pairs at temperatures from 40 to 300K.Astrophys. J. 326 (1988), 509–515.

[31] Bowles, N., Calcutt, S., Irwin, P., and Temple, J. Band parameters forself-broadened ammonia gas in the range 0.74 to 5.24 µm to support measurementsof the atmosphere of the planet Jupiter. Icarus 196 (2008), 612–624.

[32] Briggs, F. H., and Sackett, P. D. Radio observations of Saturn as a probe ofits atmosphere and cloud structure. Icarus 80 (1989), 77–103.

[33] Brooke, T., Knacke, R. F., Encrenaz, T., drossart, P., Crisp, D., and

Feuchtgruber, H. Models of the ISO 3-µm reflection spectrum of Jupiter . Icarus136 (1998), 1–13.

[34] Brown, R. H., Baines, K. H., Bellucci, G., Bibring, J.-P., Buratti, B. J.,

Capaccioni, F., Cerroni, P., Clark, R. N., Coradini, A., Cruikshank,

D. P., Drossart, P., Formisano, V., Jaumann, R., Langevin, Y., Mat-

son, D. L., Mccord, T. B., Mennella, V., Miller, E., Nelson, R. M.,

Nicholson, P. D., Sicardy, B., and Sotin, C. The Cassini Visual And In-frared Mapping Spectrometer (Vims) Investigation. Space Science Reviews 115(2004), 111–168.

BIBLIOGRAPHY 49

[35] Burgdorf, M. J., Orton, G. S., Encrenaz, T., Davis, G. R., Lellouch,

E., Sidher, S. D., and Swindyard, B. M. Far-infrared spectroscopy of the giantplanets: measurements of ammonia and phosphine at Jupiter and Saturn and thecontinuum of Neptune. Advances in Space Research 34 (2004), 2247–2250.

[36] Buriez, J. C., and de Bergh, C. A study of the atmospheres of Saturn basedon methane line profile near 1.1-µm. Astron. Astrophys. 94 (1981), 382–390.

[37] Butler, R., Sagui, L., Kleiner, I., and Brown, L. The absorption spec-trum of phosphine (ph3) between 2.8 and 3.7 µm: Line positions, intensities, andassignments. Journal of Molecular Spectroscopy 238 (2006), 178–192.

[38] Carlson, B. E., Rossow, W. B., and Orton, G. S. Cloud microphysics in thegiant planets. J. Atmos. Sci. 45 (1988), 2066–2081.

[39] Chandrasekhar, S. Radiative transfer. Oxford University Press, 1950.

[40] Clark, R. N. Planetary reflectance measurements in the region of planetary ther-mal emission. Icarus 40 (1979), 94–103.

[41] Clark, R. N., and McCord, T. B. Jupiter and Saturn: Near-infrared spectralalbedos. Icarus 40 (1979), 180–188.

[42] Colina, L., Bohlin, R. C., and Castelli, F. The 0.12-2.5 micron absoluteflux distribution of the Sun for comparison with solar analog stars. Astron. J. 112(1996), 307–315.

[43] Conrath, B. J., and Gautier, D. Saturn Helium Abundance: A Reanalysis ofVoyager Measurements. Icarus 144 (2000), 124–134.

[44] Conrath, B. J., and J., G. P. Retrieval of ammonia abundances and cloudopacities on Jupiter from Voyager IRIS spectra. Icarus 67 (1986), 456–483.

[45] Conrath, B. J., and Pirraglia, J. A. Thermal structure of Saturn from Voyagerinfrared measurements - Implications for atmospheric dynamics. Icarus 53 (1983),286–292.

[46] Courtin, R., Gautier, D., Marten, A., Bézard, B., and Hanel, R. Thecomposition of Saturn’s atmosphere at northern temperate latitude from VoyagerIRIS spectra - NH3, PH3, C2H2, C2H6, CH3D, CH4, and the Saturnian D/H isotopicratio. Astrophys. J. 287 (1984), 899–916.

[47] Craig, I. J. D., and Brown, J. C. Inverse problems in Astronomy: A guide toinversion strategies for remotely sensed data. Hilger, Bristol, 1986.

[48] Dana, V., Mandin, J.-Y., Tarrago, G., Olson, W. B., and Drossart, P.

Absolute infrared intensities in the fundamentals ν1 and ν3 of arsine. Journal ofmolecular spectroscopy 159, 22 (1993), 468–480.

BIBLIOGRAPHY 50

[49] Davis, G. R., Griffin, M. J., Naylor, D. A., Oldham, P. G., Swinyard,

B. M., Ade, P. A. R., Calcutt, S. B., Encrenaz, T., de Graauw, T.,

Gautier, D., Irwin, P. G. J., Lellouch, E., Orton, G. S., Armand, C.,

Burgdorf, M., di Giorgio, A., Ewart, D., Gry, C., King, K. J., Lim, T.,

Molinari, S., Price, M., Sidher, S., Smith, A., Texier, D., Trams, N.,

and Unger, S. J. ISO LWS measurement of the far-infrared spectrum of Saturn.Astron. Astrophys. 315 (1996), 393–396.

[50] de Graauw, T., Feuchtgruber, H., Bézard, B., Drossart, P., Encre-

naz, T., Beintema, D. A., Griffin, M., Heras, A., Kessler, M., Leech,

K., Morris, P., Roelfsema, P. R., Roos-Serote, M., Salama, A., Van-

denbussche, B., Valentijn, E. A., Davis, G. R., and Naylor, D. A. Firstresults of ISO-SWS observations of Saturn: detection of CO2, CH3C2H, C4H2 andtropospheric H2O . Astron. Astrophys. 321 (1997), L13–L16.

[51] de Haan, J. F., Bosma, P. B., and Hovenier, J. W. The adding method formultiple scattering calculations of polarized light. Astron. Astrophys. 183 (1987),371–391.

[52] Del Genio, A. D., Barbara, J. M., Ferrier, J., Ingersoll, A. P., West,

R. A., Vasavada, A. R., Spitale, J., and Porco, C. C. Saturn eddy momen-tum fluxes and convection: First estimates from Cassini images. Icarus 189 (2007),479–492.

[53] Diner, D. J., and Martonchik, J. V. Atmospheric transfer of radiation abovean inhomogeneous non-lambertian reflective ground - I. Theory. J. Quant. Spectrosc.Radiat. Transfer 31, 2 (1984), 97–125.

[54] Dougherty, M. K., Esposito, L. W., and Krimigi, S. M. Saturn fromCassini-Huygens. Springer, 2009.

[55] Dunder, T., Clapp, M. L., and Miller, R. E. Infrared spectroscopy of ho-mogeneously nucleated hydrazine aerosols: Disordered and crystalline phases. J.Geophys. Res. 98 (1993), 1213–1221.

[56] Encrenaz, T., Drossart, P., Feuchtgruber, H., Lellouch, E., Bézard,

B., Fouchet, T., and Atreya, S. K. The atmospheric composition and structureof Jupiter and Saturn from ISO observations: A preliminary review. Planet. SpaceSci. 47 (1999), 1225–1242.

[57] Fegley, B. J., and Lodders, K. Chemical models of the deep atmospheres ofJupiter and Saturn. Icarus 110 (1994), 117–154.

[58] Ferraro, J. R., Sill, G., and Fink, U. Infrared Intensity Measurements ofCryodeposited Thin Films of NH3, NH4HS, H2S, and Assignments of AbsorptionBands. Applied Spectroscopy 34 (1980), 525–533.

BIBLIOGRAPHY 51

[59] Fink, U., Larson, H. P., Bjoraker, G. L., and Johnson, J. R. The NH3

spectrum in Saturn’s 5 micron window. ApJ 268 (May 1983), 880–888.

[60] Flasar, F. M., Kunde, V. G., Abbas, M. M., K., A. R., Ade, P., Barucci,

A., Bézard, B., Bjoraker, G. L., Brasunas, J. C., Calcutt, S., Carlson,

R., Césarsky, C. J., Conrath, B. J., Coradini, A., Courtin, R., Couste-

nis, A., Edberg, S., Edgington, S., Ferrari, C., Fouchet, T., Gautier,

D., Gierasch, P. J., Grossman, K., Irwin, P., Jennings, D. E., Lellouch,

E., Mamoutkine, A. A., Marten, A., Meyer, J. P., Nixon, C. A., Orton,

G. S., Owen, T. C., Pearl, J. C., Prangé, R., Raulin, F., Read, P. L.,

Romani, P. N., Samuelson, R. E., Segura, M. E., Showalter, M. R.,

Simon-Miller, A. A., Smith, M. D., Spencer, J. R., Spilker, L. J., and

Taylore, F. W. Exploring The Saturn System In The Thermal Infrared: TheComposite Infrared Spectrometer. Space Science Reviews 115 (2004), 169–297.

[61] Fletcher, L. N. Saturn’s Atmosphere: Structure and Composition fromCassini/CIRS. PhD thesis, University of Oxford, 2007.

[62] Fletcher, L. N., Achterberg, R. K., K., G. T., Orton, G. S., Conrath,

B. J., Simon-Miller, A. A., Teanby, N., Guerlet, S., Irwin, P. G. J.,

and Flasar, F. M. Seasonal change on Saturn from Cassini/CIRS observations,2004-2009. Icarus 208 (2010), 337–352.

[63] Fletcher, L. N., Baines, K. H., Momary, T. W., Orton, G. S., Roose-

Serote, M., and Irwin, P. G. J. Saturn’s regional Properties frm Cassini/VIMS4.5-5.1 µm Spectroscopy. DPS conference, 2009.

[64] Fletcher, L. N., Baines, K. H., Momary, T. W., Showman, A. P., Ir-

win, P. G. J., Orton, G. S., Roos-Serote, M., and Merlet, C. Saturn’stropospheric composition and clouds from Cassini/VIMS 4.6-5.1 µm nightside spec-troscopy. Icarus 214 (2011), 510–533.

[65] Fletcher, L. N., Irwin, P. G. J., Teanby, N. A., Orton, G. S., Parrish,

P. D., Calcutt, S. B., Bowles, N., de Kok, R., Howett, C., and Taylor,

F. W. The meridional phosphine distribution in Saturn’s upper troposphere fromCassini/CIRS observations. Icarus 188 (2007), 72–88.

[66] Fletcher, L. N., Irwin, P. G. J., Teanby, N. A., Orton, G. S., Parrish,

P. D., de Kok, R., Howett, C., Calcutt, S. B., Bowles, N., and Taylor,

F. W. Characterising Saturn’s vertical temperature structure from Cassini/CIRS.Icarus 189 (2007), 457–478.

[67] Fletcher, L. N., Orton, G. S., Teanby, N. A., and Irwin, P. G. J. Phos-phine on Jupiter and Saturn from Cassini/CIRS. Icarus 202 (2009), 543–564.

BIBLIOGRAPHY 52

[68] Fletcher, L. N., Orton, G. S., Teanby, N. A., Irwin, P. G. J., and Bjo-

raker, G. L. Methane and its isotopologues on Saturn from Cassini/CIRS obser-vations. Icarus 199 (2009), 351–367.

[69] Franz, O. G., and Price, M. J. Uranus: Limb and polar brightening at 7300Å. The Astrophysical Journal 214 (1977), 145–146.

[70] Friedlander, S. K. Smoke, dust and haze. Wiley, New York, 1977.

[71] Friedson, A. J., Wong, A., and Yung, Y. L. Models for polar haze formationin Jupiter’s stratosphere. Icarus 158 (2002), 389–400.

[72] Fuchs, N. A. The mechanics of aerosols. Pergamon, New York, 1964.

[73] Gaddis, L., Anderson, J., Becker, T., Cook, D., Edwards, K., Elia-

son, E., Hare, T., Kieffer, H., Lee, E. M., Mathews, J., Soderblom, L.,

Sucharski, T., Torson, J., McEwen, A., and Robinson, M. An overview ofthe integrated software for imaging spectrometers (ISIS). In Lunar and planetaryinstitute science conference abstracts (1997), vol. 28, p. 387.

[74] Galilei, G. Le Opere di Galileo Galilei. Edizione Nazionale, 1890-1909.

[75] Gehrels, T., and Matthews, M. S. Saturn. University of Arizona Press, 1984.

[76] Gierasch, P. G., Ingersoll, A. P., Banfield, D., Ewald, S. P., Helfen-

stein, P., Simon-Miller, A., Vasavada, A., Breneman, H. H., Senske,

D. A., and the Galileo Imaging Team. Observation of moist convection inJupiter’s atmosphere . Nature 403 (2000), 628–630.

[77] Gierasch, P. J., Conrath, B. J., and Magalhães, J. A. Zonal mean prop-erties of Jupiter’s upper troposphere from Voyager infrared observations. Icarus 67(1986), 456–483.

[78] Gladstone, G. R., Allen, M., and Yung, Y. L. Hydrocarbon photochemistryin the upper atmosphere of Jupiter. Icarus 119 (1996), 1–52.

[79] Goody, R., and Young, Y. L. Atmospheric Radiation (Theoritcal basis), 2nd ed.Oxford University Press, 1989.

[80] Hanel, R., Conrath, B., Flasar, F. M., Kunde, V., Maguire, W., Pearl,

J. C., Pirraglia, J., Samuelson, R., Cruikshank, D. P., Gautier, D.,

Gierasch, P. J., Horn, L., and Ponnamperuma, C. Infrared observations ofthe saturnian system from Voyager 2. Science 215 (1981), 544–548.

[81] Hanel, R., Conrath, B., Flasar, F. M., Kunde, V., Maguire, W., Pearl,

J. C., Pirraglia, J., Samuelson, R., Herath, L., Allison, M., Cruik-

shank, D. P., Gautier, D., Gierasch, P. J., Horn, L., Koppany, R., and

Ponnamperuma, C. Infrared observations of the saturnian system from Voyager1. Science 212 (1981), 192–200.

BIBLIOGRAPHY 53

[82] Hanel, R. A., Conrath, B. J., Jennings, D. E., and Samuelson, R. E.

Exploration of the Solar System by Infrared Remote Sensing, 2nd ed. CambridgeUniversity Press, 2003.

[83] Hanel, R. A., Conrath, B. J., Kunde, V. G., Pearl, J. C., and Pirraglia,

J. A. Albedo, internal heat flux, and energy balance of Saturn. Icarus 53 (1983),262–285.

[84] Hansen, J. E., and Travis, L. D. Light scattering in planetary atmospheres.Space Science Reviews 16 (1974), 527–610.

[85] Houghton, J. The physics of atmospheres, 3rd ed. Cambridge U. P., 2002.

[86] Howett, C. J. A., Carlson, R. W., Irwin, P. G. J., and Calcutt, S. B.

Optical constants of ammonium hydrosulfide ice and ammonia ice. Journal of theOptical Society of America B optical Physics 24 (Jan. 2007), 126–136.

[87] Irwin, P. G. J. Giant Planets of Our Solar System: An Introduction. Springer,2006.

[88] Irwin, P. G. J., Calcutt, S. B., and Taylor, F. W. Radiative transfer modelsfor Galileo NIMS studies of the atmosphere of Jupiter. Adv. Space Res. 19, 8 (1997),1149–1158.

[89] Irwin, P. G. J., and Dyudina, U. The retrieval of cloud structure maps in theequatorial region of Jupiter using a principal component analysys of Galileo/NIMSdata. Icarus 156 (2002), 52–63.

[90] Irwin, P. G. J., Parrish, P., Fouchet, T., Calcutt, S. B., Taylor, F. W.,

Simon-Miller, A. A., and Nixon, C. A. Retrievals of jovian troposphericphosphine from Cassini/CIRS . Icarus 172 (2004), 37–49.

[91] Irwin, P. G. J., Teanbby, N. A., de Kok, R., Fletcher, L. N., Howett,

C. J. A., Tsang, C. C. C., Wilson, C. F., Calcutt, S. B., Nixon, C. A.,

and Parrish, P. D. The NEMESIS planetary atmosphere radiative transfer andretrieval tool. Journal of Quantitative Spectroscopy and Radiative Transfer 109(2008), 1136–1150.

[92] Irwin, P. G. J., Weir, A. L., Smith, F. W., Taylor, F. W., Lambert,

A. L., Calcutt, S. B., Cameron-Smith, P. J., Carlson, R. W., Baines,

K., Orton, G. S., Drossart, P., Encrenaz, T., and Roos-Serote, M.

Cloud structure and atmospheric composition of Jupiter retrieved from Galileo near-infrared mapping spectrometer real-time spectra. Journal of Geophysical Research103, 10 (1998), 23001–23021.

[93] Irwin, P. G. J., Weir, A. L., Taylor, F. W., Calcutt, S. B., and Carlson,

R. W. The origin of belt/zone contrasts in the atmosphere of Jupiter and theircorrelation with 5-µm opacity. Icarus 149 (2001), 397–415.

BIBLIOGRAPHY 54

[94] Jacobson, M. Z. Fundamentals of atmospheric modeling. Cambridge U. P., Cam-bridge, 1999.

[95] Karkoschka, E., and Tomasko, M. Saturn’s vertical and latitudinal cloudstructure 1991-2004 from HST imaging in 30 filters. Icarus 106 (2005), 428.

[96] Karkoschka, E., and Tomasko, M. G. Representation of planetary limb dark-ening data. Bull. Am. Astron. Soc. 17 (1985), 703.

[97] Karkoschka, E., and Tomasko, M. G. Saturn’s upper atmospheric hazesobserved by the Hubble Space Telescope. Icarus 106 (1992), 428.

[98] Karkoschka, E., and Tomasko, M. G. Saturn’s upper troposphere 1986-1989.Icarus 97 (1992), 161–181.

[99] Keesee, R. G. Nucleation and particle formation in the upper atmosphere. J.Geophys. Res. 94 (1989), 14683–14692.

[100] Kerola, D. X., Larson, H. P., and Tomasko, M. G. Analysis of the near-IR spectrum of Saturn: A comprehensive radiative transfer model of its midle andupper troposphere. Icarus 127 (1997), 190–212.

[101] Kim, J. H., Kim, S. J., Geballe, T. R., Kim, S. S., and Brown, L. R. High-resolution spectroscopy of Saturn at 3 microns: CH4, CH3D, C2H2, C2H6, PH3,clouds, and haze. Icarus 185 (2006), 476–486.

[102] Kleiner, I., Tarrago, G., Cottaz, C., Sagui, L., Brown, L. R., Poynter,

R. L., Pickett, H. M., Chen, P., Pearson, J. C., Sams, R. L., Blake,

G. A., Matsuura, S., Nemtchinov, V., Varanasi, P., Fusina, L., and

Di Lonardo, G. NH3 and PH3 line parameters: the 2000 HITRAN update andnew results. Journal of Quantitative Spectroscopy and Radiative Transfer 82 (2003),293–312.

[103] Lacis, A. A., and Oinas, V. A description of the correlated-k distribution methodfor modelling nongray gaseous absorption, thermal emission, and multiple scatteringin vertically inhomogeneous atmospheres. J. Geophys. Res. 96, 15 (1991), 9027–9064.

[104] Larson, H. P. Infrared spectroscopic observations of the outer planets, theirsatellites, and the asteroids. Annual Review of Astronomy and Astrophysics 18(1980), 43–75.

[105] Lewis, J. S. The clouds of Jupiter and the NH3-H2O and NH3-H2S systems. Icarus10 (1969), 365–378.

[106] Lindal, G. F., Sweetnam, D. N., and Eshleman, V. R. The atmosphere ofSaturn: An analysis of the Voyager radio occultation measurements. Astron. J. 90(1985), 1136–1146.

BIBLIOGRAPHY 55

[107] Lodders, K., and Fegley, B. The planetary scientist’s companion. Oxford U.P., Oxford, 1998.

[108] Macke, A., Mueller, J., and Raschke, E. Single scattering properties ofatmospheric ice crystals. J. Atmos. Sci. 53 (1996), 2813–2825.

[109] Marten, A., Rouan, D., Baluteau, J. P., Gautier, D., Conrath, B. J.,

Hanel, R. A., Kunde, V., Samuelson, R., Chedin, A., and Scott, N. Studyof the ammonia ice cloud layer in the Equatorial Region of Jupiter from the infraredinterferometric experiment on Voyager. Icarus 46 (May 1981), 233–248.

[110] Martonchik, J. V., Orton, G. S., and Appleby, J. F. Optical properties ofNH3 ice from the far infrared to the near ultraviolet. Applied Optics 23 (Feb. 1984),541–547.

[111] Massie, S. T., and Hunten, D. M. Conversion of para and ortho Hydrogen inthe Jovian Planets. Icarus 49 (1982), 213–226.

[112] Matson, D. L., J., S. L., and J., L. The Cassini/Huygens Mission to theSaturnian System. Space Science Reviews 104 (2002), 1–58.

[113] McCord, T. B., Coradini, A., Hibbits, C. A., Capaccioni, F., Hansen,

G. B., Filacchione, G., Clark, R. N., Cerroni, P., Brown, R. H., Baines,

K. H., Belluci, G., Bibring, J., Buratti, B. J., Bussoletti, E., Combes,

M., Cruikshank, D. P., Drossart, P., Formisano, V., Jaumann, R.,

Langevin, Y., Matson, D. L., Nelson, R. M., Nicholson, P. D., Sicardy,

B., and Sotin, C. Cassini VIMS observations of the Galilean satellites includingthe VIMS calibration procedure. Icarus 172 (Nov. 2004), 104–126.

[114] Miller, E., Klein, G., Juergens, D., Mehaffey, K., Oseas, J., Garcia,

R., Giandomenico, A., Irigoyen, B., Flickok, R., Rosing, D., Sobel, H.,

Bruce, C., Flamini, E., DeVidi, R., Reininger, F., Dami, M., Soufflot,

A., Langevin, Y., and Huntzinger, G. The visual and infrared mappingspectrometer for CASSINI. SPIE 2803 (1996), 206.

[115] Momary, T. W., and Baines, K. H. The Zoology of Saturn: The bizarre FeatureUnveiled by the 5 Micron Eyes of Cassini/VIMS. In AAS/Division for PlanetarySciences Meeting Abstracts. (Sep. 2006), vol. 38, pp. 11–21.

[116] Moreno, F., Molina, A., and Ortiz, J. L. Recent multispectral observationsof Jupiter from 0.6 to 2.4 µm. J. Geophys. Res. 98 (1993), 18837–18845.

[117] Muñoz, O., Moreno, F., Molina, A., Grodent, D., Gérard, J. C.,

and Dols, V. Study of the vertical structure of Saturn’s atmosphere usingHST/WFPC2 images. Icarus 169 (2004), 413–428.

[118] Mukai, T., and Kratschmer, W. Optical constants of the mixture of ices.Earth, Moon, Planets 36 (1985), 145–155.

BIBLIOGRAPHY 56

[119] Nixon, C. A., Irwin, P. G. J., Calcutt, S. B., Taylor, F. W., and

Carlson, R. W. Atmospheric Composition and Cloud Structure in Jovian 5-µmHotspots from Analysis of Galileo NIMS Measurements. Icarus 150 (Mar. 2001),48–68.

[120] Noll, K. S., Geballe, T. R., and Knacke, R. F. Arsine in Saturn and Jupiter.Astrophys. J. 338 (1989), L71–L74.

[121] Noll, K. S., Knacke, R. F., Geballe, T. R., and Tokunaga, A. T. Evidencefor germane in Saturn. Icarus 75 (1988), 409–422.

[122] Noll, K. S., Knacke, R. F., Geballe, T. R., and Tokunaga, A. T. Detec-tion of carbon monoxide in Saturn. ApJ Letters 309 (Oct. 1986), L91–L94.

[123] Noll, K. S., and Larson, H. P. The spectrum of Saturn from 1990-2230 cm1:abundances of AsH3, CH3D, CO, GeH4, and PH3. Icarus 89 (1990), 168–189.

[124] Noll, K. S., Larson, H. P., and Geballe, T. R. The abundance of AsH3 inJupiter. Icarus 83 (Feb. 1990), 494–499.

[125] North, J. D. Thomas Harriot and the first telescopic observations of sunspots. inThomas Harriot: Renaissance Scientist . Oxford: Clarendon Press, 1974.

[126] Ortiz, J. L., Moreno, F., and Molina, A. Saturn 1991-1993: Clouds andHazes. Icarus 119 (1996), 53–66.

[127] Ortiz, J. L., Moreno, F. J., and Molina, A. Saturn 1991-1993: Reflectivitiesand limb-darkening coefficients at methane bands and nearby continua - Temporalchanges. Icarus 117 (1995), 328–344.

[128] Orton, G., Fisher, B., Yanamandra-Fisher, P., Baines, K., Ressler, M.,

Beach-Kimball, B., B., J., Gezari, D., and Varosi, F. Pre-Cassini infraredcharacterisation of Saturn’ atmosphere. Bull. Am. Astron. Soc. 35 (2003), 1018.

[129] Orton, G. S., Appleby, J. F., and Martonchik, J. V. The effect of ammoniaice on the outgoing thermal radiance from the atmosphere of Jupiter. Icarus 52(1982), 94–116.

[130] Orton, G. S., and Ingersoll, A. P. Saturn’s atmospheric temperature structureand heat budget. Journal of Geophysical Research 85, A11 (1980), 5871–5881.

[131] Owen, T., and Gautier, D. Touring the saturnian system: The atmospheres ofTitan and Saturn. Space Science Reviews 104 (2002), 347–376.

[132] Pérez-Hoyos, S., and Sánchez-Lavega, A. On the vertical wind shear ofSaturn’s equatorial jet at cloud level. Icarus 180 (2006), 161–175.

BIBLIOGRAPHY 57

[133] Pérez-Hoyos, S., Sánchez-Lavega, A., French, R. G., and Rojas, J. F.

Saturn’s cloud structure and temporal evolution from ten years of Hubble SpaceTelescope images (1994-2003). Icarus 176 (2005), 155–174.

[134] Plass, G. N., and Kattawar, G. W. Matrix operator theory of radiative trans-fer. 1: Rayleigh scattering. Applied Optics 12 (1973), 314–329.

[135] Porco, C. C., Baker, E., Barbara, J., Beurle, K., Brahic, A., Burns,

J. A., Charnoz, S., Cooper, N., Dawson, D. D., Del Genio, A. D.,

Denk, T., Dones, L., Dyudina, U., Evans, M. W., Giese, B., Grazier,

K., Helfenstein, P., Ingersoll, A. P., Jacobson, R. A., Johnson, T. V.,

McEwen, A., Murray, C. D., Neukum, G., Owen, W. M., Perry, J.,

Roatsch, T., Spitale, J., Squyres, S., Thomas, P., Tiscareno, M., Tur-

tle, E., Vasavada, A. R., Veverka, J., Wagner, R., and West, R. CassiniImaging Science: Initial Results on Saturn’s Atmosphere. Science 307 (2005), 1243–1247.

[136] Price, M. J. Limb brightening on Uranus: An interpretation of the λ7300 Åmethane band. Icarus 35 (1978), 93–98.

[137] Rages, K., beebe, R., and Senske, D. Jovian stratospheric hazes: The highphase angle view from Galileo. Icarus 139 (1999), 211–226.

[138] Rayner, J. T., Cushing, M. C., and Vacca, W. D. The InfraRed TelescopeFacility (IRTF) spectral library: Cool stars. ApJS 185 (2009), 289–432.

[139] Robertson, C. W., and Downing, H. D. Optical constants of solid ammoniain the infrared. J. Opt. Soc. AM. 65 (1975), 432–435.

[140] Rodgers, C. D. Inverse methods for atmospheric sounding: theory and practice.Singapore: World Scientific, 2000.

[141] Roos-Serote, M., Drossart, P., Encrenaz, T., Lellouch, E., Carlson,

R. W., Baines, K. H., Kamp, L., Mehlman, R., Orton, G. S., Calcutt, S.,

Irwin, P., Taylor, F., and Weir, A. Analysis of Jupiter North Equatorial Belthot spots in the 4-5-µm range from Galileo/near-infrared mapping spectrometerobservations: Measurements of cloud opacity, water, and ammonia. Journal ofGeophysical Research 103 (Sep. 1998), 23023–23042.

[142] Roos-Serote, M., and Irwin, P. G. J. Scattering properties and location of thejovian 5-micron absorber from Galileo/NIMS limb-darkening observations. Journalof Quantitative Spectroscopy & Radiative Transfer 101 (2006), 448–461.

[143] Rossow, W. B. Cloud microphysics: Analysis of the clouds of Earth, Venus, Marsand Jupiter. Icarus 36 (1978), 1–50.

BIBLIOGRAPHY 58

[144] Rothman, L. S., Jacquemart, D., Barbe, A., Benner, D. C., Birk, M.,

Brown, L. R., Carleer, M. R., Chackerian, C., Chance, K., Coudert,

L. H., Dana, V., Devi, V. M., Flaud, J. M., Gamache, R. R., Goldman, A.,

Hartmann, J. M., Jucks, K. W., Maki, A. G., Mandin, J. Y., Massie, S. T.,

Orphal, J., Perrin, A., Rinsland, C. P., Smith, M. A. H., Tennyson, J.,

Tolchenov, R. N., Toth, R. A., Vander Auwera, J., Varanasi, P., and

Wagner, G. The HITRAN 2004 molecular spectroscopic database. Journal ofQuantitative Spectroscopy and Radiative Transfer 96 (Dec. 2005), 139–204.

[145] Sánchez-Lavega, A., Colas, F., Lecacheux, J., Laques, P., Miyazaki, I.,

and Parker, D. The great white spot and disturbances in Saturn’s equatorialatmosphere during 1990. Nature 353 (1991), 397–402.

[146] Sánchez-Lavega, A., Hueso, R., Pérez-Hoyos, S., Rojas, J. F., and

French, R. G. Saturn’s cloud morphology and zonal winds before the Cassiniencounter. Icarus 170 (2004), 519–523.

[147] Sánchez-Lavega, A., Pérez-Hoyos, S., and Hueso, R. Clouds in planetaryatmospheres: A useful application of the Clausius-Clapeyron equation. Am. J.Phys. 72, 6 (2004), 337–352.

[148] Showman, A. P., and de Pater, I. Dynamical implications of Jupiter’s tropo-spheric ammonia abundance. Icarus 174 (2005), 192–2004.

[149] Sill, G., Fink, U., and Ferraro, J. R. Absorption coefficients of solid NH3

from 50 to 7000 cm−1. J. Opt. Soc. AM. 70 (1980), 724–739.

[150] Sinton, W. M. Limb and polar brightening of Uranus at 8870 Å. The AstrophysicalJournal 176 (1972), 131–133.

[151] Slobodkin, L. S., Buyakov, I. F., Cess, R. D., and Caldwell, J. NearInfrared reflection spectra of ammonia frost: Interpretation of the upper clouds ofSaturn. J. Quant. Spectrosc. Radiat. Transfer 20 (1978), 481–490.

[152] Smith, B. A., Soderblom, L., Beebe, R. F., Boyce, J. M., Briggs, G.,

Bunker, A., Collins, S. A., Hansen, C., Johnson, T. V., Mitchell, J. L.,

Terrifle, R. J., Carr, M. H., Cook, A. F., Cuzzi, J., Pollack, J. B.,

Danielson, G. E., Ingersoll, A. P., Davies, M. E., Hunt, G. E., Ma-

sursky, H., Shoemaker, E. M., Morrison, D., Owen, T., Sagan, C., Vev-

erka, J., Strom, R., and Suomi, V. E. Encounter with Saturn - Voyager 1imaging science results. Science 212 (1981), 163–191.

[153] Sromovsky, L. A. Effects of Rayleigh-scattering polarization on reflected inten-sity: A fast and accurate approximation method for atmospheres with aerosols.Icarus 173 (2005), 284–294.

BIBLIOGRAPHY 59

[154] Sromovsky, L. A., and Fry, P. M. The source of widespread 3-µm absorptionin Jupiter’s clouds: Constraints from 2000 Cassini VIMS observations. Icarus 210(Nov. 2010), 230–257.

[155] Stam, D. M., Banfield, D., Gierasch, P. J., Nicholson, P. D., and

Matthews, K. Near-IR Spectrophotometry of Saturnian Aerosols-Meridional andVertical Distribution. Icarus 152 (2001), 407–422.

[156] Strobel, D. The photochenistry of NH3 in the Jovian atmosphere. Journal of theatmospheric sciences 30 (1973), 1205–1209.

[157] Strobel, D. F. Photochemistry of the reducing atmospheres of Jupiter, Saturnand Titan. Int. Rev. Phys. Chem. 3 (1983), 145–176.

[158] Tarrago, G., Lacome, Levy, A., Guelachvili, G., Bezard, B., and

Drossart, P. Phosphine spectrum at 4-5 µm : Analysis and line-by-line sim-ulation of 2ν2, ν2 + ν4, ν1, and ν3 bands. . Journal of molecular spectroscopy 154,1 (1992), 30–42.

[159] Temma, T., Chanover, N. J., Simon-Miller, A. A., Glenar, D. A., Hill-

man, J. J., and Kuehn, D. M. Vertical structure modeling of Saturn’s equatorialregion using high spectral resolution imaging. . Icarus 175 (2005), 464–489.

[160] Tomasko, M. G., and Doose, L. R. Polarimetry and photometry of Saturnfrom Pioneer 11: Observations and constraints on the distribution and propertiesof cloud and aerosol particles. Icarus 58 (1984), 1–34.

[161] Tomasko, M. G., Karkoschka, and Martinek, S. Observations of the limbdarkening of Jupiter at ultraviolet wavelengths and constraints on the propertiesand distribution of stratospheric aerosols. Icarus 65 (1986), 218–243.

[162] Trafton, L. Saturn: Long-term variation of H2 and CH4 absorptions. Icarus 31(1977), 369–384.

[163] Trafton, L. M. Long term changes in Saturn’s troposphere. Icarus 63 (1985),374–405.

[164] van de Hulst, H. C. Light scattering by small particles. Dover Publications, NewYork, 1981.

[165] Vasavada, A. R., HŽrst, S. M., Kennedy, M. R., Ingersoll, A. P., Porco,

C. C., Del Genio, A. D., and West, R. A. Cassini imaging of Saturn: Southernhemisphere winds and vortices. Journal of Geophysical Research 111, E10 (2006),5004.

[166] Weidenschilling, S. J., and Lewis, J. S. Atmospheric and cloud structures ofthe Jovian planets. Icarus 20 (1973), 465–476.

BIBLIOGRAPHY 60

[167] West, R. A. Spatially resolved methane band photometry of Saturn. II. Cloudstructure models at four latitudes. Icarus 53 (1983), 301–309.

[168] West, R. A. Optical properties of aggregate particles whose outer diameter iscomparable to the wavelength. Appl. Opt. 30 (1991), 5316–5324.

[169] West, R. A., and Smith, P. H. Evidence for aggregate particles in the atmo-spheres of Titan and Jupiter. Icarus 90 (1991), 330–333.

[170] West, R. A., Strobel, D. F., and Tomasko, M. G. Clouds, aerosols, andphotochemistry in the Jovian atmosphere. Icarus 65 (1986), 161–217.

[171] West, R. A., Tomasko, M. G., Smith, B. A., Wuesinghe, M., Doose,

L. R., Reitsema, H. J., and Larson, S. M. Spatially resolved methane bandphotometry of Saturn. I. Absolute reflectivity and center-to-limb variations in the6190-, 7250-, and 8900-Å bands. Icarus 51 (1982), 51–64.

[172] Witteborn, F. C., Pollack, J. B., Bergman, J. D., Goebel, H. J., Soifer,

B. T., Puetter, R. C., Rudy, R. J., and Willner, S. P. Observations ofSaturn in the 5 to 8 µm spectral region. Icarus 45 (1981), 653–660.

[173] Wong, A., Yung, Y. L., and Friedson, A. J. Benzene and haze formation inthe polar atmosphere of Jupiter. Geophysical research letter 30, 8 (2003).

[174] Wong, A. S., Lee, A., Yung, Y. L., and Ajello, J. M. Jupiter: Aerosolchemistry in the polar atmosphere. Astrophys. J. 534 (2000), 215–217.

[175] Yanamandra-Fisher, P. A., Orton, G. S., Fisher, B. M., and Sanchez-

Lavega, A. Saturn’s 5.2-µm cold spots: Unexpected cloud variability . Icarus 150(2001), 189–193.

clouds and haze in saturn’s troposphere · clouds and haze in saturn’s troposphere cécile...

Documents