Cassini radar views the surface of Titan.

Abstract

The Cassini Titan Radar Mapper imaged about 1% of Titan's surface at a resolution of approximately 0.5 kilometer, and larger areas of the globe in lower resolution modes. The images reveal a complex surface, with areas of low relief and a variety of geologic features suggestive of dome-like volcanic constructs, flows, and sinuous channels. The surface appears to be young, with few impact craters. Scattering and dielectric properties are consistent with porous ice or organics. Dark patches in the radar images show high brightness temperatures and high emissivity and are consistent with frozen hydrocarbons.

Full text

Page 1

trometer data using stellar occultations after the
13 December flyby. The chemical composition
of the upper atmosphere of Titan along the
spacecraft track, derived from mass spectrom-
eter measurements, is reported by Waite et al.
on page 982 (7).
The desire to understand the formation and
evolution of extrasolar planetary systems also
motivates the detailed study of the Saturn sys-
tem. Cassini observations will not only test and
validatemodels oftheprocessesthatshapedthe
saturnian system into its collection of rings,
satellites, and plasma; they will also provide an
improved theoretical basis for understanding
the conditions necessary for the emergence of
planetary systems around young stars.
The in situ measurements provided by
Cassini_s flight through the upper atmosphere of
Titan may constrain models of how this moon
formed (11). For example, if CO and N2were
dominant in the formation region of Saturn,
clathrates (inclusion compounds within a host
lattice such as water) of these molecules, together
withtheprimordialandchemicallyinert36Ar and
38Ar that were present during Titan_s formation,
might be incorporated into the planet. Alterna-
tively, the source of the present N2atmosphere
mightbe NH3that was later transformed into ni-
trogen. Measurements of the36Ar and38Arargon
isotopes relative to the radiogenic40Ar (released
later into this atmosphere from the decay of40K)
may distinguish between these alternatives. Like-
wise, tests of how much of Titan_s atmosphere
may have been lost over its history can come
from measurements of the isotope ratios15N/14N,
because loss to space fractionates these isotopes,
with preferential loss of the lighter14N. The
Cassini Orbiter mass spectrometer data (7), com-
bined with Huygens Probe mass spectrometer
(12) measurementsofthesesameisotoperatiosin
the deep atmosphere, will provide better observa-
tional constraints on models of Titan formation.
The Cassini Orbiter/Huygens Probe mission
providesanextraordinaryexampleofasuccessful
international collaboration in space exploration.
The navigation (9) of this sophisticated science
platform precisely to its targeted destination in
Titan_s upper atmosphere, the successful deploy-
ment and entry of the Huygens Probe, the solid
performance from the large number of instru-
ments, and the demonstration of the ability to
flawlessly plan and execute complex measure-
ment sequences are a tribute to the skill, ded-
ication, and perseverance of the multinational
Cassini team. Future Titan flybys will incremen-
tally build up the map of Titan_s surface and
sample the atmosphere at different locations and
seasons. The data from the first flyby reported
here (1–7) provide a preview of the insights that
can be expected from the next several years of ex-
ploration and are the first steps toward a substan-
tially deeper understanding of this distant world.
References
1. C. Elachi et al., Science 308, 970 (2005).
2. F. M. Flasar et al., Science 308, 975 (2005).
3. H. Backes et al., Science 308, 992 (2005).
4. D. G. Mitchell et al., Science 308, 989 (2005).
5. J. E. Wahlund et al., Science 308, 986 (2005).
6. D. E. Shemansky et al., Science 308, 978 (2005).
7. J. H. Waite Jr. et al., Science 308, 982 (2005).
8. M. G. Kivelson et al., Science 289, 1340 (2000).
9. R. D. Lorenz, J. Mitton, Lifting Titan’s Veil (Cambridge
Univ. Press, Cambridge, UK, 2002).
10. A. A. Wolf, Space Sci. Rev. 104, 101 (2002).
11. T. Owen, D. Gautier, Space Sci. Rev. 104, 347 (2002).
12. H. B. Niemann et al., Space Sci. Rev. 104, 551 (2002).
10.1126/science.1113205
R E P O R T
Cassini Radar Views the Surface of Titan
C. Elachi,1S. Wall,1* M. Allison,2Y. Anderson,1R. Boehmer,1P. Callahan,1P. Encrenaz,3E. Flamini,4G. Franceschetti,5
Y. Gim,1G. Hamilton,1S. Hensley,1M. Janssen,1W. Johnson,1K. Kelleher,1R. Kirk,6R. Lopes,1R. Lorenz,7J. Lunine,7
D. Muhleman,8S. Ostro,1F. Paganelli,1G. Picardi,9F. Posa,10L. Roth,1R. Seu,9S. Shaffer,1L. Soderblom,6
B. Stiles,1E. Stofan,11S. Vetrella,5R. West,1C. Wood,12L. Wye,13H. Zebker13
The Cassini Titan Radar Mapper imaged about 1% of Titan’s surface at a resolution of
È0.5 kilometer, and larger areas of the globe in lower resolution modes. The images
reveal a complex surface, with areas of low relief and a variety of geologic features
suggestive of dome-like volcanic constructs, flows, and sinuous channels. The surface
appears to be young, with few impact craters. Scattering and dielectric properties are
consistent with porous ice or organics. Dark patches in the radar images show high
brightnesstemperaturesand high emissivityandareconsistent withfrozenhydrocarbons.
Saturn_s largest satellite, Titan, is the only
moon and one of only four solid bodies in the
solar system to host a thick atmosphere,
uniquely distinguished by nitrogen and meth-
ane and a complex suite of organic products
of these molecules. Its primitive chemistry
may reveal clues about the prebiotic origin
of materials that ultimately gave birth to
life in our solar system. The mapping of
Titan is an especially challenging puzzle
because the most likely constituent materials
(e.g., water-ammonia and other ices, hydro-
carbons, tholins) in this chemical and temper-
ature regime are likely to exhibit different
scattering properties than at Earth and Venus,
the only other worlds mapped by spaceborne
radar (1–3).
C A S S I N I R E V E A L S T I T A N
C A S S I N I R E V E A L S T I T A N
13 MAY 2005 VOL 308SCIENCE www.sciencemag.org
970
SPECIAL SECTION

Page 2

The Cassini mission_s Titan Radar Mapper
operatesinfourmodes:radiometry,scatterome-
try, altimetry, and synthetic aperture radar
(SAR) imaging (4). SAR mode requires ranges
less than È4000 km; high-resolution SAR
(È500 m/pixel) can be collected only below
È1500km.Datareportedherearefromthefirst
close flyby, designated TA, which occurred
on 26 October 2004 at a minimum range of
1174 km, the first of 45 planned close flybys (4).
Inbound full-disk radiometry and scatterometry
scans of large regions, including the Huygens
probe landing site (5, 6), were acquired from
È25,000 km at differing angles. SAR coverage
extending through closest approach over an arc
of È100- was followed immediately by an
altimetry track covering an arc of È10-. Scat-
terometry scans and radiometry scans in
orthogonal linear polarizations were collected
outbound. Observation geometries and details,
coverage diagrams, and higher resolution
images are available in (4).
The SAR image strip (Fig. 1; Fig. 2 in
detail) reveals a variable surface with distinct
units and features. On the basis of brightness,
texture, and morphology, we define five units,
two of regional scale and three of local scale.
About one-quarter of the swath is composed
of a regional-dark, homogeneous unit with
relatively low backscatter (see the bottom half
of Fig. 2A and parts of Fig. 2, D and E),which
includes regions of variable backscatter with
indistinct boundaries and scattered SAR-bright
spots that may be small topographic features.
Most of the rest of the swath consists of a
higherbackscatter,regional-mottledunitthatis
gradational at its boundaries with the darker,
homogeneous unit. Two discrete SAR-bright
units are distinguished. A bright lobate unit
(Fig. 2E) occurs in several locations across the
swath. The lobate boundaries range from
distinct to gradational, withvariable brightness
across the unit. Both sheet-like and more
digitate examples appear, sometimes asso-
ciated with quasi-circular features. A bright,
lineated unit exhibits sharp boundaries and
bright, lineated interiors (Fig. 2C). The linea-
tions are parallel and are spaced 1 to 2 km
apart. This unit appears similar to grooved ter-
rainsseenatseveralscalesonGanymede(7, 8).
The fifth unit is SAR-dark and occurs as iso-
lated patches a few tens of kilometers or less
across (Fig. 2, B to D). One such dark patch is
near the western end of the swath. These
patches tend to be irregular, with sharp bound-
aries; many are interconnected by narrow
sections, giving the impression that their dis-
tribution is topographically controlled, perhaps
by depressions, valleys, and channels.
A large, roughly circular feature about
180 km in diameter is prominent in the SAR
strip, centered near 50-N, 87-W (Fig. 2B, at
left). The International Astronomical Union
(IAU) has provisionally approved the name
Ganesa Macula for this feature. It has sugges-
tions of a volcanic construct, possibly a dome;
it could also occupy an old impact scar. Along
its edges are SAR-bright flanks; at its center,
instead of a central peak, is an apparent central
depression, about 20 km in its longest dimen-
sion, that resembles a volcanic crater. Several
sinuouschannelsand/orridgesradiatefromthe
central feature; these appear to carry material
from the center. The 90-km sinuous features
trending southeast could be responsible for
transport of material to the SAR-bright regions
along the flank. The terrain along the eastern
margin is complex, exhibiting multiple lobate
forms that may be flows.
Aplotofthescatterometrybackscattercross
section versus incidence angle shows both
specular and diffuse-scattering echoes (Fig.
3A). The region where specular reflections
could be observed was restricted by the Titan-
spacecraft geometry to the central part of the
inbound scatterometer raster map, but diffuse
scattering was observable over a larger area,
and the plot indicates considerable heterogene-
ity. Figure 3B shows a crude classification of
the scatterometer measurements into high,
medium, and low diffuse cross-section classes,
with their locations plotted on Cassini Imaging
Science Subsystem (ISS) imagery (9). This
model-independent classification shows cross
sections clearly correlated with the near-
infrared (near-IR) albedo observed by ISS.
Figure 4A shows a map of the cross section
normalized to constant incidence angle by a
simple cosecant model, revealing a strong cor-
relation with the ISS image (Fig. 4C) in the
low incidence angle region as well. Xanadu,
the bright, continent-sized region at 90- to
150-W, shows the strongest backscatter, in-
cluding a particularly bright spot near 120-W
that is undistinguished in the near-IR.
The passive radiometer mode acquired
an inbound unresolved disk brightness mea-
surement and two low-resolution outbound
microwave brightness maps. Disk-averaged
brightness temperatures from these observa-
tions are 85 K, 85 K, and 86 K (each T5 K),
respectively, consistent with Earth-based mea-
surements (10, 11). Assuming a physical
surface temperature of 94 K, these data
Fig. 1 (on facing pages). SAR image strip acquired by Cassini’s Titan
Radar Mapper during the Titan TAflyby, with 5- latitude ? 10- west
longitude graticule. Resolution is È500 m at image center, increasing to
È1400 m at eastern end. The image has been corrected to constant
incidence angle with the use of the backscatter model shown in Fig. 3A
and logarithmically scaled. Letters indicate regions enlarged in Fig. 2.
1Jet Propulsion Laboratory, California Institute of
Technology, Pasadena, CA 91109, USA.
Institute for Space Studies, National Aeronautics and
Space Administration, New York, NY 10025, USA.
3l’Observatoire de Paris, 75014 Paris, France.4Agen-
zia Spaziale Italiana, 00198 Rome, Italy.5Facolta ´ di
Ingegneria, 80125 Naples, Italy.
Survey, Flagstaff, AZ 86001, USA.7Lunar and Plan-
etary Laboratory, University of Arizona, Tucson, AZ
85721, USA.
Sciences, California Institute of Technology, Pasade-
na, CA 91125, USA.9Universita ´ La Sapienza, 00184
Rome, Italy.10INFM and Dipartimento Interateneo di
Fisica, Politecnico di Bari, 70126 Bari, Italy.
11Proxemy Research, Laytonsville, MD 20882, USA.
12Planetary Science Institute, Tucson, AZ 85719, USA.
13Stanford University, Stanford, CA 94305, USA.
2Goddard
6U.S. Geological
8Division of Geological and Planetary
*To whom correspondence should be addressed.
E-mail: steve.wall@jpl.nasa.gov
C A S S I N I R E V E A L S T I T A N
C A S S I N I R E V E A L S T I T A N
www.sciencemag.orgSCIENCE VOL 30813 MAY 2005
971
SPECIAL SECTION

Page 3

correspond to an overall surface emissivity of
È0.9. For a dielectric surface that can be
characterized by the Fresnel equations, this is
consistent with a dielectric constant of 2 to 3.
The polarization and center-to-limb brightness
distributions of the full-disk radiometerimages
are consistent with a uniform surface temper-
ature and a bulk dielectric constant e 0 2, and
areinconsistentwithsolidwaterice(e03.1)or
ammonia ice (e 0 4.5). In particular, the
polarization differences yield broadly uniform
solutions for e 0 2.1 and 1.6, respectively, in
northernandsouthernmid-latituderegions.The
radiometry maps show considerable bright-
ness temperature variations e10 K between
regionse500kminextentthatcanbeattributed
to emissivity variations (Fig. 4B). These varia-
tionsarenegativelycorrelatedwiththoseofthe
backscatter cross section.
Altimetry data were analyzed by a simple
threshold algorithm and reduced to elevations
relative to a sphere (Fig. 5); absolute radii are
awaiting more accurate trajectory data. Relief
along the altimetry track covered in this north-
ern region is quite low, varying only È150 m
over 400 km. This is a small sample (0.02% of
the surface) and does not currently coincide
with SAR or optical coverage. However,
Cassini radiometry and scatterometry maps of
the outbound hemisphere (4) and ISS images
(9) indicate that the altimetry does not cross
any notable terrain boundaries. The SAR
images also provide information about topo-
graphicreliefofresolvedgeologicfeatures,but
without the geometric rigor of the altimetry. A
fewfeaturesintheSARswathexhibittheclose
bright-dark downrange pairing that is charac-
teristicoftopographicshading.Radarclinomet-
ric analysis allows crude estimates of heights.
The strongest sense of relief is given by small
apparent hills, 5 to 10 km in width (Fig. 2F).
Many of these show bright-dark pairing in the
downrange direction, making it unlikely that
they are flat surfaces of variable scattering.
They were imaged at a lower incidence angle
relative to most of the strip, maximizing topo-
graphic modulation.Weestimate that the relief
is only È100 m and slopes are È3- to 5- (12).
Applyingthisapproachtoaflow-likefeatureat
41-W, 47-N indicates a height of È250 m,
bounding slopes of È7-, and a flat top.
The potential pitfalls of overinterpreting
brightness as slopes may be illustrated by the
Fig. 2. Enlargements of the Cassini SAR image shown in Fig. 1. TheletterscorrespondtothelocationsshownalongtheedgeoftheSARstripinFig.1. Each
image covers an area È175 km high, with north approximately at top and illumination approximately from bottom. See text.
Fig. 3. Angular and spatial depen-
dence of Cassini Titan TAinbound
scatterometer data. (A) Plot of
backscatter cross section versus
incidence angle shows specular
behavior at small incidence angles
(blue) and dropoff at largest
incidence angles (red). Data at
intermediate angles (green) show
multiple populations (i.e., surface
units) with distinct and only
weakly incidence-dependent cross
section. Data have been classified
into low, medium, and high cross
section (dark to light green).
Curve is simple cosecant (inci-
dence) backscatter model used
to normalize images in Figs. 1
and2 andscatterometryinFig.3A.
(B) Map of scatterometer mea-
surements overlaid on Cassini ISS
image map (9), colored according
to incidence angle and backscatter
classification as in (A). High- and
low-incidence classes are geo-
graphically restricted because of
observation geometry, but cross-
section classes at intermediate
angles correlate clearly with near-
IR albedo.
C A S S I N I R E V E A L S T I T A N
C A S S I N I R E V E A L S T I T A N
13 MAY 2005 VOL 308SCIENCE www.sciencemag.org
972
SPECIAL SECTION

Page 4

large, bright, flow-like feature shown in Fig.
2E. Backscatter varies linearly across this
feature, so if it is topographic, it is roughly
parabolic in cross section, with maximum
slopes of T4-; the height of the 70-km–wide
eastern part is È1000 m. These dimensions are
comparable to thick flows on Ariel (13) and
would require an implausibly large amount of
geothermal heat to be expressed over time as
cryomagma. An alternative interpretation that
we prefer, given the diffuse appearance of the
feature, is that brightness is related to a
compositional or textural gradient on the
surface of a much thinner flow.
GivenTitan_s unique combination of atmo-
sphere and temperature, the interpretation of
these data is necessarily speculative until
coverage is increased and is obtained from
multiple instruments. Most notable is the lack
of unambiguous impact craters visible in the
SAR strip. Objects producing craters smaller
than È20 km would be screened by Titan_s
present atmosphere (14, 15). Heavily cratered
terrains on other large saturnian satellites
typically have 200 to 400 craters per 106km2
larger than È20 km (16); we should expect to
find 100 to 200 craters here if this were an
ancient surface. With no fresh craters and only
a few degraded candidates, the resurfacing rate
must be considerably higher than the impact
cratering rate.Weconclude,therefore, that this
part of Titan_s surface is geologically young.
Crater obliteration may be due to burial, either
by deposition of stratospheric photochemical
debris or direct accretion and windblown dep-
osition; pluvial erosion or deposition; tectonic
processes; or cryovolcanism. The lobate units,
channels, and other flow-like features suggest
pluvial or volcanic processes; the presence of
apparent volcanic constructs and a simple cal-
culation of available resurfacing materials (17)
favor cryovolcanism.
The SAR and scatterometer observations
are consistent with previous Earth-based data,
even though the wavelengths and the specific
areas observed differ. Arecibo (wavelength 0
13 cm) observations of Titan show a diffusely
scattered component, with a weak specular
component present for about 75% of the
observation (18), indicating smooth surfaces
with slopes as shallow as 0.2- and Fresnel
reflectivities of several percent. The specular
component has not been seen in radar obser-
vations of other icy solar system bodies; it
requires the existenceof regions that are nearly
flat at subcentimeter scales. However, diffuse
reflection of most of the echo power requires
near-surfacestructuralcomplexity(e.g.,rough-
ness or fractures and voids) at scales Q0.5 cm.
Surface roughness alone from geochemically
plausible minerals would be too weak to
produce the observed echoes. Therefore, al-
though surface roughness per se may contrib-
ute to variations in the radar brightness, we
concurwithCampbellandco-workers(18) and
attribute the bulk of Titan_s radar echo (away
from the specular peak in Fig. 3) to volume
scattering, in which case multiple scattering in
icy materials is known to be able to produce
strong returns.
Titan_s scatterometry, with both specular
and diffuse components, in some ways resem-
bles those of the Moon and the inner planets
(19), albeit at a much higher magnitude. Such
large variations in radar brightness in both
components suggest that Titan_s near-surface
structure and composition are heterogeneous.
Volume scattering depends critically on the
transparencyofthemedium(20, 21) andwould
be much greater for nearly pure water ice Eas
on the icy Galilean satellites (22)^ than if ab-
sorbing material such as ammonia or silicates
were present, as may be the case for Iapetus
(23) and comets (24). Therefore, variability in
composition remains an important candidate
to explain variations in radar reflectivity, al-
thoughwebelievevolumescatteringinlow-loss
media to be an important effect, particularly
in the SAR-bright regions.
At global scale, Titan differs from cold, icy
surfaces suchas Europa and Ganymede, where
microwave emissivities of 0.5 to 0.6 and high
radar reflectivity indicate the dominance of
Fig. 4. Comparison of radar scatterometry and radiometry with near-IR albedo for a subset of the
region in Fig. 3. Simple cylindrical projection is shown, with north at top. Star indicates approximate
landingsiteofHuygensprobe.(A) Backscatter cross section s0normalized to constant incidence angle
according to model curve from Fig. 3A. Black indicates data gaps where scan raster separation
exceeded beam width. (B) Deviation of brightness temperature T0from model, plotted with reversed
grayscale to facilitate comparison. (C) Portion of ISS image mosaic (9). Negative correlation between
scatterometry and radiometry is near-perfect, given slightly coarser resolution of latter. Radar-ISS
correlationisalsostrongbuthasnoteworthyexceptionssuchasextremelyradar-bright,low-emissivity
spot within Xanadu Regio at 120-W, which is undistinguished optically.
Fig. 5. Cassini radar altimetry
profile. Leftmost data shown are
near 10-W, 29-N, trending south-
southeast. Raw elevations relative
to a 2575-km sphere (gray) have
been smoothed by convolution
with a Gaussian function with
standard deviation È7 km (solid
black line) to suppress the point-
to-point variations (root mean
square È20 m) due to noise; bi-
directional slopes were calculated
from the smoothed elevations.
Horizontal resolution is limited by
beam width to È25 km. Only
elevation changes within the pro-
file are significant, given current
errors in trajectory information used to reduce the data.
C A S S I N I R E V E A L S T I T A N
C A S S I N I R E V E A L S T I T A N
www.sciencemag.orgSCIENCEVOL 308 13 MAY 2005
973
SPECIAL SECTION

Page 5

volume scattering from the subsurface rather
thanfromasimpledielectrichalfspace(25). A
much more complicated model for Titan, in-
volving multiple surface layers and composi-
tional heterogeneities, may be required to
reconcile the Cassini scatterometry and radi-
ometry observations. All materials posed as
candidates for Titan_s surface other than those
containing ammonia are expected to have very
low microwave absorption at temperatures
È100 K, and would have to be tens of meters
deep to produce the observed emission.
The strong correlation between radiometry
and optical data (Fig. 4, B and C) is not
understood.Thestrongcorrelationbetweenthe
SAR-bright and SAR-dark features and radio-
metrically cold and warm regions,respectively
(Fig. 4, A and B), also suggests that volume
scattering is one cause of the SAR-bright
regions. If the SAR brightness were solely
due to slopes, the correlation with brightness
temperature would be positive. Surface rough-
ness alone would have a small effect on
thermal emission at the SAR viewing angle.
A possible hypothesis is that (i) the optically
brighter and colder areas are fractured or
porous ice, producing volume scattering and
resulting low emissivity, as on the surfaces of
Europa and Ganymede; and (ii) the optically
darker, warmer, high-emissivity areas have a
lower average dielectric constant, being com-
posed of either an organic sludge (26, 27) or a
porous material of high dielectric constant,
such as water ice.
The location of reservoirs of organic liquid
is unclear, although patches of SAR-dark
material occupy È1% of the swath and may
well represent concentrations of organics. So
far, these patches have been too small to
resolve with radiometry. Shallow deposits of
liquid hydrocarbons could, however, be trans-
parent to microwaves, as is the case for cold
and pure water ice (28). Alternatively, the
liquid hydrocarbons could be masked by
floating solids. All materials likely on Titan
would sink in liquid ethane and methane, with
the exception of the polymerized product of
acetylene(polyacetylene),itselfthemostabun-
dant solid product of stratospheric photolysis
(29). Some forms of polyacetylene exhibit mi-
croporous structures with density È0.4 g/cm3,
lower than any expected combination of
methane, ethane, and nitrogen in surface liq-
uid form (28, 29). This material is a weak
semiconductor and, if suspended or floating on
liquids, it could appear relatively bright to
radar and optical sensing and would contribute
enough small-scale roughness to suppress the
specular reflection at optical wavelengths.
From even this limited view of its smallest
details, it is clear that Titan_s surface is com-
plex and young, although we note again that
only 1% of it has been imaged by SAR. As we
progress through the mission, we expect SAR
coverage of morethan25% of the planet, near-
global radiometry and scatterometry coverage,
and tens of altimetric profiles, providing
increasingly instructive information about the
processes at work on Titan.
References and Notes
1. B. Campbell, Radar Remote Sensing of Planetary
Surfaces (Cambridge Univ. Press, Cambridge, 2002).
2. J. McCauley et al., Science 218, 1004 (1982).
3. R. Saunders et al., J. Geophys. Res. 97, 13067 (1992).
4. Instrument detail, observation geometries and details,
resurfacing rate, radarclinometry methods, a coverage
diagram, and related references are available as sup-
porting material on Science Online.
5. The Huygens probe descended on 14 January 2005,
reaching the surface at about 10.5-S, 192-W. See (30).
6. Titan geographical coordinates are per the IAU (31):
In general terms, the equator is roughly Titan’s orbital
plane, and zero longitude is the subsaturn point, with a
libration of T3- due to orbital eccentricity.
7. R. T. Pappalardo et al., in Jupiter: The Planet, Satellites,
and Magnetosphere, F. Bagenal Ed. (Cambridge Univ.
Press, Cambridge, 2004), pp. 363–396.
8. J. G. Patel et al., J. Geophys. Res. 104, 24,057 (1999).
9. C. C. Porco et al., Nature 434, 159 (2005).
10. B. J. Butler, M. Gurwell, Bull. Am. Astron. Soc. 36,
1075 (2004).
11. A. Grossman, D. Muhleman, Bull. Am. Astron. Soc.
24, 954 (1992).
12. This apparent asymmetry could result from the
actual backscatterbehaviordepartingfromtheassumed
1/sin(i) form at small incidence angles and/or from the
finite resolution of the images. Even allowing for these
effects, the relief is likely not much more than 100 m,
and slopes do not approach the incidence angle, so that
the features are not severely distorted by layover.
13. D. G. Jankowski, S. W. Squyres, Science 244, 1322
(1988).
14. J. Lunine,N.A. Artemieva,R. D. Lorenz,E. Flamini, paper
presented at the 36th Annual Lunar and Planetary
Science Conference, League City, TX, 14–18 March
2005, abstract 1504.
15. S. Engel, J. I. Lunine, W. K. Hartmann, Planet. Space
Sci. 43, 1059 (1995).
16. R. Lorenz, Planet. Space Sci. 45, 1009 (1997).
17. R. Lorenz, Planet. Space Sci. 44, 1021 (1996).
18. Campbell et al. (32) conducted observations with the
Arecibo and Goldstone radio telescopes in November
and December 2001 and late in 2002, obtaining disk-
integratedecho spectraatlongitudes23- apart at26-S.
19. S. Ostro, Rev. Mod. Phys. 65, 1235 (1993).
20. K. Peters, Phys. Rev. B 46, 801 (1992).
21. G. Black, D. Campbell, P. Nicholson, Icarus 151, 167
(2001).
22. S. Ostro et al., J. Geophys. Res. 97, 18227 (1992).
23. G. J. Black, D. B. Campbell, L. M. Carter, S. J. Ostro,
Science 304, 553 (2004).
24. J. Harmon, D. Campbell, S. Ostro, M. Nolan, Planet.
Space Sci. 47, 1409 (1999).
25. D. Muhleman, A. Grossman, B. Butler, Annu. Rev.
Earth Planet. Sci. 23, 337 (1995).
26. We refer to the complex mix of nitriles and liquid
and solid hydrocarbons produced by Titan’s atmo-
spheric photochemistry as ‘‘sludge.’’ Dielectric con-
stants vary, increasing with increasing mass density,
from about 1.6 for methane to 3 or more for higher
molecular weight materials, but 2.0 to 2.4 is a rep-
resentative range for the mix. See (33).
27. R. Thompson, Icarus 86, 336 (1990).
28. Models for liquid compositions on Titan’s surface are
discussed in (34).
29. The density of polyacetylene is discussed in (35).
30. M. Tomasko et al., Nature, in press.
31. P. K. Seidelmann et al., Celest. Mech. Dynam. Astron.
82, 83 (2002).
32. D. B. Campbell, G. J. Black, L. M. Carter, S. J. Ostro,
Science 302, 431 (2003).
33. R. Lorenz et al., Planet. Space Sci. 51, 353 (2003).
34. J. Lunine, Rev. Geophys. 31, 133 (1993).
35. D. Bott, in Handbook of Conducting Polymers, T.
Skotheim, Ed. (Dekker, New York, 1986), pp. 1191–1232.
36. We gratefully acknowledge the many years of work
by many hundreds of people in the development and
design of the Cassini mission that have culminated in
the results presented herein. The Cassini Project is a
joint endeavor of NASA, the European Space Agency
(ESA), and the Italian Space Agency (ASI). Cassini is
managed by the Jet Propulsion Laboratory, California
Institute of Technology, under a contract with NASA.
Supporting Online Material
www.sciencemag.org/cgi/content/full/308/5724/970/
DC1
Materials and Methods
Figs. S1 to S4
Table S1
References
18 January 2005; accepted 25 April 2005
10.1126/science.1109919
C A S S I N I R E V E A L S T I T A N
C A S S I N I R E V E A L S T I T A N
13 MAY 2005 VOL 308SCIENCEwww.sciencemag.org
974
SPECIAL SECTION